Structure activity relationship of dendrimer microbicides with dual action antiviral activity

Abstract

Background: Topical microbicides, used by women to prevent the transmission of HIV and other sexually transmitted infections are urgently required. Dendrimers are highly branched nanoparticles being developed as microbicides. However, the anti-HIV and HSV structure-activity relationship of dendrimers comprising benzyhydryl amide cores and lysine branches, and a comprehensive analysis of their broad-spectrum anti-HIV activity and mechanism of action have not been published.

Methods and findings: Dendrimers with optimized activity against HIV-1 and HSV-2 were identified with respect to the number of lysine branches (generations) and surface groups. Antiviral activity was determined in cell culture assays. Time-of-addition assays were performed to determine dendrimer mechanism of action. In vivo toxicity and HSV-2 inhibitory activity were evaluated in the mouse HSV-2 susceptibility model. Surface groups imparting the most potent inhibitory activity against HIV-1 and HSV-2 were naphthalene disulfonic acid (DNAA) and 3,5-disulfobenzoic acid exhibiting the greatest anionic charge and hydrophobicity of the seven surface groups tested. Their anti-HIV-1 activity did not appreciably increase beyond a second-generation dendrimer while dendrimers larger than two generations were required for potent anti-HSV-2 activity. Second (SPL7115) and fourth generation (SPL7013) DNAA dendrimers demonstrated broad-spectrum anti-HIV activity. However, SPL7013 was more active against HSV and blocking HIV-1 envelope mediated cell-to-cell fusion. SPL7013 and SPL7115 inhibited viral entry with similar potency against CXCR4-(X4) and CCR5-using (R5) HIV-1 strains. SPL7013 was not toxic and provided at least 12 h protection against HSV-2 in the mouse vagina.

Conclusions: Dendrimers can be engineered with optimized potency against HIV and HSV representing a unique platform for the controlled synthesis of chemically defined multivalent agents as viral entry inhibitors. SPL7013 is formulated as VivaGel(R) and is currently in clinical development to provide protection against HIV and HSV. SPL7013 could also be combined with other microbicides.

Figure 1. Dendrimer structure and SAR of L-lysine dendrimers against HIV-1 and HSV-2.

(A). Model representation of dendrimer structure showing central core, branches in black, red and green for first (G1), second (G2) and third (G3) generations and surface groups denoted as blue round spheres. (B). Different types of cores and branches used in the synthesis of dendrimers. (C). Seven surface groups, L-glutamic acid, 2-thio-3-SO₃Na, MNAA, 4-Ph-SO₃Na, 3,5-Ph-(SO₃Na)₂, DNAA and 3,5-Ph-(CO₂Na)₂ evaluated in SAR studies. (D). SAR of L-lysine dendrimers against HIV-1 strain NL4.3 performed in MT-2 cells. G1–G5 dendrimers were evaluated for five of the seven surface groups except for MNAA where G1-G2 and G5 were not tested and 2-thio-3-SO₃Na where G1 and G5 dendrimers were not tested. The EC₅₀ values were obtained from at least two independent assays except for G1–G5 dendrimers with the glutamic surface group, G1 dendrimers with the 2-thio-3-SO₃Na, 3,5-Ph(SO₃Na)₂, DNAA and 3,5-Ph-(CO₂Na)₂ surface groups and the G3 MNAA dendrimer, where one assay was performed. (E). SAR of L-lysine dendrimers against HSV-2 performed in HEL cells. G4-G5 and G2, G4-G5 were not tested for dendrimers with the 3,5-Ph-(SO₃Na)₂ and 4-Ph-SO₃Na surface groups, respectively. EC₅₀ values were obtained from at least three independent assays. Error bars denote standard error of the mean.

Figure 2. Chemical structure of SPL7013 and SPL7115.

(A) Structure of G4 L-lysine dendrimer with DNAA surface groups and formula weight 16,582. (B) Structure of G2 L-lysine dendrimer with DNAA surface groups and formula weight 4,187.

Figure 3. Formulated SPL7013, VivaGel®, protects against HSV-2 infection and does not increase susceptibility to HSV-2 in the mouse HSV-2 vaginal transmission model.

(A) VivaGel® provides prolonged protection against HSV-2 infection. VivaGel® (or PBS control) was delivered vaginally to groups of 10 mice and the viral challenge (10 ID₅₀) was delivered later at the times indicated. The viral inoculum infected 49 out of 50 control animals. (B) VivaGel® does not increase susceptibility to HSV-2. Groups of 40 mice were exposed to a single dose of VivaGel® and HEC placebo gel and after 12 h were challenged with a low dose (0.1 ID₅₀) HSV-2 inoculum. As previously reported , animals exposed to nonoxynol-9 and several other candidate microbicide detergents become more susceptible to infection for many hours after exposure reaching a peak at about 12 h after a single exposure. The increase in susceptibility, as shown by the range bar, indicates the expected increase in vaginal transmission rate for a low-dose viral inoculum. In contrast 12 h after exposure to VivaGel® fewer animals were infected than in PBS control animals.

Figure 4. Mechanism of action studies of SPL7013 and SPL7115.

(A) Inhibition of recombinant HIV-1 RT by dendrimers as determined using a heteropolymeric template-primer in the presence of 0.1 mg/ml BSA and 0.01% IGEPAL to prevent nonspecific binding of dendrimers to the RT. Data represent the mean of four independent assays. Error bars denote standard error of the mean. (B) Time of addition study performed in the TZM-bl indicator cell line with 20 and 100 µg/ml of SPL7115 and SPL7013. AMD3100 and nevirapine were included as controls for an entry and HIV-1 RT inhibitor, respectively. Data is a representative assay from one of three independent assays. (C) Inhibition of R5 and X4 gp120 mediated cell-to-cell fusion by SPL7013 and (D) SPL7115. Data from fusion assays are the average from at least three independent assays. Error bars denote standard error of the mean.

Figure 5. Electrostatic surface views of dendrimer and gp120 models.

(A) Models of the G2 (SPL7115) and G4 (SPL7013) dendrimers taken from the last frame of a molecular dynamics simulation . Two views are shown (top and bottom) rotated by 180° around the long axis of the dendrimer. (B) Two views (rotated by 180°) of the homology model of the DRNL wild-type gp120. CD4 binding site is denoted by a yellow outline and conserved residues K121, R325, R417, K419 and K430 in the CD4i domain are indicated. (C) Side-view of the DRNL gp120 trimer with the V3 loops pointing towards the target cell membrane. The space between the gp120 monomers would normally be occupied by gp41. (D) End-on view of the DRNL gp120 trimer with the V3 loops facing the viewer. Predicted position of gp41 monomers are marked by an asterisk. In all views, the solvent-accessible surfaces are mapped with regions of negative (red) and positive (blue) electrostatic potential (charge).