A long-acting 3TC ProTide nanoformulation suppresses HBV replication in humanized mice

Abstract

Nowadays, there is a strong request for the treatment of chronic HBV-infection with direct acting antivirals. Furthermore, prevalent human immunodeficiency virus (HIV-1) and hepatitis B (HBV) co-infections highlight an immediate need for dual long-acting and easily administered antivirals. To this end, we modified lamivudine (3TC), a nucleoside analog inhibitor of both viruses, into a lipophilic monophosphorylated prodrug (M23TC). Prior work demonstrated that nanoformulation of M23TC (NM23TC) enhanced drug stability, controlled dissolution and improved access to sites of viral replication. The present study evaluated the efficacy of a NM23TC in HBV-infected chimeric liver humanized mice. Levels of HBV DNA and HBsAg in plasma were monitored up to 8 weeks posttreatment. A single intramuscular dose of 75 mg/kg 3TC equivalents of nanoformulated NM23TC provided sustained drug levels and suppressed HBV replication in humanized mice for 4 weeks. The results support further development of this long-acting 3TC nanoformulation for HBV treatment and prevention.

Single injection of nanoformulated pronucleotide 3TC suppress HBV replication in humanized liver mice for four weeks

Graphical Abstract

Background

Chronic infection with the hepatitis B virus (HBV) remains a significant worldwide healthcare problem.¹ Besides HBV monoinfection, HBV infection is the most common among people living with human immunodeficiency virus (HIV) (PLWH)² infection, those who inject illicit drugs³ and men who have sex with men.⁴ HIV-HBV coinfection is common,^{5, 6} and immune responses to viral infections often interfere with HBV clearance, thereby promoting viral persistence. HBV, a non-cytopathogenic virus, requires activation of immune response to eliminate HBV-infected hepatocytes. Concordantly, immune cells and namely, CD4⁺ T-lymphocytes and macrophages, are HIV-permissive and cause disruptions in both innate and adaptive immunity. This disbalance between high infectivity of HBV in hepatocytes and HIV-weaken immune response in the liver is a key feature of HBV-HIV coinfection pathogenesis and a suggested target for coinfection treatment.⁷ Adherence to therapy is thus a key factor for dual viral suppression and development of drug resistance, HBV viremia and end-stage liver disease in coinfected patients.^{2, 5, 6, 8} While lamivudine (3TC) is best known for its use in combination with antiretroviral therapy (cART)⁹, it remains a potent anti-HBV medicine for resource-limited settings^{10, 11} despite the known low barrier to resistance in approximately 30% of patients after one year of exposure.¹² It is also effective for the prevention of HBV acquired by HIV-infected patients.^{13, 14} Pre-exposure prophylaxes for HBV in PLWH could potentially have a great impact on public health.¹⁵ While nucleoside reverse transcriptase inhibitors have revolutionized the treatment of HBV and other viral infections, poor patient compliance to daily dosing schedules, rapid viral mutations and limited drug access to restricted sites of infection have negatively impacted treatment outcomes. These limitations could be overcome through potent long-acting (LA) medicines. However, transformation of hydrophilic nucleoside reverse transcriptase inhibitors into LA formulations have been elusive for decades. Another obstacle that is specific to nucleoside analogs rests in the abilities to directly deliver active metabolites (as monophosphates) of the compounds as LA formulations into tissue sites of active viral replication. Since the reverse transcriptase is essential for HIV and HBV replication, the development of reservoir-targeted LA formulations of potent nucleos(t)ide inhibitors could significantly improve treatment outcomes in co-infected patients. To this end, 3TC, a nucleoside analog inhibitor of HBV and HIV-1 infections was modified into a potent hydrophobic and lipophilic phosphoramide pronucleotide (M23TC) nanoformulation (NM23TC, described in Smith et al, 2019¹⁶) as a slow-release antiviral drug. A single administration of NM23TC provided sustained and high drug concentrations in blood and tissues of rodents.¹⁶ To test whether improved half-life of 3TC in blood and tissues accumulation would translate to enhanced efficacy, anti-HBV efficacy of the long acting NM23TC formulation was evaluated in chimeric liver humanized mice, a suitable model for screening anti-HBV compounds.^17–21 In this work, TK-NOG mice with humanized livers²² were infected with HBV prior treatment. Single intramuscular administration of NM23TC suppressed HBV viral load up to one month with parallel detectable prodrug in liver tissues. A proof-of-concept study of the efficacy of a long-acting 3TC formulation in HBV-infected humanized mice was demonstrated for the first time.

Methods

Nanoformulation preparation

Modified 3TC (M23TC) was synthesized as described in Smith et al.¹⁶ Specifically, M23TC (1% w/v) at a drug concentration of 15 mg/mL was dispersed in a PBS solution (pH 7.2) containing Tween 20 (0.2% w/v, Fluka Analytical, Rankankoma, NY, USA) and mPEG₂₀₀₀DSPE (0.2% w/v, Corden Pharmaceuticals, Boulder, CO, USA) stabilizers. The dispersion was stirred for 24 hours at room temperature followed by high-pressure homogenization (Avestin EmulsiFlex-C3; Avestin Inc, Ottawa, ON, Canada) at 12,000 psi for 1–2 hours to achieve a particle size of approximately 200–300 nm. The resultant nanoparticles were purified by differential centrifugation as previously described²³. The purified formulation was characterized prior to injection (Figure 1).

Figure 1. Physicochemical characterization of lipid-coated NM23TC nanoparticles.

The chemical structure of M23TC (A), size distribution by number measured by Dynamic Light Scattering (B), SEM image showing NM23TC nanoparticle morphologies with Z_av of 360 ± 55 nm (C). White arrowheads indicate nanoparticles, with a single white marker indicating a 100nm increment on the scale. X-ray diffraction of 3TC and M23TC prodrug (D).

Nanoformulation characterization

The final drug concentration in the formulation was quantified by LC-MS/MS as well as LC-UV. Quantitation was carried out using multiple reaction monitoring (MRM) at transitions of 841.60→111.99 and m/z 841.60→211.98 on a Waters ACQUITY H-class ultra-performance liquid chromatography (UPLC) system (Waters, Milford, MA) coupled to a Waters Xevo TQD micro mass spectrometer (Waters, Milford, MA). Confirmation used a Waters ACQUITY H-class UPLC coupled to a UV detector, measuring absorbance at 272 nm. Dynamic light scattering was used to measure particle size, polydispersity index (PDI), and zeta potential at a nanosuspension concentration of approximately 200 μg/mL. Nanoparticle morphologies were analyzed by scanning electron microscopy (SEM). Briefly, aqueous nanosuspensions at a concentration of 1 mg/mL were further subjected to ten-fold dilution in double distilled water prior to collection on membrane filters and air-dried overnight at room temperature. The samples were then sputter-coated with a thin layer of chromium using a Denton Desk V and examined on a Hitachi S4700 Field Emission SEM under the following conditions: “5.0kV, 10.9mm, x 50.0k, SE. Particle crystallinity was evaluated by powder X-ray diffraction (XRD) carried out in the 2θ range of 5–70 using PANalytical Empyrean diffractometer (PANalytical Inc., Westborough, MA, USA) with Cu-Kα radiation (1.5418 Å) at 40 kV, 45 mA setting. A mask of 20 mm and a divergence slit of 1/8° were used on the incident beam path. The powder samples were spread on low background quartz sample holder and the diffraction data was collected in steps of 0.013 degrees by continuously scanning the source and the solid state PIXcel3D detector with a scan rate of at the rate of 0.011 °/s. Ni-foil Kb filter was introduced in front of the detector to remove possible spurious peaks originating from Cu-Kb radiation.

NM23TC Pharmacokinetics (PK) and Biodistribution (BD) in mice

PK studies were performed in mice and rats. Specifically, Sprague Dawley rats (Jackson Labs, Bar Harbor, ME, USA) were administered a single 75 mg/kg 3TC or equivalent dose of 3TC as NM23TC intramuscularly (IM) into the caudal thigh muscle to determine PK over 4 weeks. 3TC was dissolved at 37.5 mg/mL in PBS to deliver a target injection volume of 50 μL. Similarly, NM23TC was dosed at 50 μL as an aqueous nanosuspension. For pharmacokinetics, whole blood and plasma samples were analyzed by LC-MS/MS to determine parent, prodrug and 3TC-triphosphate levels as previously described¹⁶. Briefly, 25μL blood was collected in acetonitrile and stored at −80°C until analyzed. Blood was collected in EDTA coated tubes (BD Microtainer) and centrifuged at 2000 x g for 5 minutes to separate plasma as the top layer. Liver, spleen, and lymph nodes were collected and analyzed for drug and/or 3TC-TP levels. Briefly, each tissue was homogenized in a solution of 90% methanol/10% water. After homogenization, 100 μL of the tissue was then added to 1mL of ice-cold methanol and extracted by protein precipitation as previously described¹⁶. Levels of 3TC-TP in lymph nodes and spleen were quantified by LC/MS-MS according to previously published protocols²⁴.

Generation of a humanized liver TK-NOG mouse model and HBV infection

NOD.Cg-Prkdc<scid>Il2rg<tm1Sug>Tg(Alb-UL23)7–2/ShiJicTac (TK-NOG)²² mice are specially designed strain that allows conditional depletion of mouse hepatocytes by ganciclovir injection (GCV). These are transgenic mice with expressed Herpes simplex virus tyrosine kinase (TK) under mouse albumin promoter. These animals are highly immune deficient, and their liver tissue can be repopulated by human hepatocytes injected intra-splenically. TK-NOG mice were maintained in a specific pathogen-free facility and treated humanely. All mouse studies were conducted in strict accordance with the Guide for the Care and Use of Laboratory Animals from the University of Nebraska Medical Center (UNMC). All experimental protocols were approved by the Animal Care and Use Committee of UNMC.

A liver chimeric TK-NOG mouse model was generated as previously described ²². Briefly, TK+ males 8–10 weeks of age were selected by genotyping and injected with ganciclovir 10 and 30 mg/kg seven and five days to achieve significant damage of mouse liver reflected by elevated ALT levels to 200–400 IU/ml before transplantation of human hepatocytes. Human hepatocytes were obtained from Lonza (lot #4145; Walkersville, MD, USA). Two million of hepatocytes were infused intrasplenically. Starting from one-month post transplantation, the levels of hepatocytes engraftment were monitored. The chimerism rate correlated with serum human albumin levels²² measured using the Human Albumin ELISA kit (Bethyl Laboratories Inc., Montgomery, TX).

At two months post- transplantation, animals were infected intravenously with patient-derived sera samples containing ~10⁶ HBV DNA (n=11). Treatment started at two months post-infection, and mice were observed for two months post NM23TC injection. The duration of experimental animals life reached ~9 months.

Treatments

The efficacy of the prodrug formulation was evaluated at eight weeks post-infection when HBV DNA levels in peripheral blood had been established. Specifically, HBV infected mice were administered a single intramuscular injection of NM23TC nanosuspension at a dose of 75 mg/kg 3TC equivalent and evaluated for viral suppression over one month.

Measurements of HBV DNA and hepatitis B surface antigen (HBsAg) levels in the blood

Starting one-month post infection, HBV DNA levels in mouse peripheral blood were measured using the COBAS TaqMan HBV Test (Roche Diagnostics, Switzerland) with a lower limit of detection of 20 IU/ml (1 IU=5.6 DNA copies). The samples were diluted 20–30-fold and detection limits were 400–600 IU/ml (2240–3360 DNA copies/ml).

HBV surface antigen (HBsAg) levels were measured by ELISA kit (Cell Biolabs, Inc, San Diego, CA). The plasma samples were diluted and run along with the provided standards and expressed as ng/ml as described in the assay protocol.

Immunohistochemistry

Tissues were fixed with 4% paraformaldehyde overnight at 4°C and then embedded in paraffin. Five-micron sections were cut from the paraffin blocks, mounted on glass slides, and subjected to immunohistochemical staining with mouse monoclonal antibodies for cytokeratin 18 (clone DC 10, 1:33 dilution) from ThermoFisher Scientific, hepatitis B core antigen (clone LF161, 1:50 dilution), and rabbit polyclonal antibodies to hepatitis B surface antigen (PAB361C, 1:50), both from Innovex Bioscience, Richmond, CA. Polymer-based horseradish peroxidase-conjugated anti-mouse systems were used as secondary detection reagents and were developed with 3,3′-diaminobenzidine. All paraffin-embedded sections were counterstained with Mayer’s hematoxylin. Bright field images were obtained with a Leica DM1000 LED.

Statistical Analysis

Statistical significance was determined using a one-way ANOVA with the p ≤ 0.05 being considered significant. Details are shown in Figure’s legends.

RESULTS

NM23TC physical characterization

The synthesis and characterization of M23TC (Figure 1A) was performed as described previously¹⁶. Lipid coated and stabilized nanoformulations of M23TC (NM23TC) were produced as aqueous nanosuspensions by high-pressure homogenization with a drug loading of 75%. The measured nanoparticle sizes were in the range of 225–360 nm (Figure 1B). Specifically, the measured hydrodynamic diameter particle size (Z_av), polydispersity index (PDI) and zeta potential of NM23TC nanoformulation were 360 ± 55 nm, 0.26 ± 0.01 PDI, and −35 ± 2.7 mV, respectively. The narrow polydispersity index and negative surface charge demonstrated formulation stability and homogeneity. Evaluation of nanoparticle by SEM revealed spherical morphologies for the lipid-coated nanoparticles (Figure 1C). X- ray diffraction spectra showed the changes in the crystalline structure of the prodrug (broad peak) compared to sharp diffraction peaks of the unmodified drug, a further confirmation of molecular structural changes (Figure 1D).

Pharmacokinetics

For PK studies comparing NM23TC to 3TC, male Sprague Dawley rats were injected with 75 mg/kg 3TC or equivalent as NM23TC (n=5 and 3, respectively) into the caudal thigh muscle and characterized as previously described¹⁶. Rats were selected for initial PK screening studies since they have large muscle mass compared to mice and therefore suitable for assessment of intramuscular injectable formulations. The volume of blood required for extensive drug and intracellular nucleoside reverse transcriptase inhibitor active metabolite quantitation necessitated the use of rats in addition to mice. The animals were bled at 2 hours, 1 day, 7 days, 14 days, 21 days, and sacrificed on day 28. Previous work from our laboratory demonstrated that sustained and high prodrug concentrations in blood and tissues could be achieved after a single intramuscular injection of NM23TC¹⁶. The prodrug exhibited similar decay kinetics in blood and plasma. At days 1 and 28, prodrug concentrations in blood were 711.6 and 103.1 ng/mL, respectively. At days 1 and 28 prodrug concentrations in plasma were 494.1 and 31.3 ng/mL, respectively. In contrast, 3TC provided high initial native drug concentrations of 4763.6 ng/mL after 2 hours that rapidly declined after 1 day to 17.1 ng/mL (Figure 2 A). Similarly, mice were injected with 75 mg/kg 3TC equivalent as NM23TC (n=8) into the caudal thigh muscle followed by drug quantitation in blood and plasma every two weeks and at one and two months in the liver (Figure 2B, C). The high M23TC levels in tissues compared to blood or plasma are suggestive of tissues drug depots for sustained release of M3TC. We recently demonstrated that high prodrug concentrations for NM23TC are reflective of intracellular and tissue 3TC-TP levels¹⁶. Since prodrugs are pharmacologically inactive compounds that require enzymatic activation in physiological conditions, the differences in drug concentrations between rats and mice was likely due to differences in the expression of hydrolytic enzymes²⁵.

Figure 2: Plasma, blood and tissue drug levels after a single intramuscular 75 mg/kg 3TC equivalents dose of NM23TC compared to native 3TC.

Drug levels over 4 weeks in rat plasma (A). Blood M23TC levels in mice over 8 weeks (B). Liver M23TC levels at 4 weeks in mice (C). The detailed synthesis scheme and pharmacokinetics profiles are described in Smith et al.,¹⁶.

The anti-HBV activity of NM23TC

To evaluate anti-HBV activity of NM23TC, we used humanized liver mice (Figure 3A). TK-NOG mice were transplanted with human hepatocytes, and two months later, after confirmation of human albumin (Alb) concentration in peripheral blood, animals were infected intravenously (n = 11). Following confirmation of HBV DNA in peripheral blood (week 0), animals were administered a single intramuscular dose of 75 mg/kg 3TC equivalents as NM23TC. In the set of experimental animals with HBV DNA levels in peripheral blood <10⁶ copies/ml (n = 8), single intramuscular administration of NM23TC reduced the viral load by 1.13 (p = 0.0004) and 1.17 (p = 0.0145) log₁₀ as shown in Figure 3B. HBV DNA levels of untreated (n=3) and infected control mice remained constant during the study period. In animals with initially high peripheral viral load (n = 3, Figure 3C), single administration was less effective and reduced viral load by 0.65 (p = 0.039) and 0.84 (p = 0.020) log₁₀. In these animals, we observed suppression for 4 weeks with the rebound of HBV DNA in peripheral blood at 6 weeks. The levels of human albumin remained stable (Figure 3D), and drug administration did not affect mouse body weight (data not shown). The levels of HBsAg were not changed (Figure 3E). High levels of the drug were detected in liver tissue at week 2 (7.7 and 16.2 μg/g) and at week 4 (4.8, 0.6, 1.8 μg/g). Drug levels in blood were 43.7 and 96.1 ng/ml and 1.5, 8.8, 3.2 ng/ml in animals euthanized at weeks 2 and 4 post- drug administration, respectively (Figure 3F and G). The rebound of HBV in the mouse with higher human hepatocytes engraftment levels and HBV levels before the treatment was confirmed by tissue histology for HBcAg and HBsAg staining (Figure 4). In animals with suppressed HBV peripheral viral load, rare HBc-positive cells were found (Figure 4E). In animal with the significant rebound of HBV replication, areas with multiple HBcAg-positive cells were evident (Figure 4F).

Figure 3. Evaluation of 3TC formulation efficacy.

A) Experimental setup for HBV efficacy evaluation on TK-NOG mice. B & C) Reduction of HBV peripheral blood viral load in animals with low (B) and high (C) base value (n=8 and n=3, respectively). **, * - P values 0.01 and 0.05, respectively, obtained by mixed-effect analysis with Holm-Sidak’s multiple comparison test (B) and ANOVA and Friedman multiple comparison test. C) At the two- and four-weeks reduction was statistically significant. D) Human albumin levels in blood were consistent over experimental conditions. E) HBsAg levels were also not changed. F) M23TC levels in liver and G) plasma in animals euthanized at 2- and 4-weeks post treatment. At 8 weeks no detectable concentration. The orange triangles represent animals euthanized at 2- and 4-weeks post 3TC administration, respectively (B, D, E). Green diamonds represent animals with high HBV viral load (C, D, E). Black circles represent three untreated HBV infected mice. Clear diamonds on B, D and E represent NM23TC treated animals observed for 8 weeks.

Figure 4. Detection of HBV infected human hepatocytes in NM23TC treated animals.

Representative images of liver sections of two HBV infected animals after NM23TC treatment. One with suppressed HBV replication below levels of detection (A, C, E) and second with rebounded HBV replication (B, D, F). A and B) Liver sections stained for human cytokeratin-18 confirm the presence of human hepatocytes. C and D) Staining for HBsAg shows the presence of infected human hepatocytes. E and F) Representative images of liver sections stained for HBcAg. E) The animal with suppressed HBV replication at 4 weeks post treatment and rare HBcAg-positive human cells (arrow). F) The animal with the rebound of HBV and field of hepatocytes with nuclear staining for HBcAg. Original magnification 200×.

DISCUSSION

Chronic hepatitis viral infections have emerged as one of the leading causes of morbidity and mortality among those living with HIV infections. Given the similarities between HBV and HIV polymerases that perform viral RNAs reverse transcription, nucleosides analogs, such as lamivudine (3TC), are used to manage HIV⁹ and HBV infections.^{5, 11} Lamivudine is a cytidine analogue that undergoes intracellular activation into the triphosphate form (3TC-TP), after which incorporates into the growing chain of DNA during reverse transcription of the first strand of HBV DNA and synthesis of the second strand of HBV DNA resulting in chain termination thereby inhibiting HBV DNA synthesis. Lamivudine was approved by the U.S. Food and Drug Administration for use in chronic HBV infection in 1998; however, a lifelong daily oral high dose of 100 mg is required to achieve viral suppression. Other limitations of 3TC include a low genetic barrier to resistance, poor regimen adherence and restricted drug entry into viral reservoirs. Limited distribution of therapeutic drug concentrations into sites of infection could lead to limited efficacy and emergence of drug-resistant virus strains. These underscore the need for improved formulation strategies to facilitate sustained intracellular and tissue drug delivery. Therefore, we optimized 3TC’s efficacy by developing a hydrophobic and lipophilic monophosphorylated form of the drug to facilitate the manufacture of long-acting formulations with improved in vitro efficacy and pharmacokinetic profiles in Sprague Dawley rats¹⁶. The 3TC ProTide was designed to facilitate cellular and tissue delivery of 3TC triphosphate. Intracellular delivery of monophosphorylated forms of nucleoside reverse transcriptase inhibitor analogs has been shown to improve drug potency^26–28. Following successful synthesis and characterization of the formulation, our present proof of concept study in humanized mice sought to determine whether the 3TC prodrug formulation could be used in the management of HBV infections.

We have previously shown that macrophages serve as drug depots and play a major role transporting drug nanoparticles into tissues^{29, 30}. Given that HBV infects and replicates in liver hepatocytes, optimal drug delivery strategies should target anti-HBV medicines to the liver where native drugs have limited access. Improved liver delivery of NM23TC formulation in HBV-infected chimeric mice resulted in the significant and sustained reduction in viral replication (~1 log₁₀) after a single intramuscular administration of the prodrug at a dose of 75mg/kg native 3TC equivalents. For comparison, a study by Tsuge et al.,³¹ demonstrated a 2.8 log₁₀ reduction in HBV DNA levels after daily oral treatment of HBV-infected liver chimeric mice with 30 mg/kg/day of 3TC over 6-weeks.³¹ It is worth noting that the reported total drug used over the six-week treatment period translates to 1260mg/kg for the oral dose compared to 75 mg/kg for the long-acting formulation of NM23TC over the same period of time. Prior studies have shown that efficient delivery of nucleoside analogs into cells and tissues could lead to reduced dosage, improved efficacy, and drug safety profiles³². It is presumed that nanoparticle sizes in the range of 200–300 nm were efficiently taken up by liver resident macrophages and disseminated into hepatocytes either in the form of prodrug or 3TC-TP through cell-cell contact. As expected, the treatment of infected mice with NM23TC did not affect hepatitis B surface antigen (HBsAg) concentrations in serum. It is well established that 3TC and other inhibitors of HBV polymerase have no effect on the elimination of infected hepatocytes and reduction in HBsAg secretion.³³ HBV dissemination depends on the number of human hepatocytes. The absence of human adaptive immune responses in immunodeficient mice could account for the rapid viral rebound seen in animals with the highest number of infectable cells when the drug concentration in the liver fell below therapeutic levels. The virus rebounded to pretreatment levels at 8 weeks in the mouse that showed the highest level of human hepatocyte engraftment. This observation corresponds to the known clinical data when antiviral treatment is stopped³⁴, as well as the experimental data obtained on a different model of humanized liver mice³⁵.

In conclusion, this study highlights the potential future role of long-acting formulations in the management of HBV infections. These data, taken together, provide the first proof of concept evidence that a nanoformulated long-acting 3TC prodrug could provide sustained anti-HBV responses after a single treatment. Future studies will compare the efficacy of NM23TC formulation against native 3TC in dose escalation studies to identify a dose and formulation that would achieve maximal and durable HBV suppression without adverse events.