Antiviral activity of a single-domain antibody immunotoxin binding to glycoprotein D of herpes simplex virus 2

Abstract

Despite years of research dedicated to preventing the sexual transmission of herpes simplex virus 2 (HSV-2), there is still no protective vaccine or microbicide against one of the most common sexually transmitted infections in the world. Using a phage display library constructed from a llama immunized with recombinant HSV-2 glycoprotein D, we identified a single-domain antibody VHH, R33, which binds to the viral surface glycoprotein D. Although R33 does not demonstrate any HSV-2 neutralization activity in vitro, when expressed with the cytotoxic domain of exotoxin A, the resulting immunotoxin (R33ExoA) specifically and potently kills HSV-2-infected cells, with a 50% neutralizing dilution (IC50) of 6.7 nM. We propose that R33ExoA could be used clinically to prevent transmission of HSV-2 through killing of virus-producing epithelial cells during virus reactivation. R33 could also potentially be used to deliver other cytotoxic effectors to HSV-2-infected cells.

FIG 1

Production and purification of gD2. (A) DNA encoding gD2 was amplified by PCR from the HSV-2 strain186 genome. (B and C) Purified gD2 from P. pastoris was separated by SDS-PAGE and stained with Coomassie brilliant blue (B) or detected with a polyclonal anti-gD2 antibody (R45) by Western blotting (C). The left lanes of all three gels are molecular markers, while right lanes show a single amplified PCR product (A), purified gD2 (B), or immunoreactive gD2 (C). (D) In an ELISA, gD2 was recognized by a panel of anti-gD2 antibodies: R45 (1:5,000), HSV8 (1:5,000), DL6 (1:1,000), and anti-His (1:1,000). As controls, wells were coated with gD2, and only HRP-conjugated secondary antibody (anti-rabbit for R45, anti-human for HSV8, and anti-mouse for DL6 and anti-His) was added. Error bars represent maximum and minimum values.

FIG 2

Reactivity of sera from llamas immunized with gD. A 1:100 dilution of sera collected before the initiation of immunization (naive) and after the second through fourth immunizations (Im. #2 to #4) was serially diluted 10-fold and tested in an ELISA for reactivity to gD2. The llama sera from Immunizations 2 to 4 demonstrate higher reactivity to gD2 than naive sera, demonstrating that the llama mounted an antibody response against the gD2 immunizations. Data represent the averages from three wells, and error bars are standard deviations.

FIG 3

Purification of VHH and immunotoxins and comparison of sequences from VHH of different specificities. (A and B) Representative gels that demonstrated the size and purity of purified R33 and bvR33(A) or R33ExoA and P10ExoA (B). (C) R33 and P10 amino acid sequences. VHH genes, originally amplified from the variable region of heavy-chain-only antibodies, were sequenced from VHH-phage clones and aligned to determine unique VHH sequences identified from the gD2 biopanning process. R33 is able to bind gD2, while P10 is a negative-control VHH that does not bind gD2.

FIG 4

Binding of R33 and bvR33 VHH to surface-expressed gD2 on z4/6 cells. z4/6 cells (surface expression of gD2) were stained with various VHH (R33, P10, and bvR33) and detected using fluorescence-activated cell sorting (FACS) by a FITC-conjugated anti-His secondary antibody, demonstrating that R33 and bvR33 but not P10 can bind to gD2. DL6 was used as a positive control to verify that gD2 was expressed, and a secondary antibody control (anti-His) with no VHH or primary antibody was also used as a negative control.

FIG 5

HSV-2 neutralization assay. Virus was incubated with dilutions of R33 and HSV8 (known neutralizing antibody) (A) or R33, bvR33, and P10 (B) and then plated on Vero cells to assay for HSV-2 neutralizing activity. While HSV8 is capable of neutralizing HSV-2 (IC₅₀ of 1.5 nM), R33 and bvR33 do not differ from P10 in terms of inhibiting viral infection. Each dilution was assayed in duplicate, and error bars represent maximum and minimum plaque numbers. These graphs are representative of at least three independent experiments, and results are expressed as percent inhibition compared to plaque numbers from untreated virus. Statistical significance compared to results for untreated virus was calculated by ANOVA and is indicated by asterisks (P < 0.05).

FIG 6

Specific toxicity of R33ExoA for gD2-expressing cells. Dilutions of immunotoxins were added to Vero cells (A) or z4/6 cells (B), and their cytotoxicity was measured using an MTS assay. While R33ExoA demonstrates cytotoxic activity against gD2-expressing cell lines (IC₅₀ = 0.5 nM) (95% CI, 0.1810 to 1.403), P10ExoA does not. Neither immunotoxin demonstrates activity against Vero cells, which do not express gD2. Dilutions of each protein were added to wells in triplicate, and error bars represent standard deviations.

FIG 7

(A) R33 and R33ExoA bind to recombinant gD2. An ELISA in which wells were coated with the indicated VHH or VHH immunotoxin, and gD2 was added to assay for their ability to bind gD2, was performed. While R33 and R33ExoA are able to bind to recombinant gD2, neither P10 nor P10ExoA demonstrates any reactivity. The graph is representative of three separate experiments. Each dilution was assayed in duplicate, and error bars represent maximum and minimum values. (B) R33ExoA antiviral activity in infectious center assay (ICA). An ICA with dilutions of R33ExoA and P10Exo shows that only R33ExoA has antiviral activity, with an IC₅₀ of 6.7 nM. This is a representative graph from six independent experiments. Error bars represent standard errors of the means.