Hepatitis B virus hepatotropism is mediated by specific receptor recognition in the liver and not restricted to susceptible hosts


  • Potential conflict of interest: A.A. is shareholder in Myr-GmbH, licensee of Myrcludex B. S.U. is co-applicant and co-inventor on patents protecting HBVpreS-derived lipopetides (Myrcludex B) for the use of HBV/HDV entry inhibitors. S.U., W.M., and U.H. are co-inventors on patent applications protecting the use of HBVpreS-derived lipopeptides as vehicels for liver-specific drug delivery.

  • Supported by the Bundesministerium für Bildung und Forschung (BMBF), “Innovative Therapieverfahren,” grant number 01GU0702, and Vision 7 GmbH.


The human hepatitis B virus (HBV) causes acute and chronic infections in humans and chimpanzees. HBV infects its hosts at minimal inoculation doses and replicates exclusively in hepatocytes. The viral determinants for the pronounced species specificity and the high efficacy to address hepatocytes in vivo are unknown. Previous findings showed that N-terminally myristoylated peptides constituting a receptor binding domain of the HBV large envelope (L)-protein block HBV entry in vitro and in vivo. Here we investigate the ability of such peptidic receptor ligands to target the liver. Injection of radioactively labeled HBVpreS-lipopeptides resulted in rapid accumulation in livers of mice, rats, and dogs but not cynomolgus monkeys. Without lipid moiety the peptide was excreted by renal filtration, indicating its possible retention through the lipid by serum factors. Organ distribution studies of 26 HBVpreS peptide variants revealed a correlation of HBV infection inhibition activity and the ability to target mouse livers. Together with complementary studies using primary hepatocytes of different species, we hypothesize that HBV hepatotropism is mediated through specific binding of the myristoylated N-terminal preS1-domain of the HBV L-protein to a hepatocyte specific receptor. Moreover, the restricted infectivity of HBV to human primates is not generally determined by the absence of this binding receptor in nonsusceptible hosts (e.g., mice) but related to postbinding step(s) (e.g., membrane fusion). Conclusion: HBVpreS-lipopeptides target to the liver. This observation has important clinical implications regarding the pharmacokinetic properties of Myrcludex B, the first entry inhibitor for HBV/HDV. In addition, this provides the basis for the application of the peptides as vehicles for hepatocyte-specific drug targeting. (HEPATOLOGY 2013)

See Editorial on Page 9


DMSO, dimethylsulfoxide; ge, genome equivalent; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HDV, hepatitis delta virus; L-protein, hepatitis B virus large surface protein; PHH, primary human hepatocytes; PTH, primary Tupaia hepatocytes; p.i., postinfection; RP-HPLC, reversed-phase high-performance liquid chromatography; SPECT/CT, single photon emission computed tomography/computed tomography.

The human hepatitis B virus (HBV) causes acute and chronic liver infections. Worldwide, 350 million people are persistently infected.1 Chronic HBV will remain a major global health problem, despite the availability of vaccines. Therapies (interferon-alpha [IFNα] and five nucleoside analogs) are limited and mostly noncurative. The failure to cure infections with the established therapies demands the development of novel strategies aiming at the interference with hitherto unaddressed steps in the viral life cycle, e.g., entry.2 HBV is a member of the hepadnaviridae.3 Hepadnaviruses are the smallest enveloped DNA viruses that replicate by way of reverse transcription of a pregenomic RNA (pgRNA) intermediate. During assembly the nucleocapsid acquires three viral envelope proteins termed large (L), middle (M), and small (S). They are encoded in one open reading frame and share the S-domain, which is required for membrane anchoring. In addition to the S-domain, M contains an N-terminal hydrophilic extension of 55 amino acids (preS2), while L is further extended by 107, 117, or 118 amino acids (genotype-dependent), termed preS1.4 The myristoylated preS1-domain of L plays the key role in HBV and hepatitis delta virus (HDV) infectivity through mediating attachment and specific receptor binding.5-13

Hepadnaviruses show pronounced species specificities. In addition to humans, only chimpanzees are susceptible to HBV.14 The fact that mice and rats are refractory to HBV has been attributed to the lack of either entry factor(s) or the presence of postentry restriction factors. Since delivery of plasmid-encoded HBV-genomes into hepatic cells of nonsusceptible species promote virion secretion, it is assumed that host constraints are related to early infection events.15 Another peculiarity of HBV is the efficacy to selectively infect hepatocytes in vivo, a feature that becomes particularly apparent when the virus is administered at very low inoculation doses. Injection of <10 virions establishes an infection in chimpanzees.16 The hypothesis that the species specificity and the extraordinary liver tropism are associated with an early step of HBV infection, e.g., specific receptor recognition, is attractive. However, experimental proof for this was hampered until cell culture systems for HBV and HDV, a virusoid using the HBV envelope to propagate, became available.17, 18 Using subviral particles and primary Tupaia hepatocytes (PTH), Glebe et al.13 showed that specific binding depends on the L-protein. We identified HBV L-protein-derived lipopeptides that block HBV and HDV infection of primary human hepatocytes (PHH) and HepaRG cells.7, 19, 20 The peptides are active when subcutaneously injected into PHH-transplanted urokinase plasminogen activator, severe combined immunodeficient (uPA-SCID) mice, a small, immune-deficient animal model used to study HBV infection in vivo.21 They represent the N-terminal 47 amino acids of the preS1-domain of HBV (HBVpreS/2-48myr) and include the naturally occurring modification with myristic acid. Since preincubation of cells with HBVpreS/2-48myr blocks infection they presumably address a receptor. Direct evidence, therefore, comes from in vitro binding studies using fluorescently labeled HBVpreS-derived lipopeptides (Meier et al.22).

We here followed the question of whether this interaction explains the in vivo hepatotropism of HBV. We labeled HBVpreS-lipopeptides with radioactive isotopes and investigated the in vivo distribution in several species. We demonstrate enrichment of only the inhibitory peptides in the liver of mice, rats, and dogs indicating that these animals, although not susceptible to HBV infection, express an HBV-preS-specific receptor.

Materials and Methods

Synthesis and Characterization of Peptides.

Peptides were produced by solid-phase synthesis using the fluorenylmethoxycarbonyl/t-butyl (Fmoc/tBu) chemistry on an Applied Biosystems 433A peptide synthesizer. Coupling conditions and the attachment of acyl residues were performed as described.23 Purification was achieved by semipreparative reverse-phase high-performance liquid chromatography (RP-HPLC) on a Chromolith SemiPrep RP-18e column (100 × 10 mm). Analytical analyses were performed on an Agilent 1100 HPLC system using a Chromolith Performance RP-C18e column (100 × 4.6 mm). As eluents, 0.1% trifluoroacetic acid (TFA) in water (eluent A) and 0.1% TFA in acetonitrile (eluent B) were used. Conditions: linear gradient from 0 to 100% B within 5 minutes; flow rate 4 mL/min; UV absorbance λ = 214 nm. The identity of the peptides synthesized was verified by HPLC-MS (mass spectrometry) analysis (Exactive, Thermo Fisher Scientific).

Labeling of Peptides.

For radiolabeling a 1 mM stock solution of the respective peptide in water/dimethyl sulfoxide (DMSO) was prepared. The peptides were synthesized with an artificially introduced D-tyrosine at the C-terminus. Labeling with 123I, 125I, or 131I was performed by the chloramine-T method.24 The reaction solution was purified by semipreparative HPLC using a Chromolith Performance RP-18e column (100 × 4.6 mm) applying a linear gradient of 0.1% TFA in water (eluent A) to 0.1% TFA in acetonitrile (eluent B) within 10 minutes; flow rate 2 mL/min; UV absorbance λ = 214; γ-detection.

Organ Distribution Studies of HBV PreS-Peptides.

Organ distribution studies were performed in female NMRI mice and in Wistar rats. Experiments were in compliance with the German animal protection laws. 100 μL of the peptide solution was administered subcutaneously in the hindleg or as an intravenous bolus injection into the tail vein. Three animals were sacrificed at each point in time and the radioactivity in peripheral blood, heart, lung, spleen, liver, kidneys, muscle, brain, intestine, and injection site (=tail, after intravenous injection only) was determined: The samples were weighed and the organ-associated radioactivity was quantified in a gamma counter (Berthold LB951G). The organ-associated activity was related to the injected dose (ID) and expressed as a percentage of the injected dose per gram tissue (%ID/g).

Planar Imaging and Single Photon Emission Tomography (SPECT)/Computed Tomography (CT).

After injection into anesthetized animals, scintigraphic images were obtained using a γ-camera (Gamma Imager, Biospace, France). The recording time was 5 minutes. Scintigraphic images of monkeys and dogs were performed employing a SPECT/CT camera (Millennium VG Hawkeye Gamma camera, GE Healthcare). The recording time was 10 minutes, matrix: 265 × 265 pixels. Regions of interest were determined and analyzed by the program Xeleris, ritidos 1a supplied with the camera. The sum of rvl (ventral/anterior) and ldr (dorsal/posterior) images were used for calculation. For SPECT/CT the total acquisition time took 1 hour (matrix: 128 × 128 pixels) where the dual-head rotating camera was surrounding the animal to obtain a 3D image.

Autoradiography of Liver Sections.

After administration of the labeled peptide, animals were sacrificed and the liver was dissected. It was frozen in liquid nitrogen and embedded in Tissue-Tek (Sakura). The fixed tissue was cut into sections (14 μm and 20 μm thick) in a cryostat (Reichert-Jung, 2800 Frigocut) and mounted onto glass slides. Sections were fixed with paraformaldehyde (PFA) and washed with phosphate-buffered saline (PBS) and ethanol. Autoradiography was performed by incubation of glass-mounted sections on an Amersham Hyperfilm MP.

Liver Extraction of Myrcludex B-y-131I.

To determine the peptide integrity, 131I labeled genotype C-derived preS1-lipopeptide (Myrcludex B)-y was extracted 1 hour, 4 hours, 8 hours, and 24 hours after subcutaneous injection from the liver of Wistar rats. One mL water per gram frozen liver tissue was added. Following cell disruption with a Potter S homogenizer, 1 mL acetonitrile was added. After centrifugation (10 minutes at 2,500 rpm, 4°C), the supernatant was analyzed by the HPLC system using a γ-detector. Liver samples were characterized by a liquid/liquid extraction procedure and HPLC-MS/MS analyses by Prolytic (Frankfurt, Germany).


Synthesis and Characterization of Iodinated HBVpreS-Lipopeptides.

Lipopeptides from the N-terminal part of the HBV L-protein bind PHH, PTH, and HepaRG cells and prevent productive entry of HBV in vitro6, 7, 17, 19, 20, 25 and in vivo.21 To study the distribution of these peptides in vivo we synthesized a traceable variant of HBVpreS/2-48myr, the most thoroughly studied preS-derived entry inhibitor which allows the introduction of radioactive iodine at the C-terminally fused D-Tyr (y) residue (Fig. 1A). Labeling at this site certifies that the label is either the complete peptide, free iodine (released by the action of de-iodinases), or a proteolytic degradation product lacking the N-terminal myristoyl moiety. The two latter, if they would be generated in vivo, are impaired in preS-receptor binding.7

Figure 1.

Schematic representation of the HBVpreS-derived lipopeptide HBVpreS/2-48myr, its purification, and mass spectroscopic analysis. (A) Domain structure of the HBV L-protein with its N-terminally myristoylated preS1-domain (pink), the preS2-domain (orange), and the S-domain (red) containing the four putative transmembrane domains (yellow). The acylated HBVpreS-derived lipopeptide HBVpreS/2-48myr is depicted below. An additional D-tyrosine residue (labeled with the small single code letter “y”) was added to the C-terminus to facilitate selective labeling with different radioactive iodine isotopes. (B) RP-HPLC chromatogram of unlabeled HBVpreS/2-48myr-y in comparison to the 131I-labeled lipopeptide. The shift of the retention time for the iodinated product corresponds to the dead volume between UV and γ-detector. (C) Mass spectrometric analysis of HBVpreS/2-48myr-y. The three peaks at 1,854.89 Da, 1,391.42 Da, and 1,113.34 Da represent the 3×, 4×, and 5× charged molecule. The peak at 1862.22 Da represents the sodium adduct. The two other peaks (1,848.89 Da and 1,816.55 Da) result from water loss and a small side-product with a sequence lacking asparagine. The measured monoisotopic mass of 5,558.66 Da corresponds to the calculated mass of 5,558.67 Da.

HBVpreS/2-48myr-y could be produced in >98% purity and permitted efficient labeling with radioactive iodine (Fig. 1B). The mass spectrometric analysis of the peak fraction of the unlabeled peptide shows three distinct peaks corresponding to the theoretical m/z values of [M + 5H]5+ (1,113.3 Da), [M + 4H]4+ (1,391.4 Da) and [M + 3H]3+ (1,854.9 Da). One minor peak at 1,848.9 Da (loss of water) is probably caused by aspartimide formation during synthesis.26 A lack of Asn in a minor fraction (1,816.6 Da) was also detected. Quantification of the specific activity of the labeled peptide confirmed values of 0.5-1 MBq for 125/131I/nmol and 15 MBq 123I/nmol. This corresponds to one labeled peptide molecule/700 unlabeled molecules. All other peptides (Fig. 3D) were producible in the same manner (data not shown).

HBVpreS/2-48myr-y-125I Accumulates in the Liver of Mice Following Intravenous Injection.

HBVpreS/2-48myr prevents HBV infection in PHH-transplanted uPA-SCID mice at low doses. Although the peptide was detectable in the chimeric livers of these mice, no selective enrichment in the transplanted PHH was observed.21 To investigate whether HBVpreS/2-48myr is specifically directed to the liver we performed in vivo distribution studies of HBVpreS/2-48myr-y-125I in NRMI mice (Fig. 2A). To control specificity we included a randomized preS1-peptide (HBVpreS/2-48_scrstea-y-125I) (Fig. 2B).20 As depicted in Fig. 2A, 5 minutes after intravenous application of ∼45 μg HBVpreS/2-48myr-y-125I, an almost exclusive accumulation of radioactivity in the liver was observed (>83%). The signal ceases slowly and was still detectable with ∼34% of the injected dose 24 hours postinjection (p.i.). The kinetic analysis revealed a half-life time of ∼18 hours. Apart from the pronounced liver accumulation we noticed a minor signal within the first 4 hours in the bladder. The signal was lost between 4 hours and 6 hours p.i., probably through renal clearance between repeated anesthesia. The percentage of radioactivity in the bladder after 4 hours was 16%. In contrast to the hepatic enrichment of HBVpreS/2-48myr-y-125I, intravenous injection of the randomized HBVpreS/2-48scrstea-y-125I resulted in disperse body distribution. During the first anesthesia (0-4 hours) a higher percentage of the mutant peptide appeared in the bladder (30%). Six hours p.i. most of the radioactivity was eliminated; at 24 hours p.i. no peptide was detectable. Accordingly, the radioactivity was renally filtered with a half-life of about 4 hours. This was confirmed by the autoradiographic analysis of the corresponding liver sections (Fig. 2C). While an equal distribution of radioactivity in liver lobules was detectable 24 hours p.i. after application of the wildtype peptide, no signal appeared for the mutant. This implies that the HBVpreS-lipopeptide accumulation in the liver requires the integrity of the HBVpreS1-sequence, indicating a sequence-specific binding of the peptide to an HBVpreS-receptor.

Figure 2.

Specific accumulation of HBVpreS/2-48 in the liver of mice. (A) Scintigrams of HBVpreS/2-48myr-y-125I following intravenous injection in a female NMRI mouse. At the indicated times, the anesthetized mouse was placed on the γ-camera for in vivo radioimaging. During the first 4 hours the mouse was continuously kept on the camera. At the subsequent points in time, the mouse was anesthetized shortly before the measurement. (B) Scintigrams of the scrambled peptide HBVpreS/2-48scrstea-y-125I following intravenous injection in a female NMRI mouse. The schedule for the narcoses was the same as described in (A). (C) Autoradiography of liver sections prepared 24 hours after intravenous injection of the HBVpreS-derived wildtype lipopeptide and a specific control (see Fig. 3D and Supporting Fig. 1) in NMRI mice. The autoradiograms of HBVpreS/2-48myr-y-125I (left) and the mutant peptide (HBVpreS/2-48(D11/13)myr-y-125I, right) are shown in comparison to the original liver section used for the exposure of the film.

Sequence Required for Targeting HBVpreS-Derived Lipopeptides to the Mouse Liver Colocates With the Essential Sequence Element Mandatory for HBV Infection Inhibition.

To map the sequence required for the accumulation of HBVpreS/2-48myr in the mouse liver we performed in vivo distribution analyses using the iodinated peptides depicted in Fig. 3D. These peptides have been characterized with respect to their ability to inhibit HBV infection.20 The results are summarized in the right column of Fig. 3D. To obtain quantitative measures for their ability to accumulate in different organs we injected the labeled peptides intravenously and sacrificed three mice per point in time (10 minutes, 1 hour, 4 hours, and 24 hours), extracted the intestine, brain, muscle, kidney, liver, spleen, lung, heart, and blood, counted the organ-associated radioactivity, and calculated the percentage of the injected dose per gram tissue (%ID/g). As shown in Fig. 3A, a modified (Q46K), genotype C-derived preS1-lipopeptide (Myrcludex B) locates with 71.2 %ID/g in the liver 10 minutes after intravenous injection. This corresponds to ∼87% of the total amount injected.

Figure 3.

Mapping of the requirements for liver-specific accumulation of HBVpreS peptides. (A) Biodistribution of Myrcludex B, the lead component of a modified (Q46K), genotype C-derived HBVpreS/2-48myr lipopeptide that showed the lowest IC50 in infection inhibition studies. Three NMRI mice were sacrificed at 10 minutes, 1 hour, 4 hours, and 24 hours after intravenous injection of the peptide. Organs were extracted, weighed, and the radioactivity was determined with a γ-counter. The values are given in percentages of the injected dose per gram tissue (%ID/g). (B) Biodistribution of HBVpreS/1-48 (1-48), the nonmyristoylated HBVpreS-derived peptide comprising the authentic N-terminal 48 amino acids of the genotype D sequence in NMRI mice. (C) Biodistribution of HBVpreS/2-48(G12E)stea (G12Estea) an inactive lipopeptide carrying an amino acid exchange (G-E) at position 12. (D) HBVpreS lipopeptides tested for liver specific enrichment in comparison to their HBV infection inhibition activity. Sections in light gray indicate changes to the respective sequence of the HBV genotype D. The scrambled peptide is shown in white with occasional matches depicted in gray. Bars in dark gray represent amino acids that differ from the wildtype sequence (either alanine or the respective D-amino acids). The liver factor describing the ability for hepatic enrichment was calculated in the following way: The %ID/g 1 hour p.i. was divided through the mean value of distribution [%ID/g] in the remaining organs. The higher this factor the better is the specific liver accumulation. Peptide activity to inhibit HBV infection in HepaRG cells was derived from previously published data by Schulze et al.20

Within the first 4 hours the peptide-associated radioactivity in the liver remained constant. It slowly declined to 30.5 %ID/g at 24 hours after injection. At early points in time, minor levels were detectable in the blood (at 10 minutes: 2.8 %ID/g; at 1 hour: 2.0 %ID/g), in the kidneys (at 10 minutes: 2.7 %ID/g; at 1 hour: 2.3 %ID/g) and to a lower extend in heart, lung, and spleen (at 10 minutes: 3.8 %ID/g; at 1 hour: 2.6 %ID/g). No activity was associated with the brain, indicating no crossing of the blood-brain barrier. This confirms the results of noninvasive imaging obtained with genotype D HBVpreS/2-48myr-y-125I (Fig. 2A).

Notably, the organ distribution pattern entirely changed when the N-terminal fatty acids were removed. At 10 minutes p.i. >50% of the ID/g of HBVpreS/1-48-y-131I was detectable in the kidneys (Fig. 3B). The signal declined to undetectable levels within the following 4 hours. At early timepoints higher peptide levels were detectable in the blood. No specific accumulation was observed in the liver when compared to other organs. Since a similar distribution was observed for a non-myristoylated scrambled peptide (data not shown), we conclude that myristic acid may mediate binding to a serum factor preventing the 5.4 kDa peptide from filtration in the kidney. In addition, association with a serum factor may enhance resistance against serum proteases.

To substantiate the sequence-dependence for the hepatotropism of the peptide we tested the point mutant HBVpreS/2-48stea(G12E)-y-131I. This mutant is defective in HBV infection inhibition.20 Remarkably, the single amino acid substitution completely changed the organ distribution of the peptide in NRMI mice (Fig. 3C). The pronounced association with the liver was lost and the peptide did not retain in the mouse for 24 hours. Thus, acylated HBVpreS/2-48-peptides address a homologous target in mouse and human livers with comparable binding specificities for the HBVpreS1-sequence. Finally, we performed organ distribution studies using all peptides depicted in Fig. 3D. To quantify liver association, we calculated a liver enrichment factor and compared it with the inhibitory activity of the same peptide determined in infection inhibition assays (Fig. 3D). Mutants lacking their ability to interfere with HBV infection also lost their potential to accumulate in mouse livers. Inactive peptides (e.g., those with mutations in the essential receptor binding site 9-NPLGFFP-15) behaved like the scrambled mutant, while those with a residual inhibitory activity still retained some hepatotropism. This correlation supports the hypothesis that mice harbor an HBVpreS1-specific receptor which displays the same binding specificity as its human homolog.

HBVpreS/2-48myr-y-125/123I Accumulates in the Livers of Rats and Dogs but Not Cynomolgus Monkeys.

The unexpected finding that mice harbor an HBVpreS-specific receptor in the liver prompted us to perform in vivo distribution studies in other species. Taking advantage of the SPECT/CT technology, we performed in vivo time course experiments of the distribution of HBVpreS/2-48myr-y-125I in Wistar rats, beagle dogs, and cynomolgus monkeys. As shown in Fig. 4A, intravenous injection of HBVpreS/2-48myr-y-125I into the tail vein of a rat resulted in the fast and sustained liver accumulation of the peptide. Again, a minor fraction of the radioactivity was detectable in the bladder. Urine analysis, using RP-HPLC, revealed that the renally filtered radioactivity coelutes with short C-terminal degradation products of the injected lipopeptide lacking the N-terminal myristic acid moiety (data not shown) and compares to Fig. 5C. Twenty-four hours p.i. about 28% of the maximum value was still associated with the liver, indicating stable association with a receptor. A very minor fraction of the activity was associated with the thyroid. This is probably free 125I which was released from the tyrosine residue through the action of serum or tissue deiodinases. To avoid long-term burden with radioactivity, studies in dogs and cynomolgus monkeys were performed with a 123I-labeled peptide which was applied by way of the subcutaneous route. One hour p.i. a selective accumulation of the peptide to the liver of dogs was observed. The signal persisted for >48 hours. Most of the subcutaneous injected radioactivity disappeared from the site of injection within 8 hours. Like for rat and mouse, small quantities of the label accumulated in the thyroid between 8 and 48 hours following subcutaneous injection. Because 8 hours p.i. all activity was liver-associated, we account liver-specific deiodinases to be responsible for the release of the free iodine.

Figure 4.

Time course of the distribution of HBVpreS/2-48myr in rats, dogs, and cynomolgus monkeys. (A) Scintigraphic imaging at 1 hour, 6 hours, and 24 hours of the HBVpreS/2-48myr-y-125I biodistribution in a female Wistar rat. The rat was anesthetized with sevoflurane. Subsequently, 22 MBq (corresponding to ∼37 μg) of the labeled peptide were administered intravenous into the tail vein. The two circular fields covered by the γ-camera are schematically drawn in the scintigram. The positions of the thyroid, the liver and the bladder are indicated by small circles. (B) Scintigraphic imaging at 1 hour, 8 hours, 24 hours, and 48 hours of the HBVpreS/2-48myr-y-123I distribution in a female beagle dog. The dog was anesthetized with ketamine (10 mg/kg) and xylazine (2 mg/kg). 86 MBq (corresponding to ∼14 μg) of the labeled peptide were administered subcutaneous into the left hindleg (indicated by the circle at subcutaneous depot in the left two pictures). The positions of the thyroid, the liver, and the bladder are indicated by small circles. The experiment was performed with four animals. Only one representative animal is shown. (C) Scintigraphic imaging at 1 hour, 8 hours, and 24 hours of the HBVpreS/2-48myr-y-123I distribution in a male cynomolgus monkey. The monkey was anaesthetized with ketamine (10 mg/kg, induction) and profolol (12.5 mg/kg/h, maintenance). 22 MBq (corresponding to ∼8 μg) of the labeled peptide were administered subcutaneous into the left hindleg (indicated by subcutaneous depot). The positions of the thyroid, the liver, and the bladder are marked by small circles. Note that in contrast to (A,B), no specific enrichment of the radioactivity could be detected in the liver. The experiment was performed with four animals. Only one representative animal is shown. (D) Comparison of in vitro binding results from HBVpreS/2-48myr-K-FITC to primary mouse, rat, beagle, human, and cynomolgus hepatocytes by FACS analyses (see accompanying article by Meier et al.22) with the ability of HBVpreS/2-48myr to target the liver.

Figure 5.

In vivo stability of Myrcludex B extracted from rat livers. (A) Organ distribution of radioactively labeled Myrcludex B-y-131I 10 minutes, 30 minutes, 1 hour, 4 hours, 8 hours, and 24 hours after subcutaneous administration in Wistar rats (n = 3 per point in time). The table inserted represents the results of a parallel experiment conducted with nonlabeled Myrcludex B-y. The liver concentrations [ng/g] were quantified by LC-MS analyses following extraction. (B) Radio RP-HPLC analysis of liver extracts obtained 1 hour, 4 hours, 8 hours, and 24 hours after subcutaneous administration of Myrcludex B-y-131I to Wistar rats. As a standard (std) free Myrcludex B-y-131I was used. (C) Radio RP-HPLC analysis of a urine sample collected 1 hour after subcutaneous injection of Myrcludex B-y-131I into a Wistar rat. Elution was performed with the same gradient as used in (C). (D) Comparative pharmacokinetic analysis of Myrcludex B-y-131I in the liver of Wistar rats following subcutaneous or intravenous injection. Note the exclusive targeting of the intravenous injected peptide within minutes.

Cynomolgus monkeys are commonly used for toxicity studies27 and have been proposed to be suitable for the development of an HBV animal model.28 However, HBVpreS/2-48myr does not bind to primary hepatocytes of cynomolgus monkeys (Meier et al.22). We therefore analyzed the biodistribution of HBVpreS/2-48myr-y-123I in four cynomolgus monkeys using SPECT/CT technology. In contrast to dogs (Fig. 4B) we were not able to detect any significant enrichment of HBVpreS/2-48myr-y-123I in the liver of the monkeys (Fig. 4C). The weak signal supposed to be associated with the liver 1 hour p.i. did not increase with time, even though 8 hours p.i. the peptide depot in the subcutaneous tissue was not exhausted. Instead we found a disperse distribution with a major signal associated with the bladder. This resembled the distribution pattern of the scrambled peptide in mice (Fig. 2B). Twenty-four hours after injection virtually all activity was excreted probably by renal filtration. To ensure the functionality of the tracer injected into the four animals, the liver tropism of the same preparation was verified in one NMRI mouse (data not shown). Our results demonstrate that in addition to mice, also rats and dogs harbor an HBV preS-specific receptor. Thus, expression of a functional hepatocyte-specific HBVpreS-receptor is a common feature of mammals and not unique for HBV-susceptible species. However, the lack of an HBVpreS-specific receptor in cynomolgus indicates that functionality of binding has been evolutionary lost during development of the cynomolgus branch although a closer relation to humans.

Stability of HBVpreS/2-48myr in Rat Livers After In Vivo Targeting.

To evaluate the in vivo stability of HBVpreS/2-48myrand thus the expected duration of its inhibitory potential at its target organ we investigated the integrity of a 131I-labeled Myrcludex B-y peptide in the liver of Wistar rats at several points in time after subcutaneous administration. We extracted the peptide at 1 hour, 4 hours, 8 hours, and 24 hours after subcutaneous injections from livers of three animals and analyzed its integrity by HPLC. Figure 5A shows the organ distribution of the iodine-labeled peptide at 10 minutes, 30 minutes, 1 hour, 4 hours, 8 hours, and 24 hours p.i. The results matched the quantification of the unlabeled lead substance Myrcludex B-y which was quantified by standardized LC-MS extraction (integrated table in Fig. 5A). Comparable to the results in mice (Fig. 3A), ∼50% of the amount of peptide accumulates in the liver 4 hours p.i. Following extraction and separation on a RP-column at the different points in time (Fig. 5B) we noticed, that although the total signal decreased, the majority of radioactivity elutes with the full-length peptide at a retention time of 3.2 minutes. This long in vivo half-life time indicates that the peptide might remain active for days.

When analyzing the extracts from the urine of the rat 1 hour after subcutaneous injection we detected a major labeled product eluting at a retention time of ∼0.6 minutes. Some diffuse peaks eluted between 1.0 and 1.5 minutes. No radioactivity eluted in the fractions where the hydrophobic lipopeptide was expected (retention time of 3.2 minutes). Since myristoylated HBVpreS-peptides elute at retentions times >3 minutes, the activity in the bladder represent delipidated products.

Our preceding results showed that both subcutaneous and intravenous injections resulted in liver-specific enrichment of Myrcludex B-y. To investigate whether the administration route influence the bioavailability of the peptide in the liver we performed a side-to-side comparison of both delivery pathways (Fig. 5D). While intravenous injection resulted in a rapid liver accumulation of more than 95% of the peptide within the first 10 minutes, the maximal concentration following subcutaneous injection was reached 4 hours p.i. This is probably caused by the depot effect of the subcutis. At timepoints later than 4 hours the curves approximate each other. Twenty-four hours p.i. about 15% of the injected dose was still present in the liver independent of the way of administration. Thus, subcutaneous injection delays the bioavailability of the peptide in the liver by about 4 hours but does not lead to a lower overall bioavailability.


One remarkable feature of HBV is the efficiency of how the virus infects humans and chimpanzees at extraordinary low inoculation doses. This has been observed after accidental injuries with nonsterile needles29 or in chimpanzee studies.16 Two conclusions can be drawn from these observations: (1) the majority of the genome containing HBV and HDV particles is infectious; (2) HBV and HDV must have evolved a mechanism that efficiently promotes them to the liver. The molecular basis for this liver tropism is unknown.

The work presented here suggests a mechanism on the level of receptor recognition. The data were acquired through application of chemically synthesized lipopeptide fragments of the HBV L-protein that interact with and inactivate an unknown HBV receptor. We provide evidence that the ability of HBV to address hepatocytes with high efficacy is triggered by the myristoylated N-terminal preS1-subdomain of L. The exclusive targeting of the respective lipopeptides to the liver suggests that the HBV-receptor is liver-specific and not expressed in substantial amounts in other organs.

The most remarkable finding of our study is the observation that wildtype HBVpreS1-lipopetides accumulate in livers of animals that are not susceptible for HBV. Using 26 peptides with different mutations, including exchanges of single L-amino acids with their respective D-enantiomers we demonstrated a tight correlation between the liver tropism in mice and the potency to inhibit HBV infection in vitro. Thus, receptor recognition of the HBVpreS-ligand is indistinguishable between mice and humans (and according to the data presented in Fig. 4A,B, also rats and dogs). The presence of an HBVpreS-receptor in rodents was unexpected and questions the hypothesis that the refractiveness of mice against HBV infection is caused by a deficiency in receptor-binding. However, the previous identification of Tupaia belangeri as a model for HBV infection30 implied that receptor expression is not limited to only closely related human species. The presence of an HBVpreS-specific receptor in mice and rats has important implications for the systematic development of immune competent small animal models for HBV and/or HDV. Since resistance against infection cannot solely be explained by the lack of an HBV-specific binding receptor it is probably related to the lack of either a cofactor, involved in membrane fusion (which could even be functionally associated with the same molecule), or a factor controlling a subordinated step after the release of the nucleocapsid or both. Using the transplanted uPA-SCID mouse model Lutgehetmann et al.31 demonstrated that mouse hepatocytes are not susceptible to HDV infection in vivo. Given that mouse hepatocytes bind HDV we conclude that a factor/activity required for triggering membrane fusion is missing.

The presence of an HBVpreS-specific receptor in mice should also be considered when using transplanted uPA/RAG2 mice as an in vivo infection model.32 These mice are susceptible to HBV and HDV. However, only the transplanted cells (Tupaia or human hepatocytes) replicate the viral genome and secrete progeny virions. Our findings suggest that the endogenous mouse hepatocytes, although deficient in virus propagation, influence in vivo infection. They might sequester particles thereby changing the kinetics of virus spread and the serum titers. This could explain why mice with low transplantation indices are inefficient in amplification of HBV in vivo.32

The similar pharmacokinetics of the HBVpreS-derived lipopeptides in different species has important clinical implications for Myrcludex B, the lead substance of the first in line entry inhibitor for HBV/HDV infection. (1) The absence of an HBV-specific receptor excluded cynomolgus monkeys as a model for toxicity studies. (2) The fact that Myrcludex B, besides inhibiting HBV/HDV infection with an IC50 of ∼80 pM,20 almost exclusively accumulates in the liver of mice (Fig. 3A), rats, and dogs makes it very attractive as a potential drug. The combination of an extraordinary specific activity of the peptide with an exclusive targeting to susceptible cells allows subcutaneous application of very low doses. Moreover, the remarkable serum stability of the peptide and a half-life time of about 16 hours in mouse, 10 hours in rat, and 13 hours in beagle predict its therapeutic application once every 1-3 days.

The liver is the biggest human gland and acts as an important regulator for metabolism. Accordingly, an interesting option related to the pronounced hepatotropism of the HBVpreS-derived lipopeptides is their potential as vehicles to selectively transport pharmaceutics, viral vectors, liposomes, nanoparticles, etc., to hepatocytes in vivo. Thus, any hepatocyte-related disease might be selectively addressed. Direct coupling of effectors to the peptide could be useful to induce hepatocyte-specific responses by way of the activation of surface receptors (e.g., HBVpreS-conjugated interferons). Another approach would be coupling of drugs by way of cleavable linkers. Release of the active drug at the hepatocyte surface would help to specifically deliver small molecules with unfavorable pharmocokinetic properties or systemic toxicity. Examples for such approaches would be primaquine for the treatment of malaria. A third example is related to preS1-sequences being introduced into the new generation of viral gene therapy vectors in order to render them selective for hepatocytes. Such approaches may be useful for the treatment of genetic disorders, e.g., in the urea cycle. Incorporation of HBVpreS-lipopeptides into liposomes or nanoparticles could render them universal hepatotropic carriers for the delivery of a broad spectrum of molecules. Such approaches might be suitable for the future therapeutical delivery of silencing small interfering RNAs (siRNAs). Since mice carry the HBVpreS-receptor, all these experimental approaches can be tested in the respective mouse models including transgenic or knockout mice. Finally, the criteria of an only partially species specific interaction will be helpful to identify the HBVpreS-specific receptor.


We thank Karin Leotta for the rodent imaging experiments, Tamara Becker and Janine Henrici for the handling and care of the cynomolgus monkeys, and Lothar Datan for the handling of beagle dogs.