HTS assays based on recombinant reporter parasites

An increasing knowledge of T. cruzi molecular biology and the development of plasmids and efficient parasite transfection protocols [37] permitted the construction of reliable and robust transgenic T. cruzi parasites of multiple genetic backgrounds [26], [28], [37], [38]. The availability of these microorganisms transformed the previously time-consuming, labor-intensive, observer-biased visual microscopic counting, into faster, unbiased colorimetric, fluorometric, or luminescence-based measurements, which have made feasible the screening of large compound libraries. Transgenic T. cruzi parasites that express bacterial β-galactosidase reporter enzyme [26], firefly luciferase [28], [38], or the tandem tomato fluorescent protein [28] have been made available to the scientific community. As part of their development as tools for anti–T. cruzi inhibitory assays, these genetically engineered microorganisms were shown to biologically perform like their wild-type counterparts, in terms of growth, cycle stage transitions, mice infection, and in vitro sensitivity to known drugs [26], [28]. For the recombinant parasites carrying either the β-galactosidase reporter gene or the firefly luciferase gene, a substrate addition step is required and a single end-point measurement of the reporter enzyme activity can be obtained. To detect β-galactosidase activity, colorimetric and luminescent substrates have been used (respectively, chlorophenol-red-β-D-galactopyranoside, Sigma-Aldrich [26], and GalScreen, Life Technologies [39]). It must be noted that colored compounds, which are frequently found in chemical libraries, might interfere with the readout when using colorimetric reporters [26]. This issue can be addressed by using luminescent or fluorescent-based readouts as in the first HTS campaign performed against T. cruzi at the Broad Institute [39] and in the more recent HTS made at GSK [40]. Another reason for preferring luminescence and/or fluorescence detection methodologies is their higher sensitivity, which allows for greater miniaturization of assays. An important issue to be considered in the use of transgenic parasites in the screening of chemical libraries is that reporter activity inhibitors (i.e., chemical compounds that inhibit luciferase or β-galactosidase) are selected as hits. Thus, secondary assays, such as image-based assays that rely on non-recombinant strains, are needed to identify these false hits.

Firefly luciferase–expressing parasites require the addition of luciferin before performing the readout [28], [38]. In contrast, the important advantage offered by the self-fluorescent recombinant parasite strains developed in Prof. Tarleton's lab resides on their substrate-independent detection, which allows a continuous activity measurement [28]. This property appears particularly important to investigate parasite growth kinetics, which could provide valuable insights of compounds' mechanisms of action and, for example, aid to prioritize those with high parasite lytic activity. However, the higher sensitivity of detection in vivo of methods such as PCR of specific tissues compared to luminescent or fluorescent parasites would make these methods more useful in the detection of small populations of parasites, for example, in chronically infected animals.

Recombinant T. cruzi parasites have already been used in a series of in vitro assays designed to detect drug susceptibility mainly of the intracellular amastigote life-stage [26]–[28], the parasite replicative form in the mammalian host and the preferred parasitic target stage. When screening a molecular library against intracellular amastigotes, parallel host cell toxicity assays on the retrieved hits or “active” compounds are necessary to determine whether the activity is specifically antiparasitic or either totally or partially due to a disruption of host-cell biology [27], [41], [42]. The selectivity index is then calculated using the relative IC₅₀ values of the host cell and the parasite.

The most important impact of the developed in vitro assays based on transgenic parasites has been their amenability to HTS [39], [40]. In terms of impact, the adaptation of the T. cruzi β-galactosidase–based assay [26] to HTS standards established a milestone in the anti-Chagas early drug discovery process. After a protocol improvement to reduce the assay time, but still at its original 96-well plate format, the assay showed its reliability in the screening of a 2,000 compounds chemical library [27]. The β-galactosidase reporter assay was developed into a 384-well format with a luminescent readout allowing higher throughput for the HTS campaign at the Broad Institute, where screening of 303,224 compounds (the NIH library) yielded 4,394 hits. After the confirmation of activity and determination of efficacy and cytotoxicity of the hits, 3,005 compounds with an IC50 <10 µM and>10-fold selectivity against the parasite versus the host cells were selected [39]. All information, including the chemical structures of the hits, is publicly available at Pubchem (AID 1885). Further analysis of the hits selected by this HTS revealed that a high percentage of them were not reproducible in a similar assay. It is likely that the reason for such discrepancies resided in the quality of the chemical compounds used for the original HTS [41]. Another HTS of a diversity-oriented synthesis chemical library of 77,312 compounds has been performed [42]. The most attractive confirmed hits in terms of potency, selectivity, and predicted drug-like physicochemical properties between these two HTSs were picked up for further development and the data obtained from them have served as the basis for different initiatives (Figure 1) [43]–[46]. The access to high-quality large chemical libraries will be of fundamental importance for the advance of drug discovery against neglected diseases, as it will be to get the resources and know-how to mine and test them. In this perspective, collaboration between pharmaceutical companies and academic laboratories becomes a key aspect for the successful development of future drugs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Anti–Trypanosoma cruzi compounds.

Chemical structures of anti–T. cruzi compounds currently in clinical use (*) and currently or recently in clinical trials (•), and of the most promising compounds retrieved out of the Broad Institute HTS campaign [44] and the screening of a smaller, diversity-oriented chemical library [39], [43] (#).

https://doi.org/10.1371/journal.pntd.0003259.g001

Image-based HTS assays

High content screening (HCS) allows the automated acquisition and analysis of multiple cellular features simultaneously [30], [46]. HCS is a powerful screening tool, the use of which has expanded in the last years because of high assay throughput, automatic image acquisition and analysis provided by user-friendly software interfaces and the possibility of robotization [30], [46]. The availability of this technology has already permitted the development of image-based assays to identify active compounds against Leishmania donovani [34], [47]–[49]. Similarly to HTS based on recombinant reporter parasites, automatic quantification is not biased by the observer visual parasite counting, as it is performed by an image analysis script specifically developed for this purpose [31], [32], which once set up has been shown to be extremely useful, providing very accurate readings and saving valuable time [50]. In the drug development process against T. cruzi, this methodology holds a clear advantage compared to reporter-based assays as it offers more versatility, which may allow the use, upon assay adaptation, of any parasite strain that is adapted for infection in vitro (including difficult-to-engineer strains and relevant clinical specimens). Additionally, this assay provides the information on host cell cytotoxicity of each compound in a single experiment, delivering immediate results on the selectivity index. An image-based assay developed by Engel and coworkers that has already been used to screen a FDA-approved drugs library (>900 compounds) where 55 hits were identified [31], and at least two other image-based assays have been referenced in perspective articles [33], [36]. More recently, scientists from Institute Pasteur in Korea have developed an algorithm to aid in the image analysis process of a T. cruzi infection drug inhibitory assay [32].

There are yet some disadvantages related to HCS that need to be improved, like the storage and handling of the high amount of data generated and the speed of both picture acquisition and computing analysis that can have an impact on the developed assays HTS scale up [46], [51]. A drawback specific to the reported anti–T. cruzi image-based assay is that it does not allow serial readouts, since cells must be fixed and stained at a single end-point. This is similar to β-galactosidase and luciferase reporter-based assays; however, in the imaging assays, the possibility to measure different cell biology features (high-content microscopes offer up to four laser channels and there is a large spectrum of available fluorophores and biological probes) in a single experiment should grant its further development. Besides the identification of new hits, HCS could also provide a major input in the hit-to-lead-optimization process and play a key part in deciphering new mechanisms of action [30], [51]. In the future, fluorescently tagged ligands or specific staining with antibodies could be incorporated in the assays to provide mechanistic information of compounds.

In vivo intermediate throughput screening by image acquisition and shortened protocols

A fundamental step in the drug discovery path is the selection and qualification of hits to be progressed into leads [52]. A requirement for this upgrade is consistent efficacy in animal models [29], [52]. Although various animal infection models have been studied in Chagas disease research, by far the most used has been the mouse model [53]. The mouse is the most preferred in vivo model because it adequately reproduces the main pathogenic features of human T. cruzi infection, there is accumulated knowledge about its genetics and physiology, a large amount of reagents and tools developed to work with it, and the ease of handling and storage of colonies.

Transgenic parasites have also been used with success for in vivo studies. Bio-imaging of mouse models infected with transgenic T. cruzi strains (either luciferase or fluorescent recombinant parasites) allows a rapid assessment of the infectious disease process within the animals [28], [38], [41]. Before the implementation of transgenic parasites and in vivo imaging techniques, animals had to be killed to obtain end-point data on parasite detection and organ dissemination, which was very labor-intensive, presented risk for human infection, and required the use of large animal groups, all contributing to increased costs [53], [54]. Furthermore, the techniques used (PCR, in situ hybridization, and microscopic blood or tissue sections parasitic counting) were cumbersome and presented certain limitations [38]. By means of luminescent or fluorescent transgenic parasites and the appropriate detection equipment, it is now possible to accurately evaluate in vivo the infection process and the drug responses that might inhibit it. In comparison with measurements relying on sacrifice, bio-imaging has the great advantage of providing continuous readouts on anesthetized animals, which has greatly contributed to the understanding of infection dynamics and parasite tropism [55], [56]. Interestingly, T. cruzi chronic infections in mice were focal but moved within the body from week to week. Parasites did not localize to the heart, despite these mice developing miocarditis and cardiac fibrosis. T. cruzi was found in other organs and consistently in the gastrointestinal tract [56].

Another advantage of these experimental infections for drug development is that, since progression of infection can be followed in live animals, the total number of mice required for the experiments can be drastically reduced, therefore following the 3R principle (Replacement, Reduction, and Refinement) [57], and contributing to reducing research costs [38], [54]. With such advantages, in vivo protocols have already been developed [28], [29], [38], [41], [54] and are being used to test in vitro–selected hits with the best drug-like characteristics [28], [41]. Early tagging of hits in the drug development path as in vivo–active, i.e., prior to medicinal chemistry modifications, represents considerable savings in time and funding, which are much needed to further develop the selected compounds into drugs for a disease with serious funding shortages [29]. Some of these assays are available to the community through a fee-for-service, non-profit core facility [58].

Bio-imaging assays normally model the initial acute-phase infection of Chagas disease and would select for fast-acting compounds [54]. Since the parasite detection threshold for T. cruzi in bio-imaging assays normally falls around 1,000 parasites [28], [41], [56], to determine whether a particular treatment, administered either in acute or chronic phase, results in sterile clearance, chemically induced immune suppression [59] should be used after treatment to confirm it [54].

The throughput of some in vivo assays [28] can be considered intermediate, since rapid two-day assays in low number of animals (fewer than 3 per group) can be used for screening hundreds of compounds in short periods of time [29].