|Home | About | Journals | Submit | Contact Us | Français|
Despite current prophylactic strategies, sexually transmitted infections (STIs) remain significant contributors to global health challenges, spurring the development of new multipurpose delivery technologies to protect individuals from and treat virus infections. However, there are few methods currently available to prevent and no method to date that cures human immunodeficiency virus (HIV) infection or combinations of STIs. While current oral and topical preexposure prophylaxes have protected against HIV infection, they have primarily relied on antiretrovirals (ARVs) to inhibit infection. Yet continued challenges with ARVs include user adherence to daily treatment regimens and the potential toxicity and antiviral resistance associated with chronic use. The integration of new biological agents may avert some of these adverse effects while also providing new mechanisms to prevent infection. Of the biologic-based antivirals, griffithsin (GRFT) has demonstrated potent inhibition of HIV-1 (and a multitude of other viruses) by adhering to and inactivating HIV-1 immediately upon contact. In parallel with the development of GRFT, electrospun fibers (EFs) have emerged as a promising platform for the delivery of agents active against HIV infection. In the study described here, our goal was to extend the mechanistic diversity of active agents and electrospun fibers by incorporating the biologic GRFT on the EF surface rather than within the EFs to inactivate HIV prior to cellular entry. We fabricated and characterized GRFT-modified EFs (GRFT-EFs) with different surface modification densities of GRFT and demonstrated their safety and efficacy against HIV-1 infection in vitro. We believe that EFs are a unique platform that may be enhanced by incorporation of additional antiviral agents to prevent STIs via multiple mechanisms.
Newly acquired sexually transmitted infections (STIs) affect 340 million people per year (1,–3) and exert a significant impact on global health. Human immunodeficiency virus type 1 (HIV-1) affects ~35 million people globally (2,–5), while untreated STIs, such as those caused by herpes simplex virus 2 (HSV-2), can enhance both the acquisition and the transmission of HIV and other agents of STIs by 2- to 4-fold (6, 7). In light of the findings of recent clinical trials, a specific, multipurpose prevention technology that has the ability to prevent multiple STIs using one delivery platform and that also increases user adherence urgently needs to be developed (8,–18). In the study described here, we sought to shift current topical preexposure prophylaxis (PrEP) paradigms by integrating a multipurpose biological delivery approach to debilitate and inactivate HIV.
Our long-term goal is to develop a multipurpose biologically inspired electrospun fiber (EF) prevention technology that takes cues from the innate microenvironment of the female reproductive tract to more strategically narrow the gaps in microbicide efficacy. In this work, we evaluated the potential of polymeric EF scaffolds surface modified with the potent and broad-spectrum antiviral protein griffithsin (GRFT) to protect against HIV-1 infection in vitro.
We initiated our design by contemplating the role of the female reproductive microenvironment, namely, the innate properties and roles of cervicovaginal mucus and cell surface heparan sulfate proteoglycans, in preventing and enabling virus infection, respectively. The female reproductive tract is covered by mucus, a glycoprotein gel comprised of 95% water, 1 to 2% mucin fibers, and trace constituents, including salts, DNA, lactic acid, enzymes, and other proteins (19). To establish infection, virus must travel through the mucous barrier to reach epithelial and/or subepithelial cells. With pore sizes larger than a virus (for example, HSV is ~170 nm and HIV is ~100 nm), cervicovaginal mucus has been studied for its ability to hinder virus transport (20, 21). In fact, cervicovaginal mucus is believed to provide innate protection against infection primarily through mucoadhesive mucin-glycoprotein interactions by physically decreasing the flux of pathogens to the underlying epithelium (20,–22). Similarly, cell surface heparan sulfate proteoglycan adheres to viral glycoproteins and is known to provide an initial contact point for herpes simplex virions to adhere to and surf down cell filopodia, resulting in subsequent virus fusion or endocytosis (23). These examples highlight how the natural high-affinity interactions of mucins and heparan sulfate proteoglycan with viral glycoproteins decrease pathogen flux in cervicovaginal mucus while, conversely, providing a foundation for virus attachment and entry into host cells. These adhesive interactions prompted us to consider how we might incorporate similar physicochemical attributes into the design of our EFs while still enabling the future capability of the prolonged delivery of active agents to host cells.
Just as cervicovaginal mucus poses an impediment to virus transport, often to the detriment of drug delivery, it can adversely affect the transport and, thereby, efficacy of locally delivered agents and delivery vehicles (20, 22, 24,–28). To offer physicochemical protection in a unique delivery vehicle, we envisioned and designed a fabric scaffold, comprised of polymeric EFs, that integrates specific adhesive and antiviral properties to provide virus entry inhibition (Fig. 1 and and2).2). Whereas EFs have traditionally been used to provide sustained delivery (29, 30) or scaffolds for cell growth (31, 32), they may also incorporate surface modifications to interact with cells, viruses, or their surrounding microenvironment (33). Furthermore, despite being a well-established technology that has successfully been used in numerous applications (34), EFs have only recently been explored for their ability to deliver antiretrovirals (ARVs) to the female reproductive tract (35,–40; S. E. Aniagyei, L. B. Sims, D. A. Malik, W. Kim, and J. M. Steinbach, submitted for publication). In contrast to topical prevention strategies that have incorporated ARVs within fibers, to our knowledge this is the first work to surface modify EFs with an antiviral protein to inhibit HIV infection. We believe that EFs have the potential to offer an effective new way to present and/or deliver agents to the female reproductive tract by acting as a durable, surface-functionalized delivery platform.
To create these antiviral protein-modified EFs, we selected the FDA-approved polymer poly(lactic-co-glycolic acid) (PLGA), known to impart biocompatibility, various rates of biodegradability, and sustained drug delivery, as a scaffold for modification (41,–43). For the adhesive surface moiety, the biological antiviral lectin GRFT was selected, as it has been demonstrated to bind to and to have antiviral efficacy against a variety of viruses, including HIV-1 and HSV-2 (44,–51). Previous work by our groups has demonstrated that GRFT has an exceptionally potent HIV-1-inhibitory capability as a multivalent front-line HIV entry inhibitor (49,–54), and of the biologically based inhibitors, GRFT has exceptionally potent anti-HIV activity, inactivating HIV-1 immediately upon contact (51, 52). Furthermore, a 0.1% GRFT gel was shown to protect mice against intravaginal HSV-2 challenge, making it a leading protein-based antiviral candidate providing multipurpose protection against both HSV-2 and HIV-1 STIs (46). Moreover, to the benefit of intravaginal administration, GRFT lacks the capacity to induce proinflammatory cytokines in human peripheral blood mononuclear cells, is resistant to a broad array of human proteases, and is stable in the presence of culture medium from vaginal microbes, retaining anti-HIV activity for up to 10 days, even when some of the six carbohydrate binding sites are occupied (55, 56). Mechanistically, GRFT has been shown to bind the terminal mannose N-linked glycan residues on viral envelope surfaces (51, 55, 57,–59). For HIV-1, GRFT is believed to bind to gp120 on the virion and neutralize the virus at entry, whereas for HSV-2, GRFT inhibits infection via the glycoprotein (gP) conformational changes expressed during cell-to-cell spread (46). For both HIV and HSV (among other viruses), these strong interactions between virus gPs and GRFT result in an irreversible biological effect (56) that we hypothesize inactivates the virus, resulting in physicochemical protection. Last, due to GRFT's synergistic activity with ARV combinations (53), it will be feasible to provide a platform for coadministration with encapsulated active agents in future work.
While the above-noted success of GRFT has resulted in progression to clinical trials, to date, there are no delivery platforms that integrate GRFT delivery. From a materials perspective, our long-term goal is to provide an innovative alternative delivery method to prolong the delivery of coadministered biological and chemical agents by targeting multiple stages of virus infection. In the present work, we sought to develop and characterize GRFT-modified EFs (GRFT-EFs) as a protein-based viral adhesive and inactivating platform to protect against HIV-1 infection in vitro. We hypothesize that GRFT-EFs provide a physicochemical barrier to virus entry prior to targeting of cell infection, thereby reducing and simultaneously inactivating the viral load that elicits subsequent HIV infection. In the long term, we believe that EFs may circumvent virus at its entry point and provide protection against infection in the female reproductive tract.
Carboxyl-terminated 50:50 poly(lactic-co-glycolic acid) (PLGA; 0.55 to 0.75 dl/g; molecular weight, 31,000 to 57,000) was purchased from Lactel Absorbable Polymers (Cupertino, CA). Dichloromethane (DCM), dimethyl sulfoxide (DMSO), and a 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay kit were obtained from Sigma-Aldrich (St. Louis, MO). 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP) was obtained from Fisher Scientific (Pittsburgh, PA). GRFT (molecular mass, 12.7 kDa) was produced in Nicotiana benthamiana as previously reported (60). Fetal bovine serum (FBS) and antibiotics (penicillin-streptomycin) were purchased from VWR. Dulbecco's modified Eagle's medium (DMEM) and keratinocyte serum-free medium were purchased from Invitrogen. Enzyme-linked immunosorbent assay (ELISA) kits and 1-ethyl-dimethyl-aminopropylcarbodiimide (EDC)–N-hydroxysuccinimide (NHS) were purchased from Thermo Fisher. Simulated vaginal fluid (SVF) was prepared as described in reference 61. The pseudovirus was produced in 293T/17 cells as previously described, using an envelope (Env)-expressing plasmid (CCR5-tropic clade A strain Q769.h5) and an Env-deficient HIV-1 backbone vector (pNL4.3ΔEnv-Luc) (54, 62, 63). The Env-expressing plasmid was obtained from the NIH AIDS Research and Reference Reagent Program (catalogue no. 11884).
PLGA EFs were prepared and electrospun in HFIP. Their compositions consisted of from 8 to 30% (wt/wt) drug to polymer to establish a baseline blank (unmodified) fiber formulation with a well-delineated morphology. Solutions of 10 to 30% (wt/wt) PLGA were prepared in HFIP and allowed to solubilize overnight on a shaker at room temperature. Three milliliters of each polymer solution was aspirated into and spun from a 3-ml plastic syringe on a custom-built device housed in an air-filtered Plexiglas chamber (Fig. 2). Flow rates spanning 0.5 to 3.0 ml/h were optimized over a range of voltages (15 to 27 kV), and the resulting fiber mat was collected on a rotating 25-mm-outer-diameter stainless steel mandrel located 25 cm from the blunt needle tip. The sample flow rate was monitored by an infusion pump (Fisher Scientific, Pittsburgh, PA), and the voltage was applied using a high-voltage power supply (Spellman CZE 1000R). Final electrospinning conditions applied a voltage of 27 kV and electrospun 15% (wt/wt) PLGA in HFIP with a flow rate of 2 ml/h. After electrospinning, the fibers were removed from the mandrel and dried overnight in a desiccator cabinet.
PLGA EFs were surface modified with 5, 0.5, 0.05, 0.005, 0.0005, and 0.00005 nmol GRFT per mg EF, to yield six different surface densities of the lectin GRFT on fibers. PLGA fibers were modified using EDC-NHS chemistry per the manufacturer's instructions (Thermo Fisher) (Fig. 2). Briefly, holes were punched into PLGA EF fiber sheets to create circles (surface area, 28.27 mm2), each of which weighed ~2 mg, and the circles were immersed in 6 ml 2-(N-morpholino)ethanesulfonic acid buffer (MES; pH 5 to 6) in 15-ml conical polypropylene tubes (~100 fiber circles per tube). The total mass of the fibers was recorded. Two milliliters EDC (2 mg/ml MES) and 2 ml NHS (3 mg/ml MES) were then added to the EFs. EF samples were set on a rotator at ~40 rpm for 15 min to activate and were subsequently quenched with 14 μl 2-mercaptoethanol. The supernatant was discarded, and the fibers were rinsed with 1× phosphate-buffered saline (PBS; pH 7.2 to 7.5) to remove unreacted substrates. Stock GRFT solution (10 mg/ml) was then diluted in 1× PBS, based on the total mass of fiber and the desired GRFT surface density for each, to provide six different reaction vials, one for each of the six desired surface-modified EF formulations. EDC-activated EFs and GRFT were subsequently reacted on a rotator at ~40 rpm at room temperature overnight. On the following day, 2 ml of 3.5 mg/ml hydroxylamine in PBS was added to terminate each reaction, and the fibers were rinsed twice in distilled water and placed in petri dishes. The resulting surface-modified EFs were dried in a desiccator for 2 days and stored at 4°C until use.
GRFT-EF morphologies and sizes were evaluated using scanning electron microscopy (SEM). After they were dried in a desiccator, the EFs were placed on carbon tape, sputter coated with gold, and imaged using SEM (Supra 35; Zeiss, Oberkochen, Germany). All SEM images were acquired at the appropriate magnifications to enable clear visualization of the fiber microstructure. Fiber diameters were obtained by analyzing SEM images in NIH ImageJ software. To determine the average diameter of each formulation, line elements were drawn across a minimum of 50 fibers per image. All experiments were conducted with a minimum sample size of 3. Unless otherwise noted, the error bars in all figures represent the standard deviations of the measurement means. Statistical significance was determined using a two-sided Student's t test (P < 0.05).
To determine the density of GRFT on the EF surfaces, triplicate samples of ~2 mg of EFs for each theoretical formulation (5, 0.5, 0.05, 0.005, 0.0005, and 0.00005 nmol GRFT per mg EF) were weighed and dissolved in 1 ml DMSO. The EF solutions were vortexed for 1 min and subsequently diluted 100-fold in Tris-EDTA (TE) buffer for the quantification of GRFT via a gp120-based ELISA.
To determine the in vitro release of conjugated GRFT from the EF surface, triplicate fiber pieces of ~2 mg were cut and immersed in 1 ml of SVF. Samples were incubated at 37°C and constantly shaken. At fixed time points (1, 2, 4, 8, 24, 48, and 72 h and 1, 2, 3, and 4 weeks), 1 ml of SVF from each sample was collected and replaced with another 1 ml of fresh SVF. The quantity of GRFT released from the fiber surface in the SVF was quantified by ELISA.
GRFT concentrations for surface conjugation quantification and surface release experiments were determined using a previously established ELISA method (51, 56, 64). Briefly, MaxiSorp plates (Nunc) were coated with influenza virus hemagglutinin (HA; Kentucky Bioprocessing [KBP]) at 10 μg/ml in 1× PBS and incubated overnight at 4°C. The plates were blocked with 3% (wt/vol) bovine serum albumin (BSA) in PBS containing 0.05% Tween 20 (PBS-T). Samples were diluted 1:2 in PBS and were incubated at room temperature for 1 h. Serial dilutions of purified GRFT were run in parallel to generate a standard curve. The HA-bound GRFT was detected by goat anti-GRFT antiserum (1:10,000) followed by horseradish peroxidase (HRP)-conjugated rabbit anti-goat IgG (1:10,000). The plates were developed with SureBlue TMB Microwell peroxidase substrate, and the reactions were stopped with 1 N H2SO4. The absorbance at 450 nm was measured using a BioTek Synergy HT plate reader and correlated to that of the free GRFT standard. Data are shown as the means ± standard deviations.
TZM-bl cells were obtained from the NIH AIDS Research and Reference Reagent Program (ARRRP). These cells are genetically engineered HeLa cell clones that express CD4, CXCR4, and CCR5 and contain Tat-responsive reporter genes for firefly luciferase (Luc) and Escherichia coli β-galactosidase under the regulatory control of an HIV-1 long terminal repeat (65, 66). TZM-bl cells were maintained in Gibco DMEM containing 10% heat-inactivated FBS, 25 nM HEPES, and 50 μg gentamicin per ml in a vented T-75 culture flask. Vaginal keratinocyte (VK2/E6E7 [VK2]), ectocervical (Ect1/E6E7 [Ect1]), and endocervical (End1/E6E7 [End1]) cell lines, obtained from ATCC, were evaluated in cytotoxicity experiments. These cell lines were selected because they are representative of the cell types in the female reproductive tract that would be exposed to topically applied EFs. VK2, Ect1, and End1 cells were maintained in keratinocyte serum-free medium (Invitrogen) supplemented with bovine pituitary extract (50 μg/ml), epidermal growth factor (0.1 ng/ml), penicillin (100 U/ml), and streptomycin (100 μg/ml). The medium was further supplemented with calcium chloride (CaCl2) to a final concentration of 0.4 mM. When cells were trypsinized for plating and counting of single cells, cells were neutralized using DMEM–F-12 medium with 10% FBS and 1% penicillin-streptomycin.
VK2, Ect1, and End1 cells were plated at a density of 600,000 cells per well in 12-well plates a day before treatment, to yield 50% confluence on the day of treatment. At 24 h after seeding, 2 mg unmodified EFs or GRFT-EFs was administered to VK2, Ect1, and End1 cells for 24, 48, and 72 h. Unmodified EFs and the top three most active concentrations of GRFT-modified EFs (0.05, 0.5, and 5 nmol/mg) were placed in transwell inserts in a 12-well plate with cells to assess cytotoxicity. At the corresponding time points after treatment, cytotoxicity was measured using an MTT assay. On the following day, the cells were lysed and the absorbance at 570 nm was read. All experiments were conducted with a minimum sample size of 3. Data are shown as the means ± standard deviations.
Reporter gene env-pseudotyped virus infectivity assays were used to measure the antiviral activity of GRFT-EFs. The anti-HIV activity of the GRFT-EFs was measured as a reduction in firefly luciferase (luc) reporter gene expression in the presence of GRFT-EFs compared to that of unmodified EFs after a single round of infection in TZM-bl cells, as we have done previously with free GRFT (54). Briefly, the optimal viral input dilution of each pseudovirus (CCR5-using clade A strain Q769.h5) was determined as follows. The appropriate virus concentration was established such that infection in untreated, infected cells yielded ≥150,000 relative light units (RLU) and greater than 10 times the average number of RLU of cell-only background levels, as recommended in a standardized protocol (62, 67). This virus concentration (in our experiments, ~880,000 RLU, relative to the background of ~1,900 RLU in untreated and uninfected cells) was established by diluting 1 ml of virus stock 8-fold. This diluted virus aliquot was then incubated with 2 mg of the six different GRFT-EFs in triplicate at 37°C in 96-well black solid flat-bottom culture plates. After 1.5 h, the solution containing any unbound virus (50 μl) was added to freshly trypsinized cells (10,000 cells in 100 μl of growth medium containing 10 μg/ml DEAE-dextran). The remaining volume was brought up to a final volume of 200 μl with fresh growth medium. For cells with no virus added, 100 μl of fresh growth medium was added to each well. Additionally, one set of eight control wells received cells plus virus (untreated, virus-infected control), another set of eight wells included cells plus virus incubated with blank fibers (fiber control), another set of eight wells included cells plus free GRFT, and the last set of eight wells received cells only (background control). After 48 h of incubation, 100 μl of culture medium was removed from each well and 100 μl of Bright-Glo reagent (Promega Corp.) was added to the cells. After 2 min of incubation at room temperature, luminescence was measured using the Synergy HT luminometer. The 50% inhibitory concentration (IC50) was defined as the GRFT surface modification density that caused a 50% reduction in the number of RLU compared to that for the untreated, virus-infected control wells after subtraction of the background number of RLU from the untreated, uninfected cells. Data were plotted and analyzed using one-way analysis of variance (ANOVA) with Bonferroni posttests (P < 0.05), and the IC50 was determined using the dose-response inhibition variable slope fit in GraphPad Prism (version 6.0) software. Experiments were conducted at least in triplicate, and the data are shown as means ± standard deviations.
Transwell membranes within 24-well plates were removed and replaced with unmodified PLGA EFs to assess the physical potential of EFs to prevent HIV penetration. The maximum pseudovirus titer used in the antiviral efficacy experiments was administered to the apical side of each transwell sample (with the membrane removed), which contained either no fiber or PLGA fiber. Fibers were in contact with the underlying medium, which contained TZM-bl cells, for up to 3 days. The amount of HIV infectivity relative to that with no treatment (no fiber in an empty transwell membrane) after 3 days is shown.
Two milligrams of the GRFT-EFs with 0.05, 0.5, or 5 nmol GRFT per mg was incubated with 10 dilutions (series dilutions, 3-fold stepwise) of 1 ml pseudovirus stock, beginning at the optimal viral input dilution (here, ~514,000 RLU, relative to the background in untreated and uninfected cells of ~1,800 RLU) described above. Fibers were incubated in triplicate for 1.5 h at 37°C in a total volume of 100 μl growth medium in 96-well black solid flat-bottom culture plates. The antiviral activity of GRFT-EFs was assessed using TZM-bl cells as described above. Data were plotted and analyzed in GraphPad Prism (version 6.0) software, using one-way analysis of variance with Bonferroni posttests (P < 0.05). Data are shown as the means ± standard deviations.
A schematic of the electrospinning process is shown in Fig. 2A. Using carbodiimide cross-linker chemistry, six different amounts of GRFT were conjugated to unmodified EFs (Fig. 2B and andC).C). The morphologies of EFs with different surface modifications are shown in Fig. 3. The average EF diameters were measured using ImageJ software and are shown in Fig. 3E. Relative to unmodified EFs (Fig. 3A), the modified EFs with the three highest GRFT modifications, 0.05, 0.5, and 5 nmol GRFT (Fig. 3B to toD),D), displayed no distinct morphological differences. No statistically significant differences in diameters were found between GRFT-modified and unmodified EFs, indicating that the GRFT modification does not impact the EF diameter.
To determine the density of GRFT on the EF surface, we extracted GRFT from each of the GRFT-EFs and measured the amount extracted using an ELISA (Fig. 4A). We found that the amount of extracted GRFT positively correlated with the theoretical modification density; however, at high conjugation concentrations of GRFT, we achieved very low conjugation efficiencies (Table 1). For EF modifications with 5, 0.5, 0.05, and 0.005 nmol GRFT/mg, we achieved conjugation efficiencies spanning 0.6, 4.2, 6.9, and 43.2%, corresponding to 373, 165, 42, and 40 ng GRFT per mg EF, respectively. For the modifications with the two lowest concentrations of (0.0005 and 0.00005 nmol/mg EF), the amount of GRFT was within the threshold of our detectable limit using ELISA. Figure S3 in the supplemental material depicts the results for a typical GRFT standard relative to that for GRFT extracted from a known spiked fiber sample to highlight that active GRFT was extracted from the fibers.
We next wanted to determine the amount of GRFT released from the GRFT-EFs as a function of time. After incubating the fibers in SVF for 1 h, we observed that 102, 23, and 8 ng GRFT per mg EF was released for the EFs containing 5, 0.5, and 0.05 nmol GRFT, respectively. After 4 h, we observed that a total of 30%, 41%, and 24% (113, 25, and 10 ng/mg, respectively) of the GRFT detected from the loading experiments (conjugated or adsorbed to the fibers) was released into the SVF from the EFs with 5, 0.5, and 0.05 nmol GRFT/mg, respectively (Fig. 4B). Negligible GRFT (≤1 ng GRFT released per mg of fiber) was detected in SVF for all formulations after 4 h and for GRFT modification concentrations of less than 0.05 nmol/mg. This demonstrates that the majority of the GRFT is conjugated to and retained on the fibers, whereas potentially adsorbed GRFT is released during the first 4 h. Furthermore, after 8, 24, 48, and 72 h, less than 0.5 ng per mg fiber was released. For these reasons, we terminated the experiment after 72 h. The amount released from the 1-, 2-, and 4-h release durations was summed to attain the cumulative releases, shown in Table 1, representing the total amount of GRFT released from the fiber after 4 h.
To test the antiviral activity of GRFT-EFs, we assessed the HIV-1 neutralization activity of all six surface-conjugated GRFT-EF formulations. As shown in Fig. 5, we observed a dose-dependent effect on HIV inhibition with respect to the amount of GRFT surface modification. Fibers with higher theoretical surface densities of 5 and 0.5 nmol per mg EF almost completely neutralized HIV-1 infection (100 and 97%, respectively), whereas fibers with 0.05 and 0.005 nmol/mg theoretical GRFT neutralized virus infection by 67 and 23%, respectively. Unmodified EFs inhibited HIV by ~38%, whereas the two lowest concentrations of fibers demonstrated negligible inhibition. The HIV-inhibitory activities of the top four most active GRFT-EF surface modification densities and the unmodified EFs were statistically significant (showed lower infection) compared to the results for the untreated group. In addition, the percent virus inhibition by GRFT-EFs with 5, 0.5, and 0.05 nmol GRFT was statistically significant relative to that by unmodified EFs. This statistical significance indicates that while unmodified EFs provide partial protection against HIV infection, the modification of EFs with GRFT at high surface densities significantly improves their efficacy. Using these data, we determined the IC50 of the GRFT-EFs to be 41.5 ng GRFT/mg EF, or 415 ng GRFT/ml. Similarly, we evaluated the efficacy of free GRFT as a baseline for comparison (see Fig. S1 in the supplemental material) and achieved an IC50 of 233 ± 2.7 ng GRFT/ml under similar administration conditions.
To assess the antiviral activity of these GRFT-EFs when exposed to different concentrations of HIV in vitro, we evaluated their antiviral effects across 10 different viral inocula. As shown in Fig. 6, the two highest viral inocula were the most viable and provided the most insight into the ability of our EFs to inhibit infection relative to that of no treatment. EFs with a theoretical GRFT surface density of 5 nmol per mg inhibited HIV infection by between 89 and 99%, depending on the virus concentration used. In comparison, GRFT-EFs with 0.5 and 0.05 nmol GRFT inhibited virus infection by from 64 to 99% and from 36 to 99%, respectively. For the three highest viral inocula, corresponding to those administered in the experiments whose results are shown in Fig. 5, the results for all three theoretical GRFT surface densities except 0.05 nmol/mg were statistically significant (lower levels of infection were detected) relative to the results for the untreated group. For the fourth-highest viral inoculum, only 5 and 0.5 nmol GRFT demonstrated statistically significant decreases relative to the results for untreated cells. For lower viral inocula, statistical significance was no longer observed due to the relatively low HIV infectivity observed in the untreated group, making the neutralization effect of GRFT-EFs appear to be negligible.
Between treatment groups, for the two highest viral inocula, GRFT-EFs with 5 nmol GRFT demonstrated statistically significant viral inhibition compared to the results obtained with fibers containing 0.5 or 0.05 nmol GRFT. However, for the third- and fourth-highest viral inocula, statistical significance was observed only between fibers containing 5 and 0.05 nmol and fibers containing 0.5 and 0.05 nmol. This suggests that at lower virus concentrations, both EFs containing 5 nmol GRFT and EFs containing 0.5 nmol inhibit HIV with a greater efficacy than EFs containing 0.05 nmol, and there is not a significant difference between the inhibitory potential of EFs containing 5 and 0.5 nmol. Taken together, these results indicate that GRFT-EFs are effective in inhibiting HIV-1 in a dose-dependent manner across a range of virus concentrations in vitro.
Last, as we observed that blank fibers alone partially inhibited infection, we wanted to assess the ability of blank EFs to physically prevent HIV infection. To evaluate this, we replaced transwell membranes with unmodified EFs and administered virus to the apical surfaces of these EFs, which had their basal sides in contact with the underlying medium containing TZM-bl cells. We discovered that unmodified PLGA EFs completely inhibited HIV penetration for up to 3 days under these optimal administration conditions in vitro (see Fig. S2 in the supplemental material).
To determine the biocompatibility of GRFT-EFs and to conclude that the antiviral effects of GRFT-EFs were not attributed to fiber-induced cytotoxicity, we used vaginal epithelial cell lines, VK2, Ect1, and End1, to assess the cytotoxicity of unmodified and GRFT-EFs in vitro. As shown in Fig. 7, cells of all cell lines treated with GRFT-EFs demonstrated greater than 94, 95, and 93% viability on days 1, 2, and 3, respectively, relative to the viability of negative and positive control cells receiving no treatment and treatment with 10% DMSO, respectively. No statistically significant changes in cell viability across cell lines and the duration of incubation were seen.
A significant obstacle to more efficacious microbicide delivery is the paucity of suitable biologic and drug delivery vehicles targeted to the unique microenvironment of the female reproductive tract. For intravaginal delivery, cervicovaginal mucus has been studied for its ability to hinder virus transport, as it provides innate protection against infection primarily by mucoadhesive interactions and not by steric interactions (20,–22). This phenomenon inspired us to consider the natural host defense process in the design of our material and suggested that a delivery platform that incorporates similar virus-adhesive characteristics yet that has specificity for viral glycoproteins may prevent infection by physically decreasing the flux of pathogens to the epithelium. However, just as cervicovaginal mucus poses an impediment to incoming pathogens, it can adversely affect the efficacy and transport of drug delivery. Therefore, delivery vehicle design must consider mucosal properties such that the carrier can traverse or remain stable in mucus while it ideally inhibits virus penetration and entry and simultaneously delivers the agent to target sites (virus or host cells).
Currently, most delivery vehicles provide active agent delivery—typically, ARVs—to the female reproductive tract in the form of a gel, film, or intravaginal ring (IVR) (13, 14, 17, 68,–74). However, each delivery platform has its respective hurdles to achieving optimal delivery for long-term efficacy. Regarding antiviral gels, adequate protection is provided when they are frequently applied, and they often require strict user adherence to maintain effectiveness (75, 76). However, user acceptance and adherence can be negatively impacted by the leakiness and corresponding untidiness experienced by some users. Similarly, for intravaginal films, user adherence is a common barrier to obtaining efficacy and is impacted by the perceived difficulty of film administration, as well as localized irritation after prolonged contact (76, 77). Furthermore, the rapid and complete release of the incorporated agents from intravaginal films within hours of administration may result in transient protection against STIs, thereby remaining a major hurdle to obtaining long-term delivery and efficacy (78, 79).
In comparison to gels and intravaginal films, IVRs currently provide the “gold standard” for sustained release and have been demonstrated to deliver contraceptives and provide long-term protection against STIs for over 90 days (80,–82). The ASPIRE and IPM trials recently demonstrated the feasibility of dapivirine IVRs as PrEP against HIV, resulting in a decrease in the rate of HIV infection by 27% and 31%, respectively, for up to 1 month (83). However, user adherence to the administration of IVRs remains a concern, and it is believed to contribute to the modest efficacy observed in these studies. In the ASPIRE trial, women most vulnerable to infection (those ages 18 to 21 years) were less likely to adhere to IVR administration (83), suggesting that alternative dosage forms may better address women's preferences and needs for protection. Furthermore, the ability of IVRs to incorporate more sensitive agents, such as biologics, may be a concern due to the high processing temperatures utilized during the IVR manufacturing process. Together, these challenges emphasize the need to develop alternative delivery vehicles to prevent STIs (84).
Relative to these existing technologies, EFs have more recently emerged as platforms for the delivery of antivirals with activity against HIV and other viruses (33, 35,–37, 39, 40, 85, 86). One of the primary goals in the development of EFs has been to encapsulate active agents to prolong their release and activity relative to those for other delivery platforms. Correspondingly, the advantages of EFs include high loading efficiency, enhanced agent stability relative to that after the administration of free agent, and tunable sustained release dependent on polymer choice. Furthermore, due to the variety of polymers available, virtually any compound can be incorporated, and delivery durations can be altered to suit the specific delivery requirements of a topical delivery vehicle (87, 88). Additionally, EFs have the potential to simultaneously deliver incorporated compounds, including biologics, such as proteins and oligonucleotides (89,–91), as well as traditional antiviral drugs (92,–94), making polymeric EFs an attractive platform for the delivery of multipurpose drugs with activity against STIs. Although EFs are still being explored to establish their safety and efficacy in vitro and in vivo, current in vitro studies indicate strong safety and efficacy profiles (35, 36, 40, 85, 86).
While we and others are developing EFs to encapsulate active agents within the delivery vehicle, as described above and in other work (35,–37, 39, 40, 85, 86), one overlooked advantage of EF scaffolds is that their surfaces may also be utilized to provide antiviral protection and deliver active agents. Furthermore, new biologic agents, such as the antiviral protein GRFT, hold promise as alternatives to traditionally administered small-molecule ARVs. Here, on the basis of our work and the work of others demonstrating the in vitro and in vivo efficacy of free GRFT against a diversity of infectious agents, including HIV and HSV-2, we wondered how GRFT might perform as a biomaterial surface modification. We hypothesized that PLGA EFs surface modified with the potent and broad-spectrum antiviral protein GRFT may protect against HIV infection in vitro by incorporating both a chemical and a physical barrier to adhere to and inactivate HIV. With these considerations, our goal in this work was to evaluate the potential of surface-modified PLGA EF scaffolds against HIV infection in vitro. Here our aims were to create and characterize GRFT-modified EFs, evaluate how surface modification impacts HIV infection in vitro, and assess the cytotoxicity of these EFs for vaginal and cervical epithelial cells. To our knowledge, there are no delivery platforms currently available that utilize polymer scaffolds to surface conjugate active biological agents, such as the lectin GRFT, to adhere to and inactivate HIV as a means of prevention.
The first factor that we evaluated in the design of these EFs was fiber morphology (Fig. 3). As we expected, the EF morphology remained unchanged after surface modification with a variety of GRFT concentrations. Fiber diameters averaging approximately 2 μm were obtained for all groups, with no statistically significant difference being observed between the GRFT-EF groups.
The next thing that we characterized was the efficiency of loading of GRFT on the EF surface. We observed an increase in the amount of conjugated GRFT with respect to the theoretical GRFT concentration added. However, the conjugation efficiency was rather low at a high level of theoretical GRFT modification (5 nmol per mg EF), resulting in an ~1% modification efficiency for EFs with 5 nmol GRFT and 4 to 7% modification efficiencies for EFs with 0.5 and 0.05 nmol GRFT (Fig. 4; Table 1). In contrast, when the concentration of GRFT was decreased to 0.005 nmol per mg EF, we obtained a 43% modification efficiency. At concentrations lower than 0.005 nmol per mg EF, the amount of GRFT on the fiber surface was undetectable via ELISA.
We believe that this range of conjugation efficiencies may be attributed to a number of factors. First, the GRFT used in this study has N-terminal acetylation, thus hindering the conjugation potential of the N-terminal amine group (50). Second, there are only two lysine residues, Lys6 and Lys99, in the primary sequence of GRFT (i.e., four per GRFT dimer molecule), of which the latter is buried near the interface between two monomers of the domain-swap dimer structure. Hence, there may be steric hindrance of two of the four lysine residues within the GRFT protein, resulting in the availability of an insufficient number of primary amine groups for conjugation. We suggest that this may limit efficient conjugation with the PLGA fiber carboxyl groups. To overcome these challenges, one option may be to modify GRFT with an amine at a different location that is more accessible for EF conjugation; however, studies would need to be conducted to ensure that the antiviral activity of GRFT is maintained. Conversely, the number or type of functional groups on PLGA EFs may be altered, or we may consider noncovalent attachment of GRFT to the EF surface to increase the conjugation efficiency. Both GRFT and polymer modification approaches may be pursued in future work to enhance the conjugation efficiency.
While our goal was to covalently modify EFs with GRFT, we acknowledge that the pI of GRFT (pI 5.4) may induce the surface adsorption of GRFT—rather than solely covalent binding—to the negatively charged and hydrophobic PLGA EF surface. For smaller proteins like GRFT (molecular mass of the dimer in solution, 25.4 kDa), physical adsorption can be a rapid and reversible process dependent on the protein structure (95). The net charge of a protein is determined by the pKa of its side chains, where side groups with a pKa above the pH of the physiological microenvironment have a positive charge and groups with a pKa below the pH of their environment have a negative charge (96). The net charge of the protein is a summation of these forces that contribute to protein-material, protein-cell, and protein-protein interactions in vivo. Therefore, to determine the presence of transiently surface-adsorbed GRFT and to characterize its potential release from the fiber surface, we performed a sustained-release assay over 4 weeks. Figure 4B shows that the release of what we believe to be adsorbed GRFT from the EF surface occurs within the first 4 h. During these first 4 h, approximately 113, 25, and 10 ng of GRFT was released from the EF surface, corresponding to 30, 41, and 24% release from the EFs with 5, 0.5, and 0.05 nmol GRFT per mg, respectively. After 4 h, negligible GRFT release from all formulations was detected. While the majority of GRFT was released after 1 h (102, 23, and 8 ng GRFT per mg EF for the EFs with 5, 0.5, and 0.05 nmol GRFT, respectively [Fig. 4B]), after 4 h we observed the release of less than 1 ng GRFT/mg EF from each formulation. Furthermore, after 8, 24, 48, and 72 h, less than 0.5 ng GRFT per mg of fiber was released at each time point. For these reasons, we terminated the release experiments after 72 h.
To explain these release kinetics, we considered the burst release often seen after the encapsulation of active agents within polymeric delivery vehicles. Typically, when considering the encapsulation of hydrophilic agents, it is well-known that agents near the fiber surface release within the first 24 h (34). In comparison to encapsulated agents, here we expected that surface-adsorbed GRFT would release even more rapidly, within the first few hours. Whereas reaction conditions may encourage the electrostatic binding of positively charged GRFT to negatively charged PLGA fibers at neutral pH (7.4) relative to the GRFT pI of 5.4, in the release assay, GRFT fibers were exposed to SVF (pH 4.5). When the environment is shifted to a more physiologically relevant environment with a pH below the pI of GRFT, we propose that GRFT becomes more negatively charged and is repelled from the negatively charged PLGA fiber surface, prompting rapid release.
With respect to the impact of this released (or surface-desorbed) GRFT on HIV infection, GRFT has been shown to immediately inactivate virus upon contact (50, 52). Furthermore, GRFT is known to maintain its antiviral activity even when some of its glycan-binding sites are bound (55, 56). In our studies, we preincubated HIV with GRFT fibers for 90 min and subsequently added the supernatant solution (containing unbound HIV and some fiber-released GRFT) to cells to determine the efficacy of HIV inhibition of our six surface-conjugated formulations in vitro (Fig. 5). In agreement with the loading data and regardless of the conjugation efficiency, we observed a concentration-dependent increase in inhibition corresponding to the GRFT surface modification density. Two milligrams of the GRFT-EFs with 5 and 0.5 nmol GRFT provided complete protection against HIV, whereas the same quantity of the EFs with 0.05 and 0.005 nmol GRFT per mg provided 67 and 23% inhibition, respectively. EFs with GRFT at concentrations lower than 0.005 nmol per mg had a negligible effect on HIV inhibition. Furthermore, despite released GRFT or unbound virus, when the virus solution was added to cells, full inhibition of HIV infectivity was observed in vitro for our top two most active EF formulations, indicating that unbound GRFT (and potentially unbound virus) does not detract from efficacy. From these data, we calculated that an IC50 of 41.5 ng GRFT per mg fiber, or 415 ng GRFT/ml, is required to inhibit HIV infection. Comparing this value to previous measurements in the literature of 157 ng/ml using the same virus strain (Q769.h5) (54) and with our own data (IC50 of free GRFT, ~233 ng/ml; see Data Set S1 in the supplemental material), we achieved a slight decrease in short-term efficacy utilizing surface-modified EFs relative to that achieved utilizing free GRFT.
To consider how the results of the controlled-release assay relate to observed efficacy, we note that the amount of GRFT released from the top three most active fiber modifications (5, 0.5, and 0.05 nmol/mg, corresponding to 113, 25, and 10 ng of GRFT released/mg fiber, respectively) resulted in different amounts of surface-bound GRFT, with ~260, 140, and 32 ng GRFT remaining on the fiber surface (per mg EF) after 4 h, respectively. Therefore, while quantification of the GRFT released from both the fibers with 5 nmol GRFT and the fibers with 0.5 nmol GRFT indicates that there is more free-floating GRFT for the fibers with 5 nmol GRFT than fibers with 0.5 and 0.05 nmol GRFT, more bound GRFT was also retained on the fibers with 5 nmol. These observations, combined with the small amount of GRFT that was released from the fibers with 0.5 nmol GRFT (25 ng GRFT/mg fiber) yet that showed efficacy, indicate that surface-bound GRFT is indeed playing a prominent role in inhibition. However, we cannot negate the antiviral contribution of unbound GRFT released from the fiber surface. In fact, the decrease in the IC50 seen with GRFT-modified EFs indicates that free GRFT may have more binding sites available to attach to and inhibit HIV-1 or that, despite the multivalent modification of the GRFT fiber surface, some amount of steric hindrance may be occurring, resulting in less binding and efficacy relative to those of free GRFT. To conclude that these are the exact mechanisms, however, more detailed studies would need to be conducted.
Beyond this, our closed-system in vitro experiments may highlight factors, like unbound GRFT, that may have significant ramifications on in vivo applications. Based on our efficacy and sustained-release results, we believe that in an ideal system, a mixture of free and surface-bound GRFT may prove beneficial to achieving maximum HIV inhibition. We envision that the complementary activity of free GRFT, which may more readily distribute through the vaginal lumen and mucosa, with more durable GRFT-conjugated fibers may provide a more versatile platform to provide GRFT exposure for longer durations in the female reproductive tract. While desorbed GRFT may release from the fiber and distribute similarly to GRFT gel under physiologically relevant conditions, in situations where virus may not make direct contact with the fiber, desorbed GRFT may prove to be a complementary backup strategy to ensure complete protection against infection. Considering this situation, we believe that adsorbed GRFT, while initially seemingly undesirable from a reaction yield perspective, may be beneficial here, as it can be released and distributed during early time frames of delivery to better enable virus debilitation and inactivation in the vaginal lumen and mucosa. In contrast, and despite the lower efficacy observed with GRFT fibers, we expect that GRFT fibers may offer the potential of greater availability to inhibit virus infection for longer durations.
While GRFT adsorption and desorption may complement surface modification to enhance protection, it may be desirable to modulate GRFT adsorption to the fiber surface. To minimize GRFT adsorption to the fiber prior to use, the number of washes could be increased postreaction (currently, 3 are conducted) or the fiber could be presoaked to eliminate surface-adsorbed GRFT. Furthermore, while factors such as surface roughness and fiber diameter can contribute to protein adsorption (97, 98), with these factors being constant, the role of surface chemistry often plays a significant role in protein adsorption. To this end, many groups have employed surface modification strategies to alter the surface properties of EFs to reduce protein adsorption. One strategy that has been most successful in reducing protein adsorption is conjugating polyethylene glycol (PEG) to the surface of polymeric vehicles in brush-like or other configurations (99,–101). Similarly, copolymers, such as monomethoxy-PEG–PLGA, may be used to fabricate fibers that contain similar reactive groups while eliminating nonspecific adsorption to the fiber surface.
While GRFT-modified EFs demonstrated a significant ability to inhibit HIV infection, interestingly, we observed that blank fibers alone demonstrated a 38% decrease in HIV infection, suggesting the potential for unmodified EFs to physically inhibit HIV infection (Fig. 5). This finding is consistent with those of other studies that observed HIV binding to polystyrene fibers and obstruction of sperm transport through fiber meshes (33, 35).
Although we plan to perform a more in-depth assessment of the microscopic interaction of these EFs with HIV, here we performed a preliminary study to further investigate the optimal ability of fibers to physically inhibit virus penetration. In the experiment described here, we evaluated the ability of unmodified fibers to prevent virus penetration when virus was applied to the top of a fiber suspended in a transwell plate (see Fig. S2 in the supplemental material). This experimental setup was designed to represent an ideal scenario in which all viruses make physical contact with the fiber. Virus was administered to a transwell plate (with the membrane removed) that contained either no fiber or unmodified PLGA fibers in contact with the underlying medium, which contained TZM-bl cells, for up to 3 days. We discovered that unmodified PLGA EFs completely inhibited HIV penetration for up to 3 days in vitro by acting directly with virus. From this preliminary physical inhibition study, we believe that the physical nature of multilayered porous and hydrophobic unmodified EFs may additionally promote virus immobilization or a lack of penetration and may contribute to the ability of unmodified (and modified) EFs to inhibit HIV-1 infection, if virus is directly in contact with the EFs. Similar to the above-mentioned studies, we suggest that HIV debilitation may be attributed to the large surface area and web-like, tortuous microstructure of fiber meshes available for HIV binding. In addition, the interaction between virus surface proteins and hydrophobic fibers may result in an inability to detach from or pass through fibers to infect cells.
Of note, there were clear differences in the level of virus inhibition achieved between these two assays (Fig. 5; see also Fig. S2 in the supplemental material). We attribute these differences to the assay design. In the latter physical penetration assay, virus was administered only on top of the fibers, which replaced a transwell membrane in the well. This represents an ideal scenario where all viruses are in contact with the fiber. In the former, more realistic virus inhibition assay, the fiber was placed in a well but did not completely cover the bottom of the well and instead floated in the medium. This inexact positioning, where some viruses may not come into contact with the fiber, is closer to what we may expect in vivo, highlighting the importance of direct contact and surface coverage with the fiber for the contribution of physical (and somewhat chemical) inhibition to overall efficacy. From these experiments, we hoped to elucidate the impact that direct contact of the fibers with virus, rather than a more relevant physiological mimic, where the fiber may not be in direct contact (at all times) with the virus but is simply in the same microenvironment, has on inhibition. From our efficacy data, we are encouraged that a combination of both chemical and physical effects may play a role in garnering virus protection; however, a more in-depth study needs to be conducted to clearly determine the physical mechanism(s) of inhibition.
In parallel with the above-described experiments that sought to understand how material modification or a lack thereof contributed to overall efficacy, we also wanted to investigate how the virus concentration impacts fiber efficacy. To do so, we evaluated the efficacy of the top three most active EF formulations against various virus concentrations (Fig. 6). While only the top two most infective virus concentrations produced 100% infection in untreated cells, the top five most infective virus inocula produced full or partial infections. Overall, we observed a surface density-dependent response in HIV infectivity across the exposures to the different virus inocula, demonstrating that similar trends to impart protection exist as a function of material modification and virus inoculum. Of note, we observed inhibition similar to that seen in Fig. 5, where the virus was diluted 8-fold prior to administration, relative to the inhibition seen with the third 3-fold dilution of virus in this assay (essentially a 9-fold dilution of the original virus stock).
Finally, once we characterized GRFT-EFs and established the relevant concentrations needed to achieve efficacy, we assessed the cytotoxicity of the top three most active modifications of GRFT-EFs for cells of the VK2, End1, and Ect1 cell lines at days 1, 2, and 3 (Fig. 7). We administered 2 mg/ml of the GRFT-EFs containing 5, 0.5, and 0.05 nmol GRFT per mg EFs to cells of all three cell lines. We observed that all EFs were nontoxic to cells in all three vaginal cell lines, with the cells showing greater than 93% viability after treatment, indicating their potential for future translation in vivo. We expected favorable cytotoxicity results, as a feature of this platform is that an FDA-approved polymer is used with the promising biologic GRFT, which has demonstrated outstanding safety profiles both in vitro and in vivo (46, 56). These results are particularly important for translation in vivo, as cytotoxicity and inflammation have been shown to increase the propensity for infection in clinical trials (2,–5, 7).
In this work, we demonstrated that GRFT-EFs potently inhibit HIV infection in vitro, in addition to providing a nontoxic platform for future translation. Additionally, GRFT-EFs may provide a unique and potent way to mitigate infection, enabling the use of a combination of EF surface modification strategies with encapsulated agents in future work. A topically applied microbicide such as GRFT may be customized to bind to a variety of pathogen targets (e.g., HIV gp120) and release multiple agents that disrupt diverse viral infections (e.g., genetic agents, small-molecule drugs, or antivirals), and EFs have the potential to provide exciting delivery options to the microbicide and infectious disease communities. In future work, we will seek to obtain a more complete understanding on a temporal basis of how GRFT is interacting with both the fibers and virus through biophysical analyses. This will enable us to examine the potential of released GRFT to inhibit HIV infection and the corresponding duration of inhibition or, conversely, if GRFT is attached to fibers, how the kinetics of binding and inactivation occur from a materials perspective. Various parameters, including fiber diameter, pore size, surface area, and the distribution of GRFT conjugation, can be studied in relation to virus interaction, immobilization, and inhibition. By obtaining a more in-depth understanding of the physical mechanism of the interaction of these EFs with virus, we will better understand the physical design parameters to guide our fabrication and physical design. These improvements will better enable us to take advantage of the dual modality of EFs to further enhance efficacy and target multiple STIs in one platform.
We sincerely thank Stuart Williams III for the generous use of his electrospinning equipment.
This work was supported in part by NIH grant R21/R33 AI088585 to N.M. and NIH grant U19 AI113182 to K.E.P.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AAC.00956-16.