Automated image-based assay for evaluation of HIV neutralization and cell-to-cell fusion inhibition
© Sheik-Khalil et al.; licensee BioMed Central Ltd. 2014
Received: 25 April 2014
Accepted: 18 August 2014
Published: 30 August 2014
Standardized techniques to detect HIV-neutralizing antibody responses are of great importance in the search for an HIV vaccine.
Here, we present a high-throughput, high-content automated plaque reduction (APR) assay based on automated microscopy and image analysis that allows evaluation of neutralization and inhibition of cell-cell fusion within the same assay. Neutralization of virus particles is measured as a reduction in the number of fluorescent plaques, and inhibition of cell-cell fusion as a reduction in plaque area.
We found neutralization strength to be a significant factor in the ability of virus to form syncytia. Further, we introduce the inhibitory concentration of plaque area reduction (ICpar) as an additional measure of antiviral activity, i.e. fusion inhibition.
We present an automated image based high-throughput, high-content HIV plaque reduction assay. This allows, for the first time, simultaneous evaluation of neutralization and inhibition of cell-cell fusion within the same assay, by quantifying the reduction in number of plaques and mean plaque area, respectively. Inhibition of cell-to-cell fusion requires higher quantities of inhibitory reagent than inhibition of virus neutralization.
KeywordsAutomated plaque reduction assay (APR assay) Fluorescence HIV Neutralization Fusion inhibition
The assessment of virus specific antibodies plays a central role in human immunodeficiency virus type 1 (HIV-1) vaccine development . Thus, evaluation of HIV/AIDS preventive vaccines requires the development of methods to assess different properties, including neutralizing capacity, of protective antibody responses. However, no single assay has so far been reported to detect such protective antibodies. Instead, a wide range of HIV-1 neutralization assays and variants thereof have been developed and described in the literature [2, 3]. These assays are based on different technologies but they all rely on the principle to measure reduction of virus infectivity in susceptible cells in the presence of inhibitory reagents. Likewise, the plaque reduction assay (PR) is based on virus infectivity and uses human cell lines (U87.CD4 or GHOST(3)) engineered to express HIV receptors [4, 5].
In parallel, major advances in high-throughput fluorescence microscopy and automated, high-content image analysis tools have paved the way for systematic and quantitative study of biological systems [6–10]. Fluorescence-based imaging assays have been applied to large-scale analysis to solve biological problems.
Here, we describe an automated PR assay, including sample preparation, automated image acquisition, and a computational image analysis pipeline using open-source software in order to convert the PR assay for HIV neutralization into a high-throughput, high-content assay. In addition to quantifying neutralization by reduction of plaque number, our high-content assay measures plaque area, permitting studies of cell-cell fusion inhibition. We compare this assay with manual readout and with other neutralization assays using a reference panel of inhibitory reagents, i.e. plasma and monoclonal antibodies. We demonstrate that by the use of image analysis the APR assay is converted to a high-throughput and high-content assay, where plaque area is a measure of cell-cell fusion.
Plaque Reduction (PR) assay
Infection assays were done by using HIV-1 isolates (of different subtypes and coreceptor use) and polyclonal inhibitory reagents (eight HIV-positive and one HIV-negative plasma), obtained from the Centre for AIDS Reagents (CFAR) NIBSC, UK, as previously described . We tested virus neutralization in GHOST(3)-CCR5 and -CXCR4 cell lines stably transfected with CD4, chemokine receptors and the Tat-inducible green fluorescence protein (GFP) [4, 11]. Three days following exposure of cells to the plasma-virus mixtures, plaque-forming units (PFU) under the fluorescent microscope are counted to calculate neutralization as percentage of plaque reduction (PR) in the sample containing inhibitory reagent: [1 - (PFU with inhibitory reagent/PFU without inhibitory reagent)] × 100 . For the manual readout the plaques were counted by eye in separate experiments.
Automated Plaque Reduction assay (APR assay)
Details of plaque determination
As is the case for many biological applications, the most challenging bottleneck in configuring an image analysis pipeline is the segmentation, defined as the identification and partitioning of the individual plaques in the image. The IdentifyPrimaryObjects module in the APR image analysis pipeline was adjusted to read foreground/background to match manual reading. The results of the automatic readout and what would be the corresponding manual readout within the same experiment are thus closely similar, if not identical.
Here we describe several key settings contained in the IdentifyPrimaryObjects module in the plaque identification pipeline; these are also described in the module notes within the CellProfiler user interface:
Choice of automatic thresholding method: In this context, thresholding refers to the use of intensity values to distinguish the image foreground (i.e., the GFP-expressing plaques) from the background. A number of automatic thresholding methods are provided for use in the module; the thresholding method chosen must not only reliably identify bright distinct plaques but also the smaller, dim single cells in order to be sufficiently sensitive. For the PR assay, the “Robust Background” thresholding method was used, which assumes a Gaussian intensity distribution after trimming the brightest and dimmest 5% of pixel intensities; the threshold is then calculated as the mean of this distribution plus 2 standard deviations.
Correction of the threshold value: If the automatically determined threshold is consistently too high or too low in all images, it can be further refined by adjusting a module setting which multiplies the threshold by a constant value (“threshold correction factor”). Our automatic readout was performed after fine-tuning the threshold correction factor (TCF) on plaque detection in GHOST(3)- CCR5/CXCR4 cells for HIV-1-infected and uninfected cases (as explained in the legend for Figure 4). This selection was made manually by careful review of many plates and pictures. and verified through comparison of uninfected and HIV-1 infected GHOST(3) cultures. Care is required in correction factor selection since an excessively high value will identify only the brightest plaques (i.e., yield false negatives) and underestimate the plaque area in GFP-positive images, whereas a value that is too low will detect false positives in GFP-negative images (Figure 4). A threshold correction factor of 1.6 was found to be optimal for this assay.
Selection of the lower threshold limit: In instances where the image is GFP-negative (i.e., no plaques), the automatic threshold value may be too low and therefore detects false positives. A lower threshold bound can be set as a precautionary measure by empirically estimating the GFP-negative signal across several images. For the PR assay, the lower threshold bound was set to 0.013.
Object exclusion based on plaque size: The upper and lower bounds for the typical diameter of a GFP-identified plaque can be adjusted to exclude spurious foreground regions, thereby precluding false positives. This size range was set to 4–150 pixels after calculating the mean diameter of a number of individual GHOST(3) cells in brightfield mode (data not shown).
PBMC- and pseudovirus- based neutralization assays
The detailed methodologies of the peripheral blood mononuclear cell (PBMC)- and Env (gp160) pseudotyped virus (PSV)-based neutralization assays are available on the EUROPRISE website (http://www.europrise.org/neutnet_sops.html). In brief, seven laboratories performed the PBMC-based assay, where virus isolates were used with PBMC (isolated from buffy coats of HIV-negative blood donors) as target cells. The PSV-based assay was performed by six different laboratories as a single cycle assay and by one laboratory as a multiple cycle infection assay with engineered cell lines as target cells.
The non-parametric Spearman rank statistical analysis was used to calculate correlations between findings of the automated and manual plaque reduction assays and neutralization to change in plaque area. For comparison of the three different neutralization assays in relation to virus neutralization sensitivity and plasma neutralization capacity, as well as the mean plaque area of different uninhibited HIV-1 isolates the non-parametric Kruskal Wallis test, with Dunn’s post test, was used. Normalized mean plaque area for virus-plasma combinations, at IC50, IC75 and IC90, was evaluated according to Friedman’s test and Dunn’s Post Test.
Comparison of manual and APR readout
Comparison of APR assay with other HIV-1 neutralization assays
High-content readout by the APR assay allows analysis of cell-cell fusion inhibition
For quantitative analysis of plaque reduction we introduced a new nomenclature, ICpar (inhibitory concentration for plaque area reduction). ICpar indicates the concentration at which the antibody causes a decrease in plaque area, e.g., ICpar50 is a 50% reduction in plaque area. IC90 and ICpar50 values were similar for most virus-antibody combinations (Figure 8 and Additional file 1: Table S1) indicating that inhibition of cell-to-cell fusion requires higher quantities of inhibitory reagent than inhibition of virus cell entry and replication.
Virus-induced cell-cell fusion characteristics based on plaque area readout
Here we describe a novel image-based methodology for assaying HIV neutralization in an automated high-content manner. Our high-content assay allows, for the first time, evaluation of both HIV neutralization and inhibition of cell-cell fusion within the same assay. This is achieved by quantifying the reduction in number of plaques and mean plaque area, respectively. The image analysis pipeline for the APR assay in GHOST(3) cells is based on the freely available CellProfiler software  and has now been optimized and adapted for automated reading. Automation offers many advantages: it is higher throughput, standardized, and inexpensive compared both to the manual PR assay and the classic virus infection assay performed in peripheral blood mononuclear cells (PBMC). Automated image analysis increases the objectivity of the plaque counting demonstrated by the very low intra-well variation among triplicate wells (data not shown) as well as low background in the negative controls.
A striking feature of the image-based approach is that it allows extraction of additional phenotypic features of viruses, including cell-cell fusion capacity. We observed differences between the eight HIV-1 isolates used in the present study. In fact, based on statistically significant differences in mean plaque area, viruses could be divided into two groups yielding large or small plaques. Large plaques may result from fusion of a higher number of cells than small plaques. It has been suggested that virus ability to mediate cell-cell fusion, i.e. syncytia, depends on the strength of the interaction of viral envelope protein with CD4 and coreceptors [14, 15]. In addition, Env clustering and mobility in the cell membrane may influence the outcome of syncytia formation .
Our results show that a higher concentration of inhibitory reagent was needed to get a reduction in plaque area than that needed for neutralization. This was previously proposed by Yee et al.  and given the explanation that the interaction of the envelope glycoproteins gp120/gp41 (Env) with cell membrane CD4 may be different during cell-cell fusion than during virus-cell membrane fusion . It is plausible that the molecular organization and surface density of gp120/gp41 may be different in Env-expressing cells and HIV-1 virions. Indeed, Purtscher et al.  proposed, similarly to what we show, that elevated concentrations of antibody are necessary to inhibit fusion between infected and target cells, and suggested this to be due to higher levels of gp120/gp41 expressed on the surface of infected cells as compared to the viral envelope. Interestingly, it was recently shown, in an HIV-1-infected humanized mouse model, that infected T cells form contacts with uninfected cells, and when the frequency of these contacts was reduced, plasma viremia was significantly decreased, strongly suggesting a role for cell-cell contacts in systemic viral spread . Thus, it is conceivable that cell-cell fusion inhibition capacity of an antibody has a role in mitigating pathogenesis and may also be required for an effective HIV vaccine.
Antibody assessment plays a central role in HIV-1 vaccine development where, analogous to other virus infections, virus neutralization is considered particularly important. Accordingly, international comparisons of a wide range of HIV-1 neutralization assays have been performed and showed that no assay alone detects neutralization over the entire spectrum of virus-reagent combinations [2, 21]. The APR assay has also been standardized and compared with other HIV neutralization assays within the framework of an international collaboration, Neutnet [2, 3]. As with the other assays, the sensitivity of our assay was dependent on both the neutralizing reagent and the virus. Thus the APR assay can be considered an information-rich alternative to the PBMC and PSV assays.
Here we report on a novel image-based automated plaque-reduction assay where both HIV neutralization and inhibition of cell-cell fusion can be analyzed, for the first time within the boundaries of the same assay format. This high-content assay may be used as a tool for evaluation of multiple antibody effector functions, by ways of quantifying the reduction in number of plaques and mean plaque area, respectively. The image analysis platform described herein can be further developed with the potential to study additional features of antibody-virus-cell interactions, which may prove important in antibody-based HIV vaccine design.
MAB and AEC were supported by NIH R01 GM089652.
Grants were received from the Swedish Research Council, the Swedish International Development Cooperation Agency/Department for Research Cooperation (SIDA/SAREC), the Crafoord Foundation and the European Community: EUROPRISE-Network of Excellence grant number LSHT-CT-2006-037611 and NGIN grant number 201433.
- McCoy LE, Weiss RA: Neutralizing antibodies to HIV-1 induced by immunization. J Exp Med. 2013, 210 (2): 209-223. 10.1084/jem.20121827.PubMed CentralView ArticlePubMed
- Fenyo EM, Heath A, Dispinseri S, Holmes H, Lusso P, Zolla-Pazner S, Donners H, Heyndrickx L, Alcami J, Bongertz V, Jassoy C, Malnati M, Montefiori D, Moog C, Morris L, Osmanov S, Polonis V, Sattentau Q, Schuitemaker H, Sutthent R, Wrin T, Scarlatti G: International network for comparison of HIV neutralization assays: the NeutNet report. PLoS One. 2009, 4 (2): e4505-10.1371/journal.pone.0004505.PubMed CentralView ArticlePubMed
- Heyndrickx L, Heath A, Sheik-Khalil E, Alcami J, Bongertz V, Jansson M, Malnati M, Montefiori D, Moog C, Morris L, Osmanov S, Polonis V, Ramaswamy M, Sattentau Q, Tolazzi M, Schuitemaker H, Willems B, Wrin T, Fenyö EM, Scarlatti G: International network for comparison of HIV neutralization assays: the NeutNet report II. PLoS One. 2012, 7 (5): e36438-10.1371/journal.pone.0036438.PubMed CentralView ArticlePubMed
- Lauren A, Thorstensson R, Fenyo EM: Comparative studies on mucosal and intravenous transmission of simian immunodeficiency virus (SIVsm): the kinetics of evolution to neutralization resistance are related to progression rate of disease. J Gen Virol. 2006, 87 (Pt 3): 595-606.View ArticlePubMed
- Shi Y, Albert J, Francis G, Holmes H, Fenyo EM: A new cell line-based neutralization assay for primary HIV type 1 isolates. AIDS Res Hum Retroviruses. 2002, 18 (13): 957-967. 10.1089/088922202760265623.View ArticlePubMed
- Ljosa V, Carpenter AE: Introduction to the quantitative analysis of two-dimensional fluorescence microscopy images for cell-based screening. PLoS Comput Biol. 2009, 5 (12): e1000603-10.1371/journal.pcbi.1000603.PubMed CentralView ArticlePubMed
- Pegoraro G, Bavari S, Panchal RG: Shedding light on filovirus infection with high-content imaging. Viruses. 2012, 4 (8): 1354-1371.PubMed CentralView ArticlePubMed
- Pepperkok R, Ellenberg J: High-throughput fluorescence microscopy for systems biology. Nat Rev Mol Cell Biol. 2006, 7 (9): 690-696. 10.1038/nrm1979.View ArticlePubMed
- Starkuviene V, Pepperkok R: The potential of high-content high-throughput microscopy in drug discovery. Br J Pharmacol. 2007, 152 (1): 62-71. 10.1038/sj.bjp.0707346.PubMed CentralView ArticlePubMed
- Rose PP, Hanna SL, Spiridigliozzi A, Wannissorn N, Beiting DP, Ross SR, Hardy RW, Bambina SA, Heise MT, Cherry S: Natural resistance-associated macrophage protein is a cellular receptor for sindbis virus in both insect and mammalian hosts. Cell Host Microbe. 2011, 10 (2): 97-104. 10.1016/j.chom.2011.06.009.PubMed CentralView ArticlePubMed
- Morner A, Bjorndal A, Albert J, Kewalramani VN, Littman DR, Inoue R, Thorstensson R, Fenyo EM, Bjorling E: Primary human immunodeficiency virus type 2 (HIV-2) isolates, like HIV-1 isolates, frequently use CCR5 but show promiscuity in coreceptor usage. J Virol. 1999, 73 (3): 2343-2349.PubMed CentralPubMed
- Carpenter AE, Jones TR, Lamprecht MR, Clarke C, Kang IH, Friman O, Guertin DA, Chang JH, Lindquist RA, Moffat J, Golland P, Sabatini DM: CellProfiler: image analysis software for identifying and quantifying cell phenotypes. Genome Biol. 2006, 7 (10): R100-10.1186/gb-2006-7-10-r100.PubMed CentralView ArticlePubMed
- Kamentsky L, Jones TR, Fraser A, Bray MA, Logan DJ, Madden KL, Ljosa V, Rueden C, Eliceiri KW, Carpenter AE: Improved structure, function and compatibility for CellProfiler: modular high-throughput image analysis software. Bioinformatics. 2011, 27 (8): 1179-1180. 10.1093/bioinformatics/btr095.PubMed CentralView ArticlePubMed
- Watkins BA, Crowley R, Davis AE, Louie AT, Reitz MS: Syncytium formation induced by human immunodeficiency virus type 1 isolates correlates with affinity for CD4. J Gen Virol. 1997, 78 (Pt 10): 2513-2522.View ArticlePubMed
- Shiva Naresh Mulampaka NMD: Estimating the threshold surface density of Gp120-CCR5 complexes necessary for HIV-1 envelope-mediated cell-cell fusion. PLoS One. 2011, 6 (5): e19941-10.1371/journal.pone.0019941.PubMed CentralView ArticlePubMed
- Roy NH CJ, Lambelé M, Thali M: Clustering and mobility of HIV-1 Env at viral assembly predict its propensity to induce cell-cell fusion. J Virol. 2013, 87 (13): 7516-7525. 10.1128/JVI.00790-13.PubMed CentralView ArticlePubMed
- Yee M, Konopka K, Balzarini J, Duzgunes N: Inhibition of HIV-1 Env-mediated cell-cell fusion by lectins, peptide T-20, and neutralizing antibodies. Open Virol J. 2011, 5: 44-51. 10.2174/1874357901105010044.PubMed CentralView ArticlePubMed
- Konopka K, Pretzer E, Celada F, Duzgunes N: A monoclonal antibody to the gp120-CD4 complex has differential effect on HIV-induced syncytium formation and viral infectivity. J Gen Virol. 1995, 76 (Pt 3): 669-679.View ArticlePubMed
- Purtscher M, Trkola A, Gruber G, Buchacher A, Predl R, Steindl F, Tauer C, Berger R, Barrett N, Jungbauer A, Katinger H: A broadly neutralizing human monoclonal antibody against gp41 of human immunodeficiency virus type 1. AIDS Res Hum Retroviruses. 1994, 10 (12): 1651-1658. 10.1089/aid.1994.10.1651.View ArticlePubMed
- Murooka TT, Deruaz M, Marangoni F, Vrbanac VD, Seung E, von Andrian UH, Tager AM, Luster AD, Mempel TR: HIV-infected T cells are migratory vehicles for viral dissemination. Nature. 2012, 490 (7419): 283-287. 10.1038/nature11398.PubMed CentralView ArticlePubMed
- Polonis VR, Brown BK, Rosa Borges A, Zolla-Pazner S, Dimitrov DS, Zhang MY, Barnett SW, Ruprecht RM, Scarlatti G, Fenyo EM, Montefiori DC, McCutchan FE, Michael NL: Recent advances in the characterization of HIV-1 neutralization assays for standardized evaluation of the antibody response to infection and vaccination. Virology. 2008, 375 (2): 315-320. 10.1016/j.virol.2008.02.007.View ArticlePubMed
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2334/14/472/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.