A High Throughput Protein Microarray Approach to Classify HIV Monoclonal Antibodies and Variant Antigens

Abstract

In recent years, high throughput discovery of human recombinant monoclonal antibodies (mAbs) has been applied to greatly advance our understanding of the specificity, and functional activity of antibodies against HIV. Thousands of antibodies have been generated and screened in functional neutralization assays, and antibodies associated with cross-strain neutralization and passive protection in primates, have been identified. To facilitate this type of discovery, a high throughput-screening tool is needed to accurately classify mAbs, and their antigen targets. In this study, we analyzed and evaluated a prototype microarray chip comprised of the HIV-1 recombinant proteins gp140, gp120, gp41, and several membrane proximal external region peptides. The protein microarray analysis of 11 HIV-1 envelope-specific mAbs revealed diverse binding affinities and specificities across clades. Half maximal effective concentrations, generated by our chip analysis, correlated significantly (P<0.0001) with concentrations from ELISA binding measurements. Polyclonal immune responses in plasma samples from HIV-1 infected subjects exhibited different binding patterns, and reactivity against printed proteins. Examining the totality of the specificity of the humoral response in this way reveals the exquisite diversity, and specificity of the humoral response to HIV.

Fig 1. HIV-1 protein microarray chip analysis.

HIV-1 specific antibodies were tested against a panel of recombinant multi-clade gp140 and gp120 proteins as well as gp41 and MPER analogs. Proteins were printed at a concentration of 0.1 mg/mL, which corresponds to 0.1 ng per spot. Antibodies were probed at a concentration of 10 μg/mL. The cutoff for positive binding was a signal intensity of 0.4 x 10³. Signal intensities were color-coded using green (>0.4–5.0 x 10³) for weak, yellow (>0.5–1.5 x 10⁴) for intermediate, orange (>1.5–3.0 x 10⁴) and red (>3.0 x 10⁴) for strong interaction. Non-specific binding was indicated as white (<0.4 x 10³) boxes. Data are representative of at least two independent experimental runs.

Fig 2. Binding breadth and median signal intensities of HIV-1 specific antibodies.

(A) The binding breadth is indicated by the percentage of bound proteins and color-coded as follows white (<0%), green (>0–25%), yellow (>25–50), >50–75% (orange), and >75% (red). The sample size was N = 5 for the gp140 cluster, N = 4 for the gp120 cluster, and N = 5 for the gp41/MPER cluster, as the peptide 08023 was excluded. (B) Median signal intensities of bound antibodies were calculated based on signal intensities derived from Fig 1, and plotted into a signal heat map with the same color code as in Fig 1, using green (>0.4–5.0 x 10³) for weak, yellow (>0.5–1.5 x 10⁴) for intermediate, orange (>1.5–3.0 x 10⁴) and red (>3.0 x 10⁴) for strong interaction. Non-specific binding was indicated as white (<0.4 x 10³) boxes. The cutoff for positive binding was a signal intensity of 0.4 x 10³.

Fig 3. Cluster analysis. Arrays were probed with mAbs diluted in a series of 8 half log concentrations as described in the text.

The heat map displays intensity with orange (>1.5–3.0 x 10⁴) and red (>3.0 x 10⁴) showing the strongest, green (<5.0 x 10³) the weakest, and yellow (>0.5–1.5 x 10⁴) intermediate according to Fig 1. MeV v4.6 (TM4Microarray Software Suite, www.tm4.org) was used to perform the hierarchical clustering analysis using average linkage clustering. Samples are clustered according to response against protein antigen spots (1.0 ng per spot) on the array.

Fig 4. Calculation and evaluation of EC₅₀ data.

EC₅₀ values (μg/mL) were generated and calculated (PRISM GraphPad 5.0) by antibody binding curves in ELISA (A) and in the protein microarray chip assay (B). Values were color-coded using red for low, yellow for medium, and green for high EC₅₀ values. No binding is indicated as white boxes. EC₅₀ values of both analyses are representative of two independent experiments.

Fig 5. Microarray analysis of HIV-1 infected patient samples.

The microarray chip was probed with HIV-1 patient plasma at a dilution factor of 1:100. Proteins were printed at a concentration of 0.01 mg/mL, which corresponds to 0,01 ng per spot. HIVIG and IVIG were included as a positive control and negative control, respectively. Intensities were color coded using green (>0.05–5.0 x 10³) for weak, yellow (>0.5–1.5 x 10⁴) for intermediate, orange (>1.5–3.0 x 10⁴) and red (>3.0 x 10⁴) for strong interaction. Non-specific binding was indicated as white (<0.5 x 10²) boxes. Data are representative of at least two independent experimental runs.

Fig 6. Patient plasma sample evaluation.

Plasma samples of HIV-1 infected subjects were analyzed for neutralization capacity against Tier 1 isolate HIV-1_SF162 next to median binding intensities against HIV-1_SF162 gp140, HIV-1_BaL gp120 and cross clade signal intensity within the gp140 cluster (A). Further, correlations were calculated between IC₅₀ values and 1) median HIV-1_SF162 gp140 signal intensities (B), 2) median HIV-1_BaL gp120 signal intensities (C), and 3) cross clade gp140 signal intensities (D).