Thiazolides Elicit Anti-Viral Innate Immunity and Reduce HIV Replication

Nitazoxanide (Alinia®, NTZ) and tizoxanide (TIZ), its active circulating metabolite, belong to a class of agents known as thiazolides (TZD) endowed with broad anti-infective activities. TIZ and RM-4848, the active metabolite of RM-5038, were shown to stimulate innate immunity in vitro. Because natural resistance to HIV-1 infection in HIV-exposed seronegative (HESN) individuals is suggested to be associated with strong innate immune responses, we verified whether TIZ and RM-4848 could reduce the in vitro infectiousness of HIV-1. Peripheral blood mononuclear cells (PBMCs) from 20 healthy donors were infected in vitro with HIV-1BaL in the presence/absence of TIZ or RM4848. HIV-1 p24 were measured at different timepoints. The immunomodulatory abilities of TZD were evaluated by the expression of type I IFN pathway genes and the production of cytokines and chemokines. TZD drastically inhibited in vitro HIV-1 replication (>87%). This was associated with the activation of innate immune responses and with the up-regulation of several interferon-stimulated genes (ISGs), including those involved in cholesterol pathway, particularly the cholesterol-25 hydroxylase (CH25H). TZD inhibition of HIV-1 replication in vitro could be due to their ability to stimulate potent and multifaceted antiviral immune responses. These data warrant the exploration of TZD as preventive/therapeutic agent in HIV infection.

to this virus 21,22 . TLR7 and 8 stimulation, in particular, was shown to elicit strong innate immune responses and the transcription of ISGs with antiviral properties in HESN 23 . Because in vitro exposure of PBMCs to thiazolides stimulates the activation of an immune profile that resembles that observed in HESNs, and NTZ was shown to have an antiviral activity that synergizes with that of other drugs, we verified whether in vitro susceptibility to HIV infection could be modulated by TIZ or RM-4848.

Evaluation of TZDs cytotoxicity in HIV-1-infected cells. Cell viability was analyzed in HIV-1-infected
cells cultured in the presence of different doses of TIZ or RM-4848 (1, 5, 10, 20 and 100 μ g/ml) by MTT assay. Results indicated that, with the exception of the highest dose, neither TIZ nor RM-4848 alter the viability of PBMCs (Fig. 1a,b respectively). To perform all the experiments a dose of 10 μ g/mL of compounds was chosen based on preliminary results showing that concentrations lower than 20 μ g/mL are sufficient to maintain pharmacological properties without affecting cell viability [Trabattoni D., Gnudi F. Personal Communication].

Antiviral activity of TZDs against HIV-1 BaL infection in vitro.
In vitro infection of PBMCs with HIV-1 was significantly and consistently reduced by TIZ and RM-4848; this effect was reproduced in all the experiments (n = 20) in which cells from randomly selected donors not carrying genetic markers known to reduce susceptibility to HIV infection were used (see Methods). The overall inhibition of HIV-1 replication in vitro was 87.8% for TIZ and 96.5% for RM-4848 (p-value < 0.01 vs. medium alone) at a 10μ g/mL concentration (Fig. 2a,b).
The antiviral effect of RM-4848 was further examined by adding this compound to cell cultures at different time-points: 1) for two days before HIV-1 infection and then washed away (pre-infection); 2) during the 24 hours infection period infection followed by washout; or 3) 24 hours after infection and untill the end of the cell culture period. Cells infected with HIV-1 in the presence of RM-4848 throughout infection (RM-4848) (10 μ g/mL) or in medium alone (no treatment) were considered as controls. The strongest inhibition of viral replication was observed when RM-4848 was added to the cultures at the time of infection, or one day after infection (Fig. 2c).
Analysis of anti-HIV-1 human response pathway. Results above show that thiazolides induce the up-regulation of genes such as IRF1, OAS2 and PRKRA known to obstacle HIV infectivity. To analyze more in depth whether the anti-HIV effects of TZDs stem from the modulation of cellular factors involved in the HIV-1 life cycle, we used an anti-HIV-1 human response pathway array that screens for the expression of such factors. Results indicated that the in vitro HIV-inhibitory properties of TZDs are only partially explained by the modulation of HIV-1 life cycle genes. Also in these experiments the antiviral activity of RM-4848 was more potent than that of TIZ (Fig. 4).
TZDs modulate cytokine and chemokine production in HIV-1 infected PBMCs. Cytokine and chemokine production by HIV-infected PBMCs in the presence of thiazolides was measured next. Results showed that IFNγ and IL-2 production was significantly increased in HIV-1 infected cells by both TIZ and RM-4848 (p < 0.05 vs. medium alone). Even more interestingly, the production of those CC chemokines known to be endowed with potent anti-HIV activities (MCP-1, MIP1-α , MIP1-β , RANTES) was modulated by both TIZ (Fig. 5a) and RM-4848 ( Fig. 5b) (p < 0.05 vs. medium alone). These results suggest that thiazolides obstacle viral entry and/or reduce the infection process of target cells.
TZDs up-regulate ISG involved in cholesterol metabolism. Since thiazolides up-regulate type 1 IFN and ISGs, and because ISGs involved in the metabolism and the efflux of cholesterol were recently shown to have a potent antiviral activity, we verified whether these compounds may also affect cholesterol metabolism, and in  (Fig. 6a) and RM-4848, with the effects of this latter compound once again being more potent (Fig. 6b).

Thiazolides upregulation of Interferon-Inducible Cholesterol-25-Hydroxylase (CH25H). Activation
of CH25H results in an enzymatic reaction that generates 25-HC, a natural oxysterol that mediates the efflux of cholesterol from cells and ligates the liver X receptor (LXR). To analyze possible correlations between genes involved in cholesterol efflux and in vitro inhibition of HIV-1 infection we stimulated HIV-1-infected cells with a synthetic agonist of LXR (TO-901317) and assessed viral infectivity. Results showed that agonist-mediated activation of LXR resulted in a significant decrease of cell infection (52%; p < 0.01) (Fig. 7), suggesting that the protective effects of thiazolides on HIV-1 infectivity are at least partially achieved through the modulation of cholesterol efflux mechanisms.

Discussion
The vast spectrum of antimicrobial activity of thiazolides was recently enriched by results indicating that these compounds are effective in reducing in vitro HIV infection 14,15 . We investigated the immunologic basis of this effect and show herein that incubation of PBMCs with thiazolides results in the secretion of type 1 cytokines and of CC chemokines, as well as the upregulation of type I IFN genes and interferon-stimulated genes (ISGs).    This effect was potentiated upon exposure of cells to HIV, and is associated with a significantly and reproducible reduction of HIV infectiousness of lymphocytes isolated from healthy individuals.
These results indicate that TZDs hamper HIV-1 infection of target cells and reduce viral replication using different and possibly overlapping modalities. Thus: 1) CC chemokines, whose production is stimulated by TZDs, impede the infection of target cells by CCR5 tropic viruses 24,25 , 2) Type 1 cytokines and cell-mediated immune responses contrast HIV-1 replication 26,27 , and 3) the stimulation of the transcription rate of genes endowed with generic antiviral effects, or know to have a specific ability to obstacle HIV-1 replication by TZDs, likely contributes to the anti-HIV properties of these compounds. Amongst the genes with generic antiviral effects that are upregulated by TZDs the activity of these compounds on: 1) PRKRA, a protein kinase activated by dsRNA that inhibits cellular mRNA translation and the synthesis of viral proteins 18,26 , 2) ADAR, an enzyme that destabilizes double-stranded RNA through conversion of adenosine to inosine 28 , 3) OAS2, which activates RNAses and induces the degradation of viral RNA and the inhibition of protein synthesis 29 , and 4) STAT2, a protein that activates the transcription of ISG and drives infected cell in an antiviral status 30 , is likely very important. Within genes that are upregulated by thiazolides and have a direct anti HIV-1 activity, IRF-1 a protein required for efficient HIV-1 replication 31 , and MX2 which acts as an inhibitor of HIV-1 infection by blocking viral replication at a late post-entry step, could play pivotal roles 32,33 . Notably, recent observations showed that IFN effectors work in cooperation to achieve a fully functional antiviral state, i.e., combination of different ISGs expressed together, such as what is observed in TZDs-treated cells, results in an ampler magnitude of antiviral activity than either gene alone 34 . These results would explain the greater efficacy of RM-4848 compared to TIZ in reducing HIV-1-infection, given the fact that RM-4848 results in the upregulation of a broader range of ISGs compared to what is seen in cells exposed to TIZ.
The observation that CH25H, an oxysterols-producing enzyme, is up-regulated by TIZs, is particularly intriguing. Viral infections themselves results in the rapid induction of CH25H and the consequent production of 25-HC via STAT1, a type I IFN effector. Interestingly, 25HC inhibits the growth of a wide range of enveloped viruses by inducing structural changes in the cellular membrane that impair viral entry at the virus-cell fusion step 20 . Our data show that thiazolides indeed up-regulate the expression of a number of genes that are part of cholesterol metabolism. Thus, the transcription factors of the sterol regulatory element-binding protein (SREBP) family, SREBP1 and SREBP2 are involved in the regulation of cholesterol biosynthesis 35 . Macrophage scavenger receptor-1 (MSR-1) 36 and the LDL receptor (LDL-R) 37 are responsible for the influx of lipids into macrophages. Finally, liver X receptor (LXR) α , and the peroxisome proliferated factor (PPAR) γ coordinate the removal of intracellular cholesterol 36,38 . Notably, TZDs also increased the expression of genes that encode lipid-transport proteins, such as the ATP-binding cassettes A1 (ABCA1) and G1 (ABCG1), by which intracellular excess of cholesterol is transported to extracellular acceptors 35,39 . The observations that: 1) thiazolides reduce in vitro HIV-1 infection; 2) TZDs up regulate LXR; and 3) in vitro HIV-1 infection of immune cells is significantly reduced by LXR agonists allow the speculation that the antiviral effects of these compounds are at least partially mediated by their ability to modulate proteins within the cholesterol metabolic pathway. Notably, this is a speculative hypothesis that will need to be tested in furhter experiments.
These results indicate that TZDs are endowed with strong and multifaceted antiviral activities, and confirm previous in vitro observations 14 suggesting that the use of these compounds might be envisioned in the therapy of HIV infection. In this light, preliminary results from a clinical study carried out in the United States in treatment-naïve HIV-HCV co-infected subjects are interesting. In this pilot study the activity on HIV viral loads (VL) of nitazoxanide (NTZ) alone and in combination with pegylated-interferon and ribavirin (PEG-IFN/ RBV) was evaluated. At baseline, HIV VL was available in 65 cases; after 16 weeks of NTZ in combination with PEG-IFN/RBV, HIV VL was undetectable (< 20 copies/ml) in 12 of the 14 (86%) patients in whom it could be reanalyzed. Notably, 5 of these 14 individuals were ARV-naïve throughout the study period; median HIV VL Scientific RepoRts | 6:27148 | DOI: 10.1038/srep27148 log 10 change in these patients was − 1.32; this effect was presumably mediated by NTZ (J. F. Rossignol, Personal Communication). These results should be interpreted with caution not only because of the small sample size and lack of a direct comparator, but also because the subjects had low viral loads at baseline, making the overall change in HIV VL difficult to quantify. These data nevertheless support the suggestion that NTZ may have a clinical benefit in patients infected with HIV.
Results herein show that the mechanism of activity of thiazolides is entirely new and is mediated by the activation of multiple and possibly synergistic mechanisms. Such mechanisms might confer to immune cells isolated from healthy individuals an "HESN-like" status, characterized by a drastic reduction of susceptibility to in vitro HIV infection. Additional disease models as well as clinical trials will be needed to elucidate the possible role of thiazolides in novel preventive and therapeutic approaches against HIV-1 infection and, possibly for the therapy of pathologies associated with alterations of the cholesterol metabolism.

Methods
Thiazolides. TIZ  All the procedures were carried out in accordance with the GLP guidelines adopted in our laboratory. PBMCs (2 × 10 6 cells/mL) were re-suspended in medium containing 1 ng/1 × 10 6 cells of HIV-1 BaL p24 viral input and incubated for 24 hours at 37 °C and 5% CO 2 . Cells were then washed and re-suspended in complete medium with interleukin-2 (IL-2, 15 ng/mL) (Sigma-Aldrich, St. Louise, MO, USA) and phytohemagglutinin (PHA, 7.5 μ g/mL) (Sigma-Aldrich) for 3 days. After viability assessment cells were washed and re-suspended in RPMI complete medium with IL-2 (15 ng/ml) and TIZ (10 μ g/mL), or RM-4848 (10 μ g/ml). 50% of the culture medium was renewed every 3 days. PBMCs were collected for gene expression at day 7 and 10 post-infection. In some experiments the conditions were modified as follows: 1) TZDs added to cell cultures before infection and then washed away Evaluation of the toxicity of TZDs. Cells were cultured in RPMI medium (Euroclone, Milan, Italy) supplemented with 20% FBS (Euroclone, Milan, Italy), penicillin and streptomycin, L-glutamine and gentamycin (Euroclone, Milan, Italy) alone or in presence of increasing concentration (0.5 μ g/ml, 1 μ g/ml, 10 μ g/ml, 20 μ g/ml and 100 μ g/ml) of drugs, diluted in medium culture for 1, 3, 5, 7 and 10 days. Every two days the 50% of medium was removed and replaced with fresh medium containing drugs at the corresponding concentrations. Toxicity of thiazolides was determined by a MTT based assay (Sigma Aldrich, St. Louis, MO, USA) according to manufacturer's protocol. PBMCs viability were established by dividing the absorbance reading of the formazan by the dry weight of cell cultures. Toxicity were determined by dividing viability of the drug-treated cells by viability of untreated cell control.
Cytokine and chemokine measurement. Real time PCR. cDNA quantification was performed by real-time PCR (DNA Engine Opticon 2; MJ Research, Ramsey, MN). Reactions were performed using a SYBR Green PCR mix (Promega, Fitchburg, WI, USA) and amplified according to the following thermal profile: initial denaturation (95 °C, 15 minutes) followed by 40 cycles of 15 seconds at 95 °C (denaturation) and 1 minute at 60 °C (annealing) and 20 seconds at 72 °C (extension). A Ct value of 40 or higher means no amplification and this value was not included in the calculations. Results were expressed as Δ Δ Ct and presented as ratio between the target gene and the GAPDH housekeeping mRNA. All the samples were analyzed in triplicate.
TLR, Type I IFN and HIV-1 human response pathway analysis. TLR, type I IFN and HIV-1 human response pathway were analysed in a PCR array including a set of optimized real-time PCR primer assays (SABiosciences Corporation, Frederick, MD, USA). This approach permits monitoring of mRNA expression of 84 genes related to the different pathways, plus five housekeeping genes. Results were analysed using the SABiosciences online software. Only targets showing > 2-fold modulation were considered significant. Experiments were run on the subjects included in the study pooled into distinct groups (medium alone, TIZ or RM-4848) and represent the mean value of the different targets analyzed in each group.

Statistical analysis.
Comparisons between groups were analyzed to evaluate immunological differences.
Kruskal & Wallis analysis of variance was performed for each variable; Bonferroni correction was applied to the results. Two-sided p-values were considered. Data analysis was performed using the SPSS statistical package (SPSS Inc. Chicago, IL, USA).