Small-Molecule Screening for Genetic Diseases
Sarine Markossian, Kenny K. Ang, Christopher G. Wilson, and Michelle R. Arkin
Small Molecule Discovery Center and Department of Pharmaceutical Chemistry, University of California, San Francisco, California 94143, USA; email: [email protected],
[email protected], [email protected], [email protected]

Annu. Rev. Genom. Hum. Genet. 2018. 19:15.1–15.26
The Annual Review of Genomics and Human Genetics is online at genom.annualreviews.org

https://doi.org/10.1146/annurev-genom-083117-

021452
Copyright c 2018 by Annual Reviews. All rights reserved

Keywords
drug discovery, small molecules, high-throughput screening, HTS, high-content analysis, HCA, monogenic disease, cancer
Abstract
The genetic determinants of many diseases, including monogenic diseases and cancers, have been identified; nevertheless, targeted therapy remains elusive for most. High-throughput screening (HTS) of small molecules, in- cluding high-content analysis (HCA), has been an important technology for the discovery of molecular tools and new therapeutics. HTS can be based on modulation of a known disease target (called reverse chemical genetics) or modulation of a disease-associated mechanism or phenotype (forward chemical genetics). Prominent target-based successes include modulators of transthyretin, used to treat transthyretin amyloidoses, and the BCR-ABL kinase inhibitor Gleevec, used to treat chronic myelogenous leukemia. Phe- notypic screening successes include modulators of cystic fibrosis transmem- brane conductance regulator, splicing correctors for spinal muscular atrophy, and histone deacetylase inhibitors for cancer. Synthetic lethal screening, in which chemotherapeutics are screened for efficacy against specific genetic backgrounds, is a promising approach that merges phenotype and target. In this article, we introduce HTS technology and highlight its contributions to the discovery of drugs and probes for monogenic diseases and cancer.

15.1

1. INTRODUCTION
The human genetic disease landscape is vast and growing; development of new drugs target- ing these diseases have not kept pace. Monogenic disorders, whose disease phenotype results from modification of a single gene, have the clearest genotype–phenotype correlation. It is be- coming clear, though, that even the disease phenotypes inherited in a monogenic fashion can- not be fully explained by mutations in a single locus (32). For example, although mutations in the cystic fibrosis transmembrane conductance regulator gene (CFTR) are strongly corre- lated with the cystic fibrosis (CF) phenotype, the presence of those mutations cannot solely explain the broad heterogeneity of disease severity (75). Thus, drug discovery aimed at genetic diseases, even monogenic diseases, should consider both the primary mutation and modifying determinants.
Cancer also has a genetic component, usually at the level of somatic mutations, chromo- somal fusions, and copy number variations. A subset of cancers also demonstrate a Mendelian pattern of inheritance, including hereditary breast and ovarian cancer (associated with BRCA1 and BRCA2), familial adenomatous polyposis (associated with APC), and Li–Fraumeni syndrome (associated with TP53). In the past 20 years, several successful drugs have targeted these so- matic and familial mutations, either directly or through synthetic lethal interactions. Epigenetic factors and the environment also play an important role; abnormalities in DNA methylation and histone modifications have been linked to a variety of diseases, including cancer (70, 104). In fact, drugs targeting histone-modifying enzymes, such as histone deacetylases (HDACs), as well as DNA methylation, either have received regulatory approval or are in clinical testing (101).
Almost two-thirds of approved medicines are either naturally derived or synthetic small molecules. Even with the recent development of oligonucleotide-based therapies for mono- genic diseases, such as spinal muscular atrophy (SMA) and Duchenne muscular dystrophy (DMD), and the success of antibody-based therapeutics in oncology, small molecules con- tinue to dominate drug discovery pipelines for most human genetic diseases (92, 132). Im- portant features of small-molecule drugs include chemical stability (leading to simple storage and distribution) and bioavailability through multiple routes of administration (enabling tun- able dosing regimens). Small-molecule drugs have been developed to correct a disease phe- notype (as in monogenic diseases), address drug resistance [particularly for infectious dis- eases and cancers (47, 72, 120)], and combine with existing monotherapies (for instance, to address multiple cancer-associated pathways). Historically, most small-molecule drugs inhibit the target activity; more recently, efforts have been directed toward activating aberrant pro- teins and/or restoring normal levels of protein expression. Examples in monogenic diseases include chaperone therapy to restore glucocerebrosidase activity in Gaucher disease (GD) and potentiators and correctors of CFTR for treatment of CF. Small molecules have also played an important role in deciphering mechanisms of disease and normal human biology (59, 117).
In the past 20 years, most new small-molecule medicines trace their discovery to high- throughput screening (HTS). In this review, we describe strategies and technologies used for small-molecule screening, with a focus on human genetic diseases. We first discuss HTS tech- nologies and small-molecule libraries (Section 2). We then highlight examples where HTS led to the development of highly impactful small-molecule drugs for human genetic diseases, in- cluding both monogenic diseases (Section 3) and genetic alterations in cancer (Section 4). Fi- nally, we provide a perspective on the current challenges and future avenues for HTS for genetic diseases.

15.2 Markossian et al.

2. SMALL-MOLECULE SCREENING: STRATEGY AND TECHNOLOGY
2.1. High-Throughput Screening in Drug Discovery
HTS is a platform technology employing common core components—usually supported by robotic automation—that enable rapid testing of 103–106 chemicals in a miniaturized assay format. Because it is amenable to a wide array of biochemical, cell-based, and model-organism-based as- says, HTS can be applied to a broad range of biological problems. It is important to note, though, that screening is just the beginning of a drug discovery and development pipeline. Following hit identification by HTS, significant medicinal chemistry is required to optimize compounds into leads by improving potency, selectivity, in vivo exposure, and safety. Clinical or development can- didates are then evaluated through a set of toxicology, formulation, and manufacturing hurdles before they are ready for testing in humans as an investigational new drug. In the last two decades, the majority of investigational new drug candidates and new chemical entities trace their origins to HTS (83).
HTS is not without problems. First, attrition is frequently high; experimental therapeutics derived from HTS fail in significant numbers at various stages of clinical development, particularly for first-in-class drugs (102). Second, the increased efficiency of HTS has not led to increases in the number of new chemical entities across all disease areas. Third, several articles have emphasized the importance of carefully interpreting HTS results, since assays can be prone to different types of artifacts, and compounds in HTS libraries can act through non-drug-like mechanisms (e.g., owing to chemical instability) (8, 66). Despite these well-founded concerns, HTS remains a proven means to discover new drug leads and enable impactful biomedical research (83). All drug discovery, regardless of the hit-identification approach, relies on the selection of disease-relevant targets and models; failure often arises from our incomplete understanding of the disease. In fact, small- molecule probes allow (in)validation of potential targets and disease models using the very types of molecules that could become drug leads. HTS therefore also plays an important role in the development of tool molecules to dissect disease states.
Once exclusively found in large pharmaceutical companies, HTS is now accessible to academics and small biotech companies. The increase in academic HTS has been driven by an interest in developing both high-quality chemical probes and first-in-class therapeutics for novel targets and understudied diseases. The popularization of HTS is at least partially due to the development of the National Institutes of Health (NIH)–funded Molecular Libraries Probe Production Centers Network (117), the lowered costs for automated equipment, the availability of chemical libraries, and the hiring of pharmaceutically trained scientists in academia. Like our Small Molecule Dis- covery Center at the University of California, San Francisco (6), well over 100 laboratories have emerged in the past decade to bridge the gap between biomedical research and the development of therapeutics [e.g., the Academic Drug Discovery Consortium (http://addconsortium.org)].

2.2. Assay Selection
HTS assays can be designed to measure the binding of a compound to a protein, changes in the activity of a protein, or alterations to a cellular phenotype. Terms such as “targeted” and “bio- chemical” are commonly used to describe assays using purified components, while “phenotypic” refers to assays that use cells or model organisms. Cellular or organismal assays are highly di- verse, however, and can measure broad phenotypic changes or modulation of specific proteins and pathways within the cellular context. It is also helpful to think of HTS assays by analogy to genetics. In this context, reverse chemical genetics refers to the selection of compounds based on their effect on a particular target; these compounds are then used to define the function of their

www.annualreviews.org • Small-Molecule Drug Discovery 15.3

a Reverse chemical screening b Forward chemical screening

Biochemical assays

Purified target enzyme

Product

Cell-based assays or animal-based model systems for disease

Inhibition of enzyme activity

Target- based screening

Phenotypic Screening

Hit selection and optimization

Orthogonal phenotypic validation

Hit to lead

X Y

Cell-based assays or animal-based model systems for disease

Hit selection

Hit to lead

Figure 1

Comparison of high-throughput screening workflows for reverse and forward chemical screening. (a) In a reverse or target-based chemical screening, biochemical assays are used to identify compounds that modulate the activity of a purified target protein of interest (the example in the figure shows an enzyme activity assay). Once bioactive hits are selected, the activity and/or the phenotypic consequences of those hits are usually tested in cell-based orthogonal assays. Iterative rounds of this process coupled with medicinal chemistry usually lead to a selective bioactive compound that modulates the activity of the protein of interest in vivo. (b) In a forward or phenotypic chemical screen, a cell- or animal-based assay is developed, usually in a microplate format. This assay is then used to identify hits from a library of chemicals that cause a phenotypic alteration and show changes in the assay readout compared with the control (the example in the figure shows an epifluorescence assay readout). Iterative rounds of hit validation and optimization through medicinal chemistry are then performed, with subsequent assays used to prioritize a lead compound. The target of this bioactive compound is then identified through either genetics or biochemistry.

target in cells (Figure 1a). Conversely, forward chemical genetics selects compounds based on a phenotypic effect; subsequent identification of the compounds’ target(s) then provides insight into the biological system (Figure 1b).
Assay development for HTS is an area of continuous innovation to address new targets and take advantage of innovations in spectroscopy, microscopy, and cell biology. Various contemporary detection technologies employed by HTS have been reviewed by Janzen (67) and are summarized in Table 1. Emerging examples include technologies that interrogate conformational changes in a protein upon small-molecule binding (27), multiparametric phenotypic assays to profile the biological effect of hit compounds (28), and model-organism assays that simultaneously evaluate compound efficacy and potential toxicity in vivo (19, 22, 43, 103, 125). Because of the diversity of assays, a detailed description of assay development is beyond the scope of this review. Nevertheless,

15.4 Markossian et al.

Table 1 Comparison of assay technologies commonly used in chemical screening and validation

Assay technology Plate reader Imaging Biophysical Gel based
Example readouts ABS, FINT, FRET,
FP, and LUM Epi-FINT and FACS SPR, MS, and NMR Western blot
Throughput Very high Moderate/high Moderate Low
Data content Moderate Very high High High
Likelihood of artifacts/interference High Moderately low Moderately low Moderately low
Multiplexing Moderate High Moderate Moderate

Abbreviations: ABS, absorbance; FACS, fluorescence-activated cell sorting; FINT, fluorescent intensity; FP, fluorescence polarization; FRET, fluorescence resonance energy transfer; LUM, luminescence; MS, mass spectrometry; NMR, nuclear magnetic resonance; SPR, surface plasmon resonance.

good assay design is critical to the success of HTS, and many excellent tutorials are available, including the Assay Guidance Manual, a regularly updated e-book available through the National Center for Biotechnology Information’s Bookshelf site (124).

2.2.1. Target-based biochemical assays. Protein-based (or nucleic acid–based) biochemical assays measure the engagement of a compound with the target, usually through changes in bio- chemical activity (Figure 1a). Many enzyme inhibitors have been developed through target-based drug discovery; biochemical assays can also measure protein–protein or protein–nucleic acid in- teractions. Other target-based assays utilize whole cells, particularly for transmembrane proteins such as G-protein-coupled receptors and ion channels, but measure the biochemical activity of the membrane-embedded protein. Common technologies used for detection include absorbance, fluorescence, and luminescence. Fluorescence-based detection is the most versatile; three typical modes of fluorescent readouts are fluorescent intensity, fluorescence resonance energy transfer, and fluorescence polarization. These formats can be designed to monitor losses or gains in signal. For instance, protease substrates are modified to contain a dye that becomes fluorescent after proteolysis; protease inhibitors would lead to a loss of fluorescence relative to a well containing uninhibited enzyme. Target-based screens have the important advantage that the molecular target of the screening hit is known. However, they can also suffer from their simplicity and the loss of biological context; target-based screens are therefore followed by cell-based assays.
A specialized type of targeted screen measures the binding of compounds to the target, using a biophysical assay such as nuclear magnetic resonance, surface plasmon resonance, or X-ray crystallography. While these methods provide high-resolution data, they are also relatively reagent intensive and/or low throughput compared with typical HTS formats. Binding assays are generally used for fragment-based drug discovery, in which small molecules roughly half the mass of a drug (e.g., <250 Da) are tested. In principle, fragment-based drug discovery allows efficient sampling of chemical diversity space using a small number of molecules (typically hundreds to thousands); in practice, it has been successful for targeting “undruggable” protein classes and obtaining selective kinase inhibitors (52) (see Section 4.2).

2.2.2. Cell-based phenotypic assays. Phenotypic assays, utilized in forward chemical genetic screens (Figure 1b), have a good track record in identifying first-in-class molecules (127). The designs are highly diverse but can be divided into cell-by-cell assays and well-averaged assays. Cell-by-cell assays are usually accomplished by high-content analysis (HCA). Bioluminescence and fluorescence readouts are most common for well-averaged assays, with examples includ- ing luciferase-based reporter gene assays, alamarBlue (fluorescence) cell viability assays, and

www.annualreviews.org • Small-Molecule Drug Discovery 15.5

fluorescence resonance energy transfer or protein complementation assays to detect the dynamics of protein interactions (126). One promising new technology is the cellular thermal shift assay, which screens for small-molecule target engagement in cells (85).
HCA combines a cell-based assay with automated high-throughput fluorescent imaging. The ability to analyze many morphological features with high throughput has made HCA a preeminent screening method (14). HCA typically uses cells engineered with fluorescent reporters or protein fusions, but antibody-based immunofluorescence is also possible. One important advantage of HCA is the ability to reanalyze acquired images. A multiparametric HCA records a wide array of cellular features; reanalysis can then be carried out on the stored images to select key features or by principal component analysis to explore complex phenotypes (95). Recent improvements in hardware include LED light sources for uniform intense illumination and large-format cameras; image-analysis software offers multiple feature-extraction algorithms and machine-learning capa- bility (21, 115). HCA has been particularly fruitful for screening primary and patient-derived cells for genetic diseases such as DMD (Section 3.3). For instance, Ketley et al. (74) used imaging to identify small molecules that inhibited the accumulation of sequestered Muscleblind-like (MBNL) proteins in the nuclei of cells derived from myotonic dystrophy patients.
Cell-based assays offer the attractiveness of directly identifying small-molecule modulators that elicit desired cellular effects. Critics of cell-based phenotypic screening point to the lack of direct evidence of target engagement, something that target-based biochemical screens are designed to address. However, cell-based assays can be designed to read out activities proximal to the disease phenotype and therefore can be used to identify new drug targets or new ways of modulating the genetically altered proteins, particularly for monogenic diseases (18, 41). This screening approach has been termed mechanism-informed phenotypic drug discovery (90). Use of this method with a fluorescent membrane-potential assay led to the discovery of the CF drug ivacaftor (VX-770), which targets a mutated form of CFTR (138, 140) (Section 3.1).

2.3. Chemical Library Selection
It is important to screen an appropriate collection of compounds (library) to maximize the discovery of target-relevant hits by HTS. Compound libraries are available from commercial vendors, and some groups will have proprietary collections synthesized in-house or curated from other sources. Libraries fall into five broad categories: (a) approved or experimental drugs; (b) bioactives with known mechanisms of action; (c) targeted chemotypes designed to bind to classes of proteins, such as G-protein-coupled receptors or kinases; (d ) diverse synthetic compounds; and (e) natural products, which can be mixtures or pure compounds (63). Targeted HTS efforts tend to favor targeted chemotypes and diverse libraries, whereas phenotypic screens can utilize these sets as well as bioactives, drugs, and natural products. Bioactives, drugs, and targeted chemotypes can provide hypotheses for target identification after phenotypic screening.
Clinical drug sets and bioactive chemical collections typically contain fewer than 2,000 com- pounds and are often used during a pilot screen. The purpose of a pilot screen is twofold. First, the pilot provides an evaluation of the assay performance metrics such as signal-to-background ratio, robustness of controls, reagent stability, hit-rate estimation, and possible artifacts (see Section 2.4). Second, drug libraries can be used to repurpose an approved agent for a new disease. In some cases, existing drugs have been taken directly to clinical trials; at a minimum, drugs or experimental therapeutics (i.e., compounds still in clinical trials) can serve as late-stage leads that require medicinal chemistry fine-tuning to be used for the new clinical indication. Buoyed by the promise of rapid clinical translation at low cost, drug repurposing has been particularly popular for rare diseases, cancers, and infectious diseases of the developing world (26, 36, 105, 122).

15.6 Markossian et al.

Diversity-based libraries are most commonly used for HTS. Libraries of 105–106 compounds are selected from multiple vendors, and the same compound set is used for many HTS cam- paigns. When purchasing a compound library, the laboratory will first evaluate the compounds using a set of filters for drug likeness. This approach is meant to remove compounds that have known liabilities for drug development, such as chemically reactive or unstable moieties and features associated with toxicity or poor bioavailability. Features include chemical functionali- ties and overall physicochemical features (e.g., size and hydrophobicity) (8, 66). Diverse libraries allow for the discovery of novel chemical scaffolds against the target of interest. By contrast, tar- geted libraries, such as kinase-inhibitor-like scaffolds, allow smaller-sized screens but can miss out on chemical (and perhaps mechanistic) novelty. Other types of diverse compound libraries include low-molecular-weight (molecular weight <250) fragments (51), natural products (63), and DNA-encoded chemical libraries (14). DNA-encoded libraries are highly complex mixtures often containing 108 or more compounds; these are selected for binding to a target, and the active compounds are identified by sequencing their associated DNA barcodes.
In recent years, HTS libraries have become available for precompetitive research. Through consortia and open innovation platforms, researchers can access pharmaceutical compound col- lections and screening resources (12). While these platforms can have a strong impact on drug and probe discovery for genetic diseases, there are also caveats to consider, including data transparency, intellectual property, and restrictions imposed on the use of active compounds (54).

2.4. Hit Identification and Validation
Initial hits are almost always obtained from HTS; typically, 0.03%–0.5% of compounds show potential activity (up to 5% for bioactive or drug collections). Compounds are usually considered active if they alter the assay signal by at least 3 standard deviations from the mean values of untreated controls (150). The challenge, however, is to separate the real hits that act in a drug-like manner from false positives. To be a probe or drug, a small molecule should bind to its target at a single site, modulate activity in a dose-dependent manner, and not interact with the other assay components. Additionally, most HTS campaigns are looking for reversible binders and target- or phenotype-selective compounds.
Following HTS, the team enters the phase of hit validation, where a series of assays are used to select hits for further consideration. First, active compounds are retested in the primary as- say at multiple doses to assess potency and the quality of the dose–response curve; a very steep dose–response curve can indicate artifacts such as compound aggregation (123) or nonspecific tox- icity. Second, many compounds interact with the assay components themselves, and it is critical to remove these false positives by testing in an orthogonal assay format. For instance, colored or fluo- rescent compounds can affect absorbance or fluorescence-based assays; luciferase-based reporter gene assays can be affected by compounds that inhibit luciferase itself (128). Historical screening data (e.g., using in-house data or databases such as PubChem and ChEMBL) can identify whether a compound is frequently active in assays with a certain readout or active in unrelated assays. Such frequent hitters are also called pan-assay interference compounds (PAINS) (7). Some common chemical scaffolds are considered PAINS; many laboratories remove such structures from further consideration, while others advocate for a data-driven approach (30, 118). Ultimately, chemi- cal structures of confirmed hits are compared to identify structure–activity relationship trends. Compounds and closely related analogs are then synthesized or purchased, and detailed mecha- nistic studies are initiated. Hits that show appropriate mechanisms of action are further improved through iterative cycles of compound design and testing.

www.annualreviews.org • Small-Molecule Drug Discovery 15.7

3. SMALL-MOLECULE DISCOVERY FOR MONOGENIC DISEASES
Since the early stages of chemical genetics in the early 1990s, multiple reverse and forward che- mical genetic screens have elucidated pathophysiological mechanisms of genetic disorders and thus have been established as a supplementary approach to classical genetics. In recent years, there has been significant progress in developing drugs for CF, SMA, and transthyretin amy- loidosis; therapies for DMD and GD have also advanced toward the clinic. In the following sections, we discuss examples of screens (Table 2) that yielded novel drugs or drug leads for those diseases.

3.1. Cystic Fibrosis
Among Caucasians, CF is the most common autosomal recessive disorder that is inherited in a Mendelian monogenic fashion (75). The disease phenotype is caused primarily by mutations in CFTR, encoding a protein kinase A–regulated chloride channel. Multiple classes of CFTR mutations are observed, causing defects in protein production, trafficking, function, or stability. Deletion of phenylalanine 508 (F508del), which leads to misfolding of the CFTR protein and premature degradation via the endoplasmic reticulum–associated degradation pathway, is the most prevalent disease-causing mutation (110, 111) (Figure 2a,b). Mutations that cause gating (ion flux) defects are also observed in CF patients. The most prevalent gating mutant is G551D-CFTR, observed in approximately 4–5% of CF patients (Figure 2a,b).
Therapeutic strategies have focused mainly on identifying either correctors or potentiators of mutant CFTR. Corrector therapies promote trafficking of ∆F508-CFTR to the cell surface. Potentiators improve the chloride conductance of the CFTR channel. There are currently two US Food and Drug Administration (FDA)–approved small-molecule drugs targeting CFTR— ivacaftor and lumacaftor—both which were identified through cell-based HTS (Figure 2). Vertex Pharmaceuticals developed a cell line containing ∆F508-CFTR and tracked changes in CFTR function in two ways. First, they added compounds for 16 hours, removed the compounds, and measured ion flux using fluorescent dyes; this assay was designed to monitor increased levels of active ion channel resulting from improved trafficking (correcting). Second, they measured ion flux in the presence of compounds after a short treatment, thereby measuring improvements in ion flux resulting from compound-induced activation (potentiation) of the channel (140). A screen of 228,000 chemically diverse compounds with the potentiator assay, followed by medicinal chemistry, identified ivacaftor (VX-770), the first CFTR channel gating modulator approved by the FDA and European Medicines Agency (EMA) for patients carrying the G551D-CFTR mutation (110, 138) (Figure 2c). A screen of 164,000 small molecules with the corrector assay led to the development of lumacaftor (VX-809), which corrected the folding and processing defect of ∆F508- CFTR (139). VX-809 showed only modest effects in clinical trials (38) and was subsequently approved by the FDA as a combination therapy with ivacaftor (20) (Figure 2c). Interestingly, the efficacy of these channel correctors and potentiators was demonstrated using cultured human bronchial epithelial cells isolated from the lungs of patients with CF (138, 139). This study provided an early example of the use of human-relevant in vitro models in lieu of animal models for drug development.
Alternative assay formats have sought to measure CFTR trafficking directly or to bypass CFTR
mutations. In one interesting example, an mCherry-Flag-∆F508-CFTR trafficking reporter was expressed under an inducible promoter. Expressed protein throughout the cell was visualized by mCherry fluorescence; CFTR that was successfully trafficked to the cytoplasmic membrane was visualized by immunofluorescent staining using an anti-Flag antibody. The researchers could use this assay to detect increases in plasma membrane localization of mCherry-Flag-∆F508-CFTR

15.8 Markossian et al.

Table 2 Examples of FDA- or EMA-approved drugs or lead compounds identified through small-molecule screening in monogenic diseases

Disease
Pathway affected
Gene mutated

Chemotherapy Clinical
development status

Target

HTS assay

Reference(s)
Cystic fibrosis Transepithelial ion transport CFTR Ivacaftor (VX-770) Approved G551D-CFTR MIPDD: cell-based FRET assay for membrane
potential 138, 140
Lumacaftor (VX-809) and
ivacaftor (VX-770) Approved ∆F508-CFTR MIPDD: cell-based FRET assay for membrane
potential 139, 140
Tezacaftor (VX-661) and ivacaftor Phase 3 ∆F508-CFTR MIPDD: cell-based FRET assay for membrane
potential 98, 140
Spinal muscular atrophy Function and life of α-motor neurons SMN1 RG7916 Phase 2 SMN2 splicing MIPPD: cell-based
luciferase reporter assay for
SMN2 splicing modifiers 94
Branaplam Phase 2 SMN2 splicing MIPPD: cell-based
luciferase reporter assay for
SMN2 splicing modifiers 100
LDN-75654 Preclinical SMN2 splicing MIPPD: cell-based
Luciferase reporter assay for SMN2 splicing modifiers 33, 34
LDN-76070 Preclinical SMN2 splicing MIPPD: cell-based
Luciferase reporter assay for SMN2 splicing modifiers 33, 34
Arylpiperidines Preclinical SMN2 splicing MIPPD: cell-based
Luciferase reporter assay for SMN2 splicing modifiers 147
Olesoxime Phase 2/3 Neuroprotection Phenotypic cell-based assay
for motor neuron survival 17, 29
(Continued)

Table 2 (Continued)

Disease
Pathway affected
Gene mutated

Chemotherapy Clinical development
status

Target

HTS assay

Reference(s)
Duchenne muscular dystrophy Sarcolemma structural integrity Dystrophin Ataluren (PTC124) Phase 3 Nonsense
mutations of dystrophin MIPPD: cell-based
luciferase reporter assay for suppressors of nonsense
mutations of dystrophin 15, 60, 145
SMTC1100 Phase 1/2 Utrophin MIPPD: cell-based
luciferase reporter assay for utrophin promoter activation 60, 91, 129
Transthyretin amyloidoses Protein folding Transthyretin Tafamidis Approved Transthyretin misfolding Target-based screening using a biochemical
fibril-formation assay 39
Gaucher disease Lysosomal storage GBA1 NCGC607 Preclinical Mutant gluco- cerebrosidase Target-based assay with tissue extract of a patient
with Gaucher disease 1, 61
NCGC758 Preclinical Mutant gluco- cerebrosidase Target-based assay with tissue extract of a patient
with Gaucher disease 2, 61
Huntington’s disease Polyglutamine aggregation HTT Y-39983 Preclinical Rho kinase Phenotypic cell-based
aggregation assay 10
C2-8 Preclinical Unknown Phenotypic cell-based
aggregation assay 152

Abbreviations: EMA, European Medicines Agency; FDA, US Food and Drug Administration; FRET, fluorescence resonance energy transfer; HTS, high-throughput screening; MIPDD, mechanism-informed phenotypic drug discovery.

a
Cl– Cl–

Class III: G551D-CFTR

c
Mechanism-informed phenotypic screen

Recombinant cell-based assay

FRET readout measuring membrane- potential changes

b

CFTR
mutation Worldwide prevalance Defect Consequence Approved therapy Mechanism of action
G551D ~4–5% Gating Reduced protein levels Ivacaftor Corrector
ΔF508 ~70% Protein trafficking Reduced protein function Lumacaftor/ ivacaftor Corrector/ potentiator
Figure 2

Summary of the workflow for cystic fibrosis drug discovery through high-throughput screening (HTS). (a) Functional classes of the two CFTR mutations with approved targeted therapy. ∆F508-CFTR is a class II mutation that affects trafficking of CFTR into the plasma membrane. G551D-CFTR is a class III mutation that causes a gating defect in the channel that impairs its function. (b) Table summarizing key details about the two CFTR mutations along with their approved therapies. (c) Workflow of drug discovery through HTS for ivacaftor and lumacaftor. Both drugs were discovered by mechanism-informed phenotypic drug discovery. Ivacaftor was
discovered through HTS designed to discover CFTR potentiators (blue); lumacaftor was discovered through HTS designed to discover CFTR correctors (orange). Additional abbreviation: FRET, fluorescence resonance energy transfer.

levels in cells treated with the CF corrector VX-809 (18). In an orthogonal approach aimed at reducing sodium hyperabsorption and increasing airway surface liquid hydration, researchers targeted the epithelial sodium channel ENaC, a downstream effector of CFTR. A live-cell, loss- of-function small interfering RNA (siRNA) screen using a FLIPR membrane-potential voltage- sensitive fluorescent dye in combination with the specific ENaC blocker amiloride identified ENaC regulators and additional potentially druggable targets whose modulation could normalize airway surface liquid homeostasis (4).

www.annualreviews.org • Small-Molecule Drug Discovery 15.11

3.2. Spinal Muscular Atrophy
SMA is an orphan genetic disease and a valuable example of how chemical genetics can be used to understand and eventually treat orphan diseases. It is an autosomal recessive pediatric disease and the leading cause of infant mortality in the United States that results from a genetic condition. Mortality is due to the rapid and severe degeneration and loss of α-motor neurons in the anterior horn of the spinal cord (81). Degeneration is due to the lack of the survivor of motor neuron (SMN) protein, which is normally ubiquitously and abundantly expressed in neurons. Humans possess two nearly identical copies of the SMN genes, SMN1 and SMN2 (77), which differ by alternative splicing: SMN1 produces full-length mRNA and a functional protein, whereas SMN2 produces mostly an alternatively spliced mRNA that lacks exon 7 and is translated into a truncated protein that is readily degraded (35, 80). SMA patients have either a mutation in SMN1 or a homozygous deletion of this gene (29, 81). Therefore, SMA patients rely solely on the SMN2 product and hence cannot produce enough full-length SMN protein to preserve the function and life of α-motor neurons.
Current therapeutic strategies have relied heavily on identifying selective SMN2 splicing mod- ifiers that promote exon 7 inclusion. Nusinersen (Spinraza), an antisense oligonucleotide that increases exon 7 inclusion in SMN2 mRNA transcripts, is the first therapy approved by the FDA to treat children and adults with SMA (99). Nusinersen is administered directly to the central ner- vous system using an intrathecal injection. There are also multiple small-molecule SMN2 splicing modifiers that are either in clinical trials or in preclinical stages of development for the treatment of SMA. If approved, these small molecules will possess the advantage of oral administration and a broad peripheral and central nervous system distribution (146).
These selective, small-molecule-based SMN2 splicing modifiers were discovered by similar cell-based HTS campaigns. An SMN2 minigene (SMN2mg) reporter construct, designed so that only the full-length SMN2mg mRNA containing exon 7 produced active firefly luciferase, was stably transfected into HEK293H cells. PTC Therapeutics used this cell line to screen a li- brary of approximately 200,000 compounds. Semiquantitative end-point reverse transcription PCR of SMN2mg was subsequently used for hit validation. Further optimization of the hits led to three orally administered compounds (the coumarins SMN-C1 and SMN-C2 and the pyrido[1,2- α]pyrimidin-4-one SMN-C3) that increased full-length SMN2 mRNA levels in SMA type I patient fibroblasts with nanomolar potency. The splicing modifiers led to an increase in SMN protein levels and prevented neuromuscular pathology in ∆7 mice, a model of severe SMA (94). The clinical candidate RG7916 is currently in phase 2 clinical trials for treatment of SMA (PTC Therapeutics, Roche Pharmaceuticals, and the SMA Foundation).
A similar screen was run in NSC34 motor neuron cells with the Novartis compound library (∼1.4 × 106 compounds). Positive hits from this screen were assessed via quantitative PCR to confirm the splicing variant and via the enzyme-linked immunosorbent assay to confirm increased
SMN protein levels. Further optimization of these compounds in both mouse myoblasts and SMA patient fibroblasts led to the production of NVS-SM1 and NVS-SM2. These compounds elevated full-length SMN protein expression and promoted survival in a mouse model of severe SMA. The mechanism of action of these molecules involved their binding to and stabilizing the complex of U1 small nuclear ribonucleoproteins with the SMN2 mRNA (100). From this program, branaplam (Novartis) is currently in phase 2 clinical trials for treatment of SMA.
Two additional programs are in preclinical testing. A robust luciferase reporter assay like the ones used in the above-described screens for splicing regulators for HTS was also designed by Cherry et al. (33). This assay was developed in HEK293 cell lines and screened against 115,000 small molecules from diverse sources. Two novel compounds (LDN-75654 and LDN-76070) identified from this screen increased endogenous SMN protein levels in patient fibroblasts.
15.12 Markossian et al.

LDN-76070 also increased SMN protein levels in a mouse model of severe SMA, which resulted in enhanced motor function and an increased life span in these mice (33, 34). Using this same cell-based SMN2-luciferase reporter assay, the NIH Chemical Genomics Center screened their Molecular Libraries Small Molecule Repository of 210,386 compounds. After further hit-to-lead optimization, they identified two arylpiperidines as novel modulators of SMN protein (147).
Neuroprotection could provide an alternate, though nonspecific, approach to treating SMA and other motor neuron and neurodegenerative diseases. The idea is that therapies could com- pensate for the lack of SMN by enhancing the survival of motor neurons (109). A phenotypic cell-based assay was used to score motor neuron cell death in vitro when screened with a collec- tion of approximately 40,000 compounds from Trophos. The screen was conducted in the hopes of identifying potential drug candidate for the treatment of amyotrophic lateral sclerosis. The screen revealed olesoxime (TRO19622) as a molecule that promoted motor neuron survival in the absence of trophic support. This compound also showed promising levels of neuroprotection and neuroregeneration for motor neurons in several animal models (17). Unfortunately, it did not significantly improve amyotrophic lateral sclerosis patient survival during a phase 3 clinical trial, but it is currently in phase 2/3 trials for the treatment of SMA (29).

3.3. Other Monogenic Diseases
CF and SMA are not the only two monogenic diseases with approved therapies. Extensive research in the last decade has provided successful therapies or potential avenues for the treatment of multi- ple other monogenic diseases, including DMD, transthyretin amyloidosis, GD, and Huntington’s disease. Therapeutic advances for those diseases are briefly discussed in the following sections.

3.3.1. Duchenne muscular dystrophy. DMD is the most common dystrophic disorder in hu- mans and is caused by mutations or deletion of the X-linked dystrophin gene. Dystrophin allows the sarcolemma to maintain structural integrity, and its loss leads to multiple pathophysiolo- gies, including perturbation in calcium homeostasis, oxidative stress, inflammation, fibrosis, and ultimately muscle fiber degeneration and death. Approximately two-thirds of the time, disease mutations are large deletions that usually cause a translational frame shift, resulting in an unstable transcript and dystrophin protein (15). The presence of a premature stop codon accounts for the loss of dystrophin expression in another 13% of patients. Current therapeutic research aims to restore dystrophin expression, upregulate its autosomal homolog utrophin, or compensate for the lack of dystrophin (60).
These strategies have identified multiple compounds in different stages of clinical develop- ment for the treatment of DMD. Restoring dystrophin expression in patients that carry nonsense mutations can be achieved by read-through therapy. Multiple screens aiming to identify com- pounds capable of suppressing nonsense mutations have been conducted (49, 50, 145). Ataluren (PTC124), a read-through compound identified through HTS by PTC Therapeutics, is currently in phase 3 clinical trials for DMD and has been conditionally approved in Europe (15, 60). Such read-through therapy could also be used for other genetic diseases, such as CF (91). Another route to achieving dystrophin expression is through antisense-mediated exon skipping. Exon skipping can be applicable to up to 83% of all DMD mutations and thus is one of the most promising treatment strategies (121). Multiple HTS studies have yielded promising compounds capable of enhancing antisense oligonucleotide–mediated dystrophin exon skipping (65, 73, 96). Multiple compounds have also been screened for utrophin promoter activation. From one such screen, an orally bioavailable compound named SMTC1100 is currently in phase 1/2 clinical trials (60, 91, 129). Finally, HTS efforts are being pursued to identify compounds that can activate and

www.annualreviews.org • Small-Molecule Drug Discovery 15.13

differentiate muscle stem cells in the hopes of compensating for the lack of dystrophin by regen- erating diseased muscle in DMD patients (13, 60, 95).
3.3.2. Transthyretin amyloidoses. Pharmacological chaperones are compounds that bind to a disease-causing protein and thermodynamically stabilize it, keeping it from degrading or ag- gregating. An example of chaperone therapy is the treatment of the rare protein misfolding and aggregation genetic disorders called transthyretin amyloidoses. Transthyretin is a transporter present in blood as a tetramer. Mutations in transthyretin destabilize the tetramer, promoting monomer misfolding and aggregation, leading to amyloidogenesis. Tafamidis, which kinetically stabilizes the tetramer, slows disease progression and has been approved by the EMA for the treatment of transthyretin amyloidoses (39).
3.3.3. Gaucher disease. GD can also benefit from chaperone therapy. GD results from mutations in both alleles of the GBA1 gene, leading to deficient glucocerebrosidase activity, accumulation of glycolipid substrates, and defective lysosomal function. There are multiple types of GD that fall along a spectrum of clinical phenotypes, including nonneuronopathic type 1 Gaucher dis- ease (GD1) and neuronopathic types 2 and 3 Gaucher disease (GD2 and GD3) (89). Therapeutic strategies for GD include chaperone therapy (137), enzyme replacement therapy, and substrate reduction therapy (141). Chaperone therapy is an attractive strategy to treat all types of GD be- cause the chaperone compound can pass through the blood–brain barrier and restore the mutant glucocerebrosidase activity in neurons (137). Stabilizing glucocerebrosidase could also be a treat- ment for Parkinsonism, since mutations in GBA1 are the most frequent known genetic risk factor for Parkinson’s disease (87, 137). Currently, there are multiple enzyme replacement therapies and substrate reduction therapies to treat GD1 but no approved chaperone therapy for GD (141). A recent target-based HTS of 250,000 compounds identified modulators of mutant glucocerebrosi- dase activity. Further validation of the chaperone activity of the lead compounds using a cell-based orthogonal assay identified promising novel, noninhibitory chaperones of mutant glucocerebrosi- dase (61). From this screen, the chaperone NCGC607 restored protein levels and the activity of glucocerebrosidase, leading to a reduction in glycolipid storage in both induced pluripotent stem cell (iPSC)–derived macrophages and dopaminergic neurons from patients with GD1 and GD2. NCGC607 also reduced α-synuclein levels in iPSC-derived dopaminergic neurons from patients with parkinsonism (1).
3.3.4. Huntington’s disease. Huntington’s disease is an autosomal dominant, progressive neu- rodegenerative disease caused by expansion of a CAG repeat in exon 1 of the huntingtin gene (HTT ), leading to a polyglutamine expansion in the N terminus of the huntingtin protein. Poly- glutamine aggregates made of insoluble mutant huntingtin (mhtt), which can be readily detected within neurons in the brain, is a hallmark of Huntington’s disease. There are currently no cures for this disease (119). Multiple cell-based viability screens (3, 131, 142, 144) as well as in vitro (11, 64, 143) and cell-based (10, 37, 44, 57, 106, 131, 152) aggregation assays have identified a number of small molecules, but none of the compounds have been successfully developed into drugs. iPSC-based drug screening for Huntington’s disease is currently on the rise and has the potential to find better therapeutics, since it may include previously undefined cellular activities that cooperate in causing disease (151).

4. SMALL-MOLECULE SCREENING IN CANCER
Through immense research efforts and advances, multiple targeted therapies have been approved for use in cancer treatment in the past three decades. Both reverse and forward chemical genetics

15.14 Markossian et al.

have played pivotal and synergistic roles in the discovery of several FDA-approved cancer drugs (Table 3). In the following sections, we present selected examples of drugs that were discovered through reverse (e.g., Gleevec, vemurafenib, and olaparib) or forward (e.g., HDAC inhibitors) chemical genetic screens as well as synthetic lethal screening (e.g., olaparib).

4.1. Discovery of Gleevec
Chronic myeloid leukemia is characterized by the presence of a BCR-ABL fusion gene, which is the cause of pathogenesis. The constitutive tyrosine kinase activity of the BCR-ABL oncogene is essential to transform hematopoietic cells in vitro and in vivo, suggesting that inhibiting this activity would be an attractive therapeutic strategy (42). Indeed, this strategy led to the discovery of the FDA-approved drug Gleevec (Novartis) for the treatment of chronic myeloid leukemia (42). Gleevec was identified through target-based HTS that monitored biochemical inhibition of protein kinase C (42, 58, 153). Subsequent medicinal chemistry led to compounds that had enhanced activity against tyrosine kinases but lost their activity against protein kinase C (42). Further optimization to improve solubility and oral bioavailability led to Gleevec, which showed the highest selectivity for growth inhibition of BCR-ABL-expressing cells and was deemed the most promising compound for clinical development (42). Subsequent in vitro and in vivo profiling of Gleevec showed that this drug is a selective inhibitor of the Abl tyrosine kinase that suppresses the growth of BCR-ABL-positive cells (24, 25, 31, 48). Gleevec entered phase 1 clinical trials in 1998 and was approved by the FDA in 2001 for the treatment of chronic myeloid leukemia because of its extraordinary efficacy in patients (42). Gleevec is a transformative drug for chronic myeloid leukemia and ushered in the age of so-called targeted therapies for cancer.

4.2. Discovery of Vemurafenib
Melanomas and many additional cancers contain a V600E mutation in BRAF kinase (BRAFV600E). Raf kinases are key downstream effectors of the Ras GTPase in the Ras-Raf-MEK-ERK pathway. Dimerization of Raf kinases is essential for their Ras-dependent activation. The V600E mutation, which is in the active loop of BRAF, increases its kinase activity and subsequently drives the proliferation of cancer cells. Selective inhibition of the V600E-mutated BRAF kinase activity was thought to be a promising therapeutic strategy to treat those cancers. Vemurafenib, the first drug designed using fragment-based lead discovery, was also the first BRAFV600E kinase inhibitor approved by the FDA to treat late-stage melanoma (133, 149).
The discovery of vemurafenib was described in a 2012 review by Bollag et al. (16). Briefly, a library of 20,000 fragments (150–350 Da), selected for their chemical properties and diversity, was screened at a high concentration against multiple divergent kinases. Compounds that could inhibit at least three of the five kinases were chosen for further studies. Cocrystallization of the kinase and the hit compound was then used to identify true binding interactions. The 7- azaindole core was one of the nonspecific hits identified from the first round of screens. Analogs were then built around the 7-azaindole core to optimize potency and selectivity. Subsequent cycles of screening, crystallography, and structure-guided design with BRAFV600E and wild-type BRAF led to the discovery of PLX4720, an analogue of vemurafenib. PLX4720 demonstrated remarkable selectivity for BRAFV600E over wild-type BRAF in melanoma and colorectal cancer cell lines. PLX4720 was also efficacious in B-RafV600E-dependent tumor xenograft models. Plexxikon, in collaboration with Roche Pharmaceuticals, initiated clinical trials with vemurafenib in 2006, culminating in the approval of vemurafenib in the United States (in 2011) and Europe (in 2012) to treat late-stage melanoma. Since its discovery, vemurafenib has been essential for understanding

www.annualreviews.org • Small-Molecule Drug Discovery 15.15

Annu. Rev. Genom. Hum. Genet. 2018.19. Downloaded from www.annualreviews.org Access provided by Australian National University on 05/28/18. For personal use only.

Table 3 Examples of selected breakthrough lead compounds and FDA-approved drugs that hit milestones in the past two decades in
oncology and were identified through small-molecule screening

Drug

Milestone
Cancer type Targeted disease etiology

Target
Mechanism of action

Screen type HTS assay for lead compound discovery

References
Gleevec First personalized medicine/targeted therapy drug that revolutionized treatment and prognosis of chronic myeloid leukemia Chronic myeloid leukemia BCR-ABL
fusion gene ABL
tyrosine kinase ATP analogue Target-based screen Biochemical kinase inhibition assay measured by enzymatic transfer of radiolabeled
phosphate 42, 153
Vemurafenib First drug designed using fragment-based lead discovery Late-stage melanoma V600E
activating mutation in BRAF V600E- BRAF
Kinase ATP analogue Target-based screen: scaffold- or fragment-
based screening approach Biochemical kinase inhibition assay: AlphaScreen technology, Zr-LYTE kinase assays 16, 133, 149
Olaparib One of the first drugs designed to exploit synthetic lethality Germline
BRCA-
mutated advanced ovarian
cancer Mutations in BRCA1 or BRCA2 PARP Blocks the binding of the substrate to PARP Target-based HTS Biochemical PARP enzymatic activity inhibition assay measured by FlashPlate scintillation
proximity assay 9, 45, 88
Vorinostat One of few drugs discovered by forward chemical genetics Cutaneous T cell lymphoma Hypoacetylation/ aberrant transcription regulation HDAC Binds to the pocket of the catalytic site Phenotypic screen Cell-based assay: survival and terminal differentiation of
transformed erythroid cells 56, 84, 114
Romidepsin (FR901228) One of few drugs discovered by forward chemical genetics Cutaneous T cell lymphoma Hypoacetylation/ aberrant transcription regulation HDAC Binds to the pocket of the catalytic site Phenotypic screen Cell-based assay: chemical screening of microbial compounds exhibiting antitumor
properties against tumor cell lines 93, 134–136
PRLX 93936
(phase 1/2 clinical trials) First drug lead against an indirect synthetic lethal target for the nondruggable oncogenic RAS Multiple myeloma Oncogenic RAS Mitochondrial VDAC Erastin analogue Phenotypic synthetic
lethality screen Cell-based assay: synthetic lethality screen using
engineered tumorigenic cells expressing mutated
oncogenic RAS 46, 148

Abbreviations: FDA, US Food and Drug Administration; HDAC, histone deacetylase; HTS, high-throughput screening; PARP, poly(ADP-ribose) polymerase; VDAC, voltage-dependent anion channel.

the biology of BRAF signaling and dimerization (107, 108). Resistance to vemurafenib is mediated by induced dimerization of an aberrantly spliced variant of BRAFV600E (107). Next-generation BRAF inhibitors that are currently in either preclinical or clinical stages of development include compounds that inhibit both monomeric and dimeric forms of RAF as well as peptides that block BRAF dimer formation (55, 71).

4.3. Discovery of Histone Deacetylase Inhibitors
In addition to genetic changes, aberrant transcriptional regulation in cancer is at least partially due to alterations in epigenetic regulation of the genome. HDACs modify histone tails, promoting closed chromatin structure, and hence regulate global transcription (104). A priori, the rationale to inhibit HDACs as a therapeutic strategy for cancer is weak, since HDACs are involved in basically all cell functions. Nevertheless, HDAC inhibitors have antitumor activity in vitro (84). There are multiple FDA-approved HDAC inhibitors for the treatment of cutaneous T cell lymphoma, peripheral T cell lymphoma, and multiple myeloma as well as many others in preclinical and clinical stages of development.
The first two approved HDAC inhibitors were discovered by purely phenotypic approaches (90). Vorinostat, the first HDAC inhibitor approved by the FDA for treatment of cutaneous T cell lymphoma (in 2006), predates HTS but was optimized through a phenotypic assay (84). In 1971, Friend et al. (56) observed that dimethyl sulfoxide can cause growth arrest and terminal differentiation of transformed erythroid cells. Although the target was not identified throughout the process, extensive hypothesis-based structure–activity studies led to the discovery of suberoy- lanilide hydroxamic acid (SAHA, i.e., vorinostat), which caused the growth arrest and subsequent death of a variety of transformed cells both in vitro and in animals (84, 114). SAHA was effica- cious in concentrations not toxic to normal cells (114). Further work led to the discovery that HDACs are the molecular targets of SAHA (84, 113). The second HDAC inhibitor approved by the FDA for the treatment of cutaneous T cell lymphoma (in 2009) was romidepsin (FR901228). Romidepsin was initially identified through chemical screening of microbial compounds derived from Chromobacterium violaceum (134–136). The screen was designed to identify compounds that induced morphological reversion of an H-ras-transformed NIH3T3 cell line and exhibited anti- tumor properties against murine and human tumor cell lines. FR901228 was later shown to be a novel HDAC inhibitor (93). These HDAC inhibitors highlight how phenotypic assays can lead to the discovery of new cancer drugs as well as new target classes.

4.4. Discovery of Olaparib by Exploiting Synthetic Lethality
The concept of synthetic lethality is an emerging approach for anticancer therapy and has been discussed in depth (69, 97). Historically, one of the bottlenecks for drug discovery in cancers with complex and dynamic genetic determinants has been in the failure to identify compounds that selectively kill cancer cells at doses that do not impair normal cell proliferation. In the post- Gleevec era, there is also an increased emphasis on preselecting patients who would benefit from chemotherapy. However, the frequently mutated, oncogenic, druggable targets have already been extensively screened by HTS; therefore, there is a clear unmet need for mining drugs against synthetic lethal targets to provide a solution for cancers that do not respond to known targeted therapies. Moreover, synthetic lethal screens may identify an indirect target that may be druggable even when the oncogenic target itself is considered undruggable. Thus, designing HTS assays that exploit synthetic lethality can identify compounds that selectively kill cancer cells and—if the genetic determinants of synthetic lethality are known—can be readily translated in the clinic.

www.annualreviews.org • Small-Molecule Drug Discovery 15.17

Olaparib, a poly(ADP-ribose) polymerase (PARP) inhibitor used in the treatment of ovarian cancers, is the first FDA-approved drug discovered by exploiting synthetic lethal interactions (68). Mutations in the BRCA1 or BRCA2 genes greatly increase the risk of developing breast and ovarian cancer in women and prostate cancer in men. Genetic synthetic lethal screens identified the PARP gene; subsequent testing of small-molecule PARP inhibitors validated the notion that BRCA inactivation strongly sensitized cells to PARP inhibition (23, 53). Synthetic lethality was due to the failure of cells carrying BRCA mutations to repair the DNA damage caused by PARP inhibition. Lead compounds for PARP inhibition were identified through a target-based screen. Briefly, HTS using a scintillation proximity assay identified phthalazinones as lead PARP inhibitors (9, 45). Subsequent medicinal chemistry optimization led to the identification of cell-active PARP inhibitors with good metabolic stability and nanomolar inhibition of both PARP-1 and PARP-2 with toxicity against BRCA1-deficient breast cancer cell lines (79, 88). Finally, high-content image- based screens that measured spindle polarity and centrosome-clustering phenotypes following treatment with AstraZeneca’s collection of phthalazinone PARP inhibitors demonstrated that the PARP 1/2/6 inhibitor AZ0108 (olaparib) was effective at low nanomolar concentrations (68). Olaparib was the first drug approved by the FDA and EMA to treat ovarian cancer patients with deleterious or suspected germline BRCA-mutated advanced ovarian cancer (79). There are now
multiple PARP inhibitors approved for treatment or in clinical trials.
The concept of synthetic lethality has gained traction for other cancer targets (such as oncogenic mutant RAS) and other DNA repair pathway components (such as CHK1 and ATR) (40, 82, 112). RAS mutations are present in 10–20% of human cancers, but RAS has been termed an undruggable target. Several screens have therefore targeted RAS mutant cancers by synthetic lethality (40, 46); the furthest advanced of these programs is PRLX93936 (Prolexys Pharmaceuticals), an analog of the HTS hit erastin (46, 148).

5. CHALLENGES AND POTENTIAL OF SMALL-MOLECULE SCREENING FOR HUMAN GENETIC DISEASES
The last decade has seen a significant expansion in the landscape of human genetic diseases. The Online Mendelian Inheritance in Man (OMIM) database currently contains more than 15,000 genes linked to identified Mendelian disorders (5). The NIH estimates that there are approxi- mately 7,000 orphan genetic diseases, with an average of 25–30 million Americans living with a rare disease. Moreover, evidence also suggests that the disease phenotypes of monogenic diseases represent the result of combined actions of multiple loci (32). Epigenetic factors and the envi- ronment are also thought to modulate disease phenotype. Hence, with the promise of genetic information comes the challenge of interpreting it for the purposes of drug discovery. Cancer, a complex human disease with a genetic component, lies at the other end of the genetic spectrum, as it generally shows a wide diversity of genetic changes (32). Diversity occurs both between cancers of a common tissue origin and in the hundreds of mutations seen within a given tumor cell. It is therefore a huge challenge to define the key mutations that should be targeted to treat any given patient.
The genetic and contextual complexities in human disease suggest that both target-based screening and phenotypic screening with laboratory cell lines may miss key factors and poten- tially druggable mechanisms. Recent advances in cellular technologies are driving the HTS field toward in vitro and in vivo models that may be more relevant to human disease. The availability of patient-specific iPSCs allows mechanistic testing and even HTS for monogenic, complex, and epigenetic disorders. In GD, for example, the development of an iPSC model for neuronopathic GD from patient dermal fibroblasts (130) and an iPSC-derived macrophage model exhibiting

15.18 Markossian et al.

the GD phenotypes (2) has enabled chemical screening for small molecules that can restore the heritable deficiency. Genetic engineering (e.g., with CRISPR/Cas9) further allows the design of patient-derived cells with the genetic mutation corrected in an otherwise identical background. Finally, the use of three-dimensional cell clusters, cocultures, and organoids is starting to have a dramatic impact in the screening of engineered and patient-derived cells for cancer (14). Com- binatorial screening of drugs is now also feasible and may play an important role in tackling the complexity of the pathways involved in cancer initiation and progression (86). The development of model organisms for in vivo HTS is also reaching maturity. For example, genome-edited ze- brafish models have recently become very popular in assessing cancer heterogeneity, progression, and relapse through chemical genetics (76, 78). Zebrafish have also recently been screened for Dravet syndrome (62) and other genetic disorders (103).
Phenotypic screening using such in vitro and in vivo relevant models faces the challenge of identifying the compound’s target. However, given the complexity of human biology, a favorable case can be made for mechanism-informed phenotypic drug discovery, as employed in cancer programs and in the discovery of ivacaftor and lumacaftor (90, 138, 140). To address this challenge, there have been significant advances in developing technologies to identify targets after chemical genetic screens (116).
When possible, patient-derived cells should be included in the drug discovery assay funnel, to ensure that the mechanism of the HTS-derived hit is relevant in the disease context. In some cases, these primary cell models have been considered more relevant and predictive than animal models and have even replaced animal models of efficacy for selecting clinical candidates (39, 138, 139). Cellular technologies will have an impact on toxicology as well, through the adoption of screening in iPSC-derived cardiac myocytes and more complex organ-on-a-chip technologies earlier in the drug discovery pipeline.
Tremendous collective efforts in the past decade have addressed the need for drug discovery in orphan genetic diseases as well as cancer. Future efforts toward drug discovery for genetic diseases not only will help find cures for those specific diseases but will also help uncover disease mechanisms and perhaps find cures for nongenetic forms of disease. In fact, some rare genetic diseases are specialized forms of common diseases. For instance, mutations in tau or amyloid precursor protein cause early-onset Alzheimer’s disease, and GD protein also predisposes people to Parkinson’s disease. Hence, the development of assay systems, chemical probes, and drugs for these rare diseases carries the promise of uncovering common mechanisms for the nongenetic forms of disease and may lead to therapies for a much wider set of patients.

DISCLOSURE STATEMENT
The authors are not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.

LITERATURE CITED
1. Aflaki E, Borger DK, Moaven N, Stubblefield BK, Rogers SA, et al. 2016. A new glucocerebrosidase chaperone reduces α-synuclein and glycolipid levels in iPSC-derived dopaminergic neurons from pa- tients with Gaucher disease and Parkinsonism. J. Neurosci. 36:7441–52
2. Aflaki E, Stubblefield BK, Maniwang E, Lopez G, Moaven N, et al. 2014. Macrophage models of Gaucher disease for evaluating disease pathogenesis and candidate drugs. Sci. Transl. Med. 6:240ra73
3. Aiken CT, Tobin AJ, Schweitzer ES. 2004. A cell-based screen for drugs to treat Huntington’s disease.
Neurobiol. Dis. 16:546–55

www.annualreviews.org • Small-Molecule Drug Discovery 15.19

4. Almaca J, Faria D, Sousa M, Uliyakina I, Conrad C, et al. 2013. High-content siRNA screen reveals global ENaC regulators and potential cystic fibrosis therapy targets. Cell 154:1390–400
5. Amberger JS, Bocchini CA, Schiettecatte F, Scott AF, Hamosh A. 2015. OMIM.org: Online Mendelian Inheritance in Man (OMIMⓍR ), an online catalog of human genes and genetic disorders. Nucleic Acids Res. 43:D789–98
6. Arkin MR, Ang KK, Chen S, Davies J, Merron C, et al. 2014. UCSF Small Molecule Discovery Center: innovation, collaboration and chemical biology in the Bay Area. Comb. Chem. High Throughput Screen. 17:333–42
7. Baell JB, Holloway GA. 2010. New substructure filters for removal of pan assay interference compounds (PAINS) from screening libraries and for their exclusion in bioassays. J. Med. Chem. 53:2719–40
8. Baell JB, Walters MA. 2014. Chemical con artists foil drug discovery. Nature 513:481–83
9. Banasik M, Komura H, Shimoyama M, Ueda K. 1992. Specific inhibitors of poly(ADP-ribose) synthetase and mono(ADP-ribosyl)transferase. J. Biol. Chem. 267:1569–75
10. Bard J, Wall MD, Lazari O, Arjomand J, Munoz-Sanjuan I. 2014. Advances in Huntington disease drug discovery: novel approaches to model disease phenotypes. J. Biomol. Screen. 19:191–204
11. Berthelier V, Wetzel R. 2006. Screening for modulators of aggregation with a microplate elongation assay. Methods Enzymol. 413:313–25
12. Besnard J, Jones PS, Hopkins AL, Pannifer AD. 2015. The Joint European Compound Library: boosting precompetitive research. Drug Discov. Today 20:181–86
13. Billin AN, Bantscheff M, Drewes G, Ghidelli-Disse S, Holt JA, et al. 2016. discovery of novel small molecules that activate satellite cell proliferation and enhance repair of damaged muscle. ACS Chem. Biol. 11:518–29
14. Bittker JA, Ross NT, eds. 2017. High Throughput Screening Methods: Evolution and Refinement. Cambridge, UK: R. Soc. Chem.
15. Blat Y, Blat S. 2015. Drug discovery of therapies for Duchenne muscular dystrophy. J. Biomol. Screen.
20:1189–203
16. Bollag G, Tsai J, Zhang J, Zhang C, Ibrahim P, et al. 2012. Vemurafenib: the first drug approved for
BRAF-mutant cancer. Nat. Rev. Drug Discov. 11:873–86
17. Bordet T, Buisson B, Michaud M, Drouot C, Galea P, et al. 2007. Identification and characterization of cholest-4-en-3-one, oxime (TRO19622), a novel drug candidate for amyotrophic lateral sclerosis.
J. Pharmacol. Exp. Ther. 322:709–20
18. Botelho HM, Uliyakina I, Awatade NT, Proenca MC, Tischer C, et al. 2015. Protein traffic disorders: an effective high-throughput fluorescence microscopy pipeline for drug discovery. Sci. Rep. 5:9038
19. Bowman TV, Zon LI. 2010. Swimming into the future of drug discovery: in vivo chemical screens in zebrafish. ACS Chem. Biol. 5:159–61
20. Boyle MP, Bell SC, Konstan MW, McColley SA, Rowe SM, et al. 2014. A CFTR corrector (lumacaftor) and a CFTR potentiator (ivacaftor) for treatment of patients with cystic fibrosis who have a phe508del CFTR mutation: a phase 2 randomised controlled trial. Lancet Respir. Med. 2:527–38
21. Bray MA, Vokes MS, Carpenter AE. 2015. Using CellProfiler for automatic identification and measure- ment of biological objects in images. Curr. Protoc. Mol. Biol. 109:14.17.1–13
22. Bruni G, Lakhani P, Kokel D. 2014. Discovering novel neuroactive drugs through high-throughput behavior-based chemical screening in the zebrafish. Front. Pharmacol. 5:153
23. Bryant HE, Schultz N, Thomas HD, Parker KM, Flower D, et al. 2005. Specific killing of BRCA2- deficient tumours with inhibitors of poly(ADP-ribose) polymerase. Nature 434:913–17
24. Buchdunger E, Cioffi CL, Law N, Stover D, Ohno-Jones S, et al. 2000. Abl protein-tyrosine kinase inhibitor STI571 inhibits in vitro signal transduction mediated by c-kit and platelet-derived growth factor receptors. J. Pharmacol. Exp. Ther. 295:139–45
25. Buchdunger E, Zimmermann J, Mett H, Meyer T, Muller M, et al. 1996. Inhibition of the Abl protein- tyrosine kinase in vitro and in vivo by a 2-phenylaminopyrimidine derivative. Cancer Res. 56:100–4
26. Bulman CA, Bidlow CM, Lustigman S, Cho-Ngwa F, Williams D, et al. 2015. Repurposing auranofin as a lead candidate for treatment of lymphatic filariasis and onchocerciasis. PLOS Negl. Trop. Dis. 9:e0003534

15.20 Markossian et al.

27. Butko MT, Moree B, Mortensen RB, Salafsky J. 2016. Detection of ligand-induced conformational changes in oligonucleotides by second-harmonic generation at a supported lipid bilayer interface. Anal. Chem. 88:10482–89
28. Caicedo JC, Cooper S, Heigwer F, Warchal S, Qiu P, et al. 2017. Data-analysis strategies for image-based cell profiling. Nat. Methods 14:849–63
29. Calder AN, Androphy EJ, Hodgetts KJ. 2016. Small molecules in development for the treatment of spinal muscular atrophy. J. Med. Chem. 59:10067–83
30. Capuzzi SJ, Muratov EN, Tropsha A. 2017. Phantom PAINS: problems with the utility of alerts for Pan-Assay INterference CompoundS. J. Chem. Inf. Model. 57:417–27
31. Carroll M, Ohno-Jones S, Tamura S, Buchdunger E, Zimmermann J, et al. 1997. CGP 57148, a tyrosine kinase inhibitor, inhibits the growth of cells expressing BCR-ABL, TEL-ABL, and TEL-PDGFR fusion proteins. Blood 90:4947–52
32. Chakravorty S, Hegde M. 2017. Gene and variant annotation for Mendelian disorders in the era of advanced sequencing technologies. Annu. Rev. Genom. Hum. Genet. 18:229–56
33. Cherry JJ, Evans MC, Ni J, Cuny GD, Glicksman MA, Androphy EJ. 2012. Identification of novel compounds that increase SMN protein levels using an improved SMN2 reporter cell assay. J. Biomol. Screen. 17:481–95
34. Cherry JJ, Osman EY, Evans MC, Choi S, Xing X, et al. 2013. Enhancement of SMN protein levels in a mouse model of spinal muscular atrophy using novel drug-like compounds. EMBO Mol. Med. 5:1103–18
35. Cho S, Dreyfuss G. 2010. A degron created by SMN2 exon 7 skipping is a principal contributor to spinal muscular atrophy severity. Genes Dev. 24:438–42
36. Chong CR, Chen X, Shi L, Liu JO, Sullivan DJ Jr. 2006. A clinical drug library screen identifies astemizole as an antimalarial agent. Nat. Chem. Biol. 2:415–16
37. Chopra V, Fox JH, Lieberman G, Dorsey K, Matson W, et al. 2007. A small-molecule therapeutic lead for Huntington’s disease: preclinical pharmacology and efficacy of C2-8 in the R6/2 transgenic mouse. PNAS 104:16685–89
38. Clancy JP, Rowe SM, Accurso FJ, Aitken ML, Amin RS, et al. 2012. Results of a phase IIa study of VX-809, an investigational CFTR corrector compound, in subjects with cystic fibrosis homozygous for the F508del-CFTR mutation. Thorax 67:12–18
39. Coelho T, Merlini G, Bulawa CE, Fleming JA, Judge DP, et al. 2016. Mechanism of action and clinical application of tafamidis in hereditary transthyretin amyloidosis. Neurol. Ther. 5:1–25
40. Cox AD, Fesik SW, Kimmelman AC, Luo J, Der CJ. 2014. Drugging the undruggable RAS: mission possible? Nat. Rev. Drug Discov. 13:828–51
41. De Matteis MA, Vicinanza M, Venditti R, Wilson C. 2013. Cellular assays for drug discovery in genetic disorders of intracellular trafficking. Annu. Rev. Genom. Hum. Genet. 14:159–90
42. Deininger M, Buchdunger E, Druker BJ. 2005. The development of imatinib as a therapeutic agent for chronic myeloid leukemia. Blood 105:2640–53
43. Delvecchio C, Tiefenbach J, Krause HM. 2011. The zebrafish: a powerful platform for in vivo, HTS drug discovery. Assay Drug Dev. Technol. 9:354–61
44. Desai UA, Pallos J, Ma AA, Stockwell BR, Thompson LM, et al. 2006. Biologically active molecules that reduce polyglutamine aggregation and toxicity. Hum. Mol. Genet. 15:2114–24
45. Dillon KJ, Smith GC, Martin NM. 2003. A FlashPlate assay for the identification of PARP-1 inhibitors.
J. Biomol. Screen. 8:347–52
46. Dolma S, Lessnick SL, Hahn WC, Stockwell BR. 2003. Identification of genotype-selective antitumor agents using synthetic lethal chemical screening in engineered human tumor cells. Cancer Cell 3:285–96
47. Drilon A, Somwar R, Wagner JP, Vellore NA, Eide CA, et al. 2016. A novel crizotinib-resistant solvent- front mutation responsive to cabozantinib therapy in a patient with ROS1-rearranged lung cancer. Clin. Cancer Res. 22:2351–58
48. Druker BJ, Tamura S, Buchdunger E, Ohno S, Segal GM, et al. 1996. Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells. Nat. Med. 2:561–66
49. Du L, Damoiseaux R, Nahas S, Gao K, Hu H, et al. 2009. Nonaminoglycoside compounds induce readthrough of nonsense mutations. J. Exp. Med. 206:2285–97
www.annualreviews.org • Small-Molecule Drug Discovery 15.21

50. Du L, Jung ME, Damoiseaux R, Completo G, Fike F, et al. 2013. A new series of small molecular weight compounds induce read through of all three types of nonsense mutations in the ATM gene. Mol. Ther. 21:1653–60
51. Erlanson DA, Fesik SW, Hubbard RE, Jahnke W, Jhoti H. 2016. Twenty years on: the impact of fragments on drug discovery. Nat. Rev. Drug Discov. 15:605–19
52. Erlanson DA, Jahnke W, eds. 2016. Fragment-Based Drug Discovery: Lessons and Outlook. Weinheim, Ger.:
Wiley-VCH
53. Farmer H, McCabe N, Lord CJ, Tutt AN, Johnson DA, et al. 2005. Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature 434:917–21
54. Ferrins L, Pollastri MP. 2018. The importance of collaboration between industry, academics, and non-
profits in tropical disease drug discovery. ACS Infect Dis. 4:445–48
55. Freeman AK, Ritt DA, Morrison DK. 2013. Effects of Raf dimerization and its inhibition on normal and disease-associated Raf signaling. Mol. Cell 49:751–58
56. Friend C, Scher W, Holland JG, Sato T. 1971. Hemoglobin synthesis in murine virus-induced leukemic
cells in vitro: stimulation of erythroid differentiation by dimethyl sulfoxide. PNAS 68:378–82
57. Fuentealba RA, Marasa J, Diamond MI, Piwnica-Worms D, Weihl CC. 2012. An aggregation sensing reporter identifies leflunomide and teriflunomide as polyglutamine aggregate inhibitors. Hum. Mol. Genet. 21:664–80
58. Gambacorti-Passerini C. 2008. Part I: milestones in personalised medicine—imatinib. Lancet Oncol. 9:600
59. Garbaccio RM, Parmee ER. 2016. The impact of chemical probes in drug discovery: a pharmaceutical industry perspective. Cell Chem. Biol. 23:10–17
60. Gintjee TJ, Magh AS, Bertoni C. 2014. High throughput screening in Duchenne muscular dystrophy:
from drug discovery to functional genomics. Biology 3:752–80
61. Goldin E, Zheng W, Motabar O, Southall N, Choi JH, et al. 2012. High throughput screening for small molecule therapy for Gaucher disease using patient tissue as the source of mutant glucocerebrosidase. PLOS ONE 7:e29861
62. Griffin A, Hamling KR, Knupp K, Hong S, Lee LP, Baraban SC. 2017. Clemizole and modulators of
serotonin signalling suppress seizures in Dravet syndrome. Brain 140:669–83
63. Harvey AL, Edrada-Ebel R, Quinn RJ. 2015. The re-emergence of natural products for drug discovery in the genomics era. Nat. Rev. Drug Discov. 14:111–29
64. Heiser V, Engemann S, Brocker W, Dunkel I, Boeddrich A, et al. 2002. Identification of benzothiazoles
as potential polyglutamine aggregation inhibitors of Huntington’s disease by using an automated filter retardation assay. PNAS 99(Suppl. 4):16400–6
65. Hu Y, Wu B, Zillmer A, Lu P, Benrashid E, et al. 2010. Guanine analogues enhance antisense
oligonucleotide-induced exon skipping in dystrophin gene in vitro and in vivo. Mol. Ther. 18:812–18
66. Irwin JJ, Duan D, Torosyan H, Doak AK, Ziebart KT, et al. 2015. An aggregation advisor for ligand discovery. J. Med. Chem. 58:7076–87
67. Janzen WP. 2014. Screening technologies for small molecule discovery: the state of the art. Chem. Biol.
21:1162–70
68. Johannes JW, Almeida L, Daly K, Ferguson AD, Grosskurth SE, et al. 2015. Discovery of AZ0108, an orally bioavailable phthalazinone PARP inhibitor that blocks centrosome clustering. Bioorg. Med. Chem. Lett. 25:5743–47
69. Kaelin WG Jr. 2005. The concept of synthetic lethality in the context of anticancer therapy. Nat. Rev.
Cancer 5:689–98
70. Kaniskan HU, Konze KD, Jin J. 2015. Selective inhibitors of protein methyltransferases. J. Med. Chem.
58:1596–629
71. Karoulia Z, Gavathiotis E, Poulikakos PI. 2017. New perspectives for targeting RAF kinase in human cancer. Nat. Rev. Cancer 17:676–91
72. Katayama R, Kobayashi Y, Friboulet L, Lockerman EL, Koike S, et al. 2015. Cabozantinib overcomes
crizotinib resistance in ROS1 fusion-positive cancer. Clin. Cancer Res. 21:166–74
73. Kendall GC, Mokhonova EI, Moran M, Sejbuk NE, Wang DW, et al. 2012. Dantrolene enhances antisense-mediated exon skipping in human and mouse models of Duchenne muscular dystrophy. Sci. Transl. Med. 4:164ra60

15.22 Markossian et al.

74. Ketley A, Chen CZ, Li X, Arya S, Robinson TE, et al. 2014. High-content screening identifies small molecules that remove nuclear foci, affect MBNL distribution and CELF1 protein levels via a PKC- independent pathway in myotonic dystrophy cell lines. Hum. Mol. Genet. 23:1551–62
75. Knowles MR, Drumm M. 2012. The influence of genetics on cystic fibrosis phenotypes. Cold Spring Harb. Perspect. Med. 2:a009548
76. Langenau DM, ed. 2016. Cancer and Zebrafish: Mechanisms, Techniques, and Models. Cham, Switz.: Springer
77. Lefebvre S, Burglen L, Reboullet S, Clermont O, Burlet P, et al. 1995. Identification and characterization of a spinal muscular atrophy-determining gene. Cell 80:155–65
78. Liu J, Zhou Y, Qi X, Chen J, Chen W, et al. 2017. CRISPR/Cas9 in zebrafish: an efficient combination for human genetic diseases modeling. Hum. Genet. 136:1–12
79. Lord CJ, Ashworth A. 2017. PARP inhibitors: synthetic lethality in the clinic. Science 355:1152–58
80. Lorson CL, Hahnen E, Androphy EJ, Wirth B. 1999. A single nucleotide in the SMN gene regulates splicing and is responsible for spinal muscular atrophy. PNAS 96:6307–11
81. Lunn MR, Stockwell BR. 2005. Chemical genetics and orphan genetic diseases. Chem. Biol. 12:1063–73
82. Ma CX, Cai S, Li S, Ryan CE, Guo Z, et al. 2012. Targeting Chk1 in p53-deficient triple-negative breast cancer is therapeutically beneficial in human-in-mouse tumor models. J. Clin. Investig. 122:1541–52
83. Macarron R, Banks MN, Bojanic D, Burns DJ, Cirovic DA, et al. 2011. Impact of high-throughput screening in biomedical research. Nat. Rev. Drug Discov. 10:188–95
84. Marks PA, Breslow R. 2007. Dimethyl sulfoxide to vorinostat: development of this histone deacetylase inhibitor as an anticancer drug. Nat. Biotechnol. 25:84–90
85. Martinez Molina D, Nordlund P. 2016. The cellular thermal shift assay: a novel biophysical assay for in situ drug target engagement and mechanistic biomarker studies. Annu. Rev. Pharmacol. Toxicol. 56:141–61
86. Mathews Griner LA, Guha R, Shinn P, Young RM, Keller JM, et al. 2014. High-throughput combi- natorial screening identifies drugs that cooperate with ibrutinib to kill activated B-cell-like diffuse large B-cell lymphoma cells. PNAS 111:2349–54
87. McMahon B, Aflaki E, Sidransky E. 2016. Chaperoning glucocerebrosidase: a therapeutic strategy for both Gaucher disease and Parkinsonism. Neural Regen. Res. 11:1760–61
88. Menear KA, Adcock C, Boulter R, Cockcroft XL, Copsey L, et al. 2008. 4-[3-(4- cyclopropanecarbonylpiperazine-1-carbonyl)-4-fluorobenzyl]-2H-phthalazin-1-one: a novel bioavail- able inhibitor of poly(ADP-ribose) polymerase-1. J. Med. Chem. 51:6581–91
89. Mistry PK, Lopez G, Schiffmann R, Barton NW, Weinreb NJ, Sidransky E. 2017. Gaucher disease: progress and ongoing challenges. Mol. Genet. Metab. 120:8–21
90. Moffat JG, Rudolph J, Bailey D. 2014. Phenotypic screening in cancer drug discovery—past, present and future. Nat. Rev. Drug Discov. 13:588–602
91. Moorwood C, Khurana TS. 2013. Duchenne muscular dystrophy drug discovery—the application of utrophin promoter activation screening. Expert Opin. Drug Discov. 8:569–81
92. Mullard A. 2017. 2016 FDA drug approvals. Nat. Rev. Drug Discov. 16:73–76
93. Nakajima H, Kim YB, Terano H, Yoshida M, Horinouchi S. 1998. FR901228, a potent antitumor antibiotic, is a novel histone deacetylase inhibitor. Exp. Cell Res. 241:126–33
94. Naryshkin NA, Weetall M, Dakka A, Narasimhan J, Zhao X, et al. 2014. SMN2 splicing modifiers improve motor function and longevity in mice with spinal muscular atrophy. Science 345:688–93
95. Nierobisz LS, Cheatham B, Buehrer BM, Sexton JZ. 2013. High-content screening of human primary muscle satellite cells for new therapies for muscular atrophy/dystrophy. Curr. Chem. Genom. Transl. Med. 7:21–29
96. O’Leary DA, Sharif O, Anderson P, Tu B, Welch G, et al. 2009. Identification of small molecule and genetic modulators of AON-induced dystrophin exon skipping by high-throughput screening. PLOS ONE 4:e8348
97. O’Neil NJ, Bailey ML, Hieter P. 2017. Synthetic lethality and cancer. Nat. Rev. Genet. 18:613–23
98. Ong T, Ramsey BW. 2016. New therapeutic approaches to modulate and correct cystic fibrosis trans- membrane conductance regulator. Pediatr. Clin. N. Am. 63:751–64
99. Ottesen EW. 2017. ISS-N1 makes the first FDA-approved drug for spinal muscular atrophy. Transl.
Neurosci. 8:1–6

www.annualreviews.org • Small-Molecule Drug Discovery 15.23

100. Palacino J, Swalley SE, Song C, Cheung AK, Shu L, et al. 2015. SMN2 splice modulators enhance U1-pre-mRNA association and rescue SMA mice. Nat. Chem. Biol. 11:511–17
101. Pattenden SG, Simon JM, Wali A, Jayakody CN, Troutman J, et al. 2016. High-throughput small molecule screen identifies inhibitors of aberrant chromatin accessibility. PNAS 113:3018–23
102. Paul SM, Mytelka DS, Dunwiddie CT, Persinger CC, Munos BH, et al. 2010. How to improve R&D productivity: the pharmaceutical industry’s grand challenge. Nat. Rev. Drug Discov. 9:203–14
103. Plantie E, Migocka-Patrzalek M, Daczewska M, Jagla K. 2015. Model organisms in the fight against muscular dystrophy: lessons from Drosophila and zebrafish. Molecules 20:6237–53
104. Plass C, Pfister SM, Lindroth AM, Bogatyrova O, Claus R, Lichter P. 2013. Mutations in regulators of the epigenome and their connections to global chromatin patterns in cancer. Nat. Rev. Genet. 14:765–80
105. Pollak M. 2014. Overcoming drug development bottlenecks with repurposing: repurposing biguanides to target energy metabolism for cancer treatment. Nat. Med. 20:591–93
106. Pollitt SK, Pallos J, Shao J, Desai UA, Ma AA, et al. 2003. A rapid cellular FRET assay of polyglutamine aggregation identifies a novel inhibitor. Neuron 40:685–94
107. Poulikakos PI, Persaud Y, Janakiraman M, Kong X, Ng C, et al. 2011. RAF inhibitor resistance is mediated by dimerization of aberrantly spliced BRAF(V600E). Nature 480:387–90
108. Poulikakos PI, Zhang C, Bollag G, Shokat KM, Rosen N. 2010. RAF inhibitors transactivate RAF dimers and ERK signalling in cells with wild-type BRAF. Nature 464:427–30
109. Pruss RM, Giraudon-Paoli M, Morozova S, Berna P, Abitbol JL, Bordet T. 2010. Drug discovery and development for spinal muscular atrophy: lessons from screening approaches and future challenges for clinical development. Future Med. Chem. 2:1429–40
110. Ramsey BW, Davies J, McElvaney NG, Tullis E, Bell SC, et al. 2011. A CFTR potentiator in patients with cystic fibrosis and the G551D mutation. N. Engl. J. Med. 365:1663–72
111. Ramsey BW, Welsh MJ. 2017. Progress along the pathway of discovery leading to treatment and cure of cystic fibrosis. Am. J. Respir. Crit. Care Med. 195:1092–99
112. Reaper PM, Griffiths MR, Long JM, Charrier JD, MacCormick S, et al. 2011. Selective killing of ATM- or p53-deficient cancer cells through inhibition of ATR. Nat. Chem. Biol. 7:428–30
113. Richon VM, Emiliani S, Verdin E, Webb Y, Breslow R, et al. 1998. A class of hybrid polar inducers of transformed cell differentiation inhibits histone deacetylases. PNAS 95:3003–7
114. Richon VM, Webb Y, Merger R, Sheppard T, Jursic B, et al. 1996. Second generation hybrid polar compounds are potent inducers of transformed cell differentiation. PNAS 93:5705–8
115. Schindelin J, Rueden CT, Hiner MC, Eliceiri KW. 2015. The ImageJ ecosystem: an open platform for biomedical image analysis. Mol. Reprod. Dev. 82:518–29
116. Schirle M, Jenkins JL. 2016. Identifying compound efficacy targets in phenotypic drug discovery. Drug Discov. Today 21:82–89
117. Schreiber SL, Kotz JD, Li M, Aube J, Austin CP, et al. 2015. Advancing biological understanding and therapeutics discovery with small-molecule probes. Cell 161:1252–65
118. Senger MR, Fraga CA, Dantas RF, Silva FP Jr. 2016. Filtering promiscuous compounds in early drug discovery: Is it a good idea? Drug Discov. Today 21:868–72
119. Shannon KM, Fraint A. 2015. Therapeutic advances in Huntington’s disease. Mov. Disord. 30:1539–46
120. Sharma J, Pareek V, Liu H, Cheng H. 2016. Emerging treatment for ALK-positive lung cancer. Expert Opin. Emerg. Drugs 21:147–55
121. Shimizu-Motohashi Y, Miyatake S, Komaki H, Takeda S, Aoki Y. 2016. Recent advances in innovative therapeutic approaches for Duchenne muscular dystrophy: from discovery to clinical trials. Am. J. Transl. Res. 8:2471–89
122. Shiryaev SA, Mesci P, Pinto A, Fernandes I, Sheets N, et al. 2017. Repurposing of the anti-malaria drug chloroquine for Zika virus treatment and prophylaxis. Sci. Rep. 7:15771
123. Shoichet BK. 2006. Screening in a spirit haunted world. Drug Discov. Today 11:607–15
124. Sittampalam GS, Coussens NP, Brimacombe K, Grossman A, Arkin M, eds. 2017. Assay Guid- ance Manual. Bethesda, MD: Eli Lilly and Natl. Cent. Adv. Transl. Sci. https://www.ncbi.nlm.nih. gov/books/NBK53196

15.24 Markossian et al.

125. Sleigh JN, Buckingham SD, Esmaeili B, Viswanathan M, Cuppen E, et al. 2011. A novel Caenorhabditis elegans allele, smn-1(cb131), mimicking a mild form of spinal muscular atrophy, provides a convenient drug screening platform highlighting new and pre-approved compounds. Hum. Mol. Genet. 20:245–60
126. Specht EA, Braselmann E, Palmer AE. 2017. A critical and comparative review of fluorescent tools for live-cell imaging. Annu. Rev. Physiol. 79:93–117
127. Swinney DC. 2013. The contribution of mechanistic understanding to phenotypic screening for first- in-class medicines. J. Biomol. Screen. 18:1186–92
128. Thorne N, Shen M, Lea WA, Simeonov A, Lovell S, et al. 2012. Firefly luciferase in chemical biology: a compendium of inhibitors, mechanistic evaluation of chemotypes, and suggested use as a reporter. Chem. Biol. 19:1060–72
129. Tinsley JM, Fairclough RJ, Storer R, Wilkes FJ, Potter AC, et al. 2011. Daily treatment with SMTC1100, a novel small molecule utrophin upregulator, dramatically reduces the dystrophic symptoms in the mdx mouse. PLOS ONE 6:e19189
130. Tiscornia G, Vivas EL, Matalonga L, Berniakovich I, Barragan Monasterio M, et al. 2013. Neuronopathic Gaucher’s disease: induced pluripotent stem cells for disease modelling and testing chaperone activity of small compounds. Hum. Mol. Genet. 22:633–45
131. Titus SA, Southall N, Marugan J, Austin CP, Zheng W. 2012. High-throughput multiplexed quantitation of protein aggregation and cytotoxicity in a Huntington’s disease model. Curr. Chem. Genom. 6:79–86
132. Torre BG, Albericio F. 2017. The pharmaceutical industry in 2016. An analysis of FDA drug approvals from a perspective of the molecule type. Molecules 22:368
133. Tsai J, Lee JT, Wang W, Zhang J, Cho H, et al. 2008. Discovery of a selective inhibitor of oncogenic B-Raf kinase with potent antimelanoma activity. PNAS 105:3041–46
134. Ueda H, Manda T, Matsumoto S, Mukumoto S, Nishigaki F, et al. 1994. FR901228, a novel antitu- mor bicyclic depsipeptide produced by Chromobacterium violaceum no. 968. III. Antitumor activities on experimental tumors in mice. J. Antibiot. 47:315–23
135. Ueda H, Nakajima H, Hori Y, Fujita T, Nishimura M, et al. 1994. FR901228, a novel antitumor bicyclic depsipeptide produced by Chromobacterium violaceum no. 968. I. Taxonomy, fermentation, isolation, physico-chemical and biological properties, and antitumor activity. J. Antibiot. 47:301–10
136. Ueda H, Nakajima H, Hori Y, Goto T, Okuhara M. 1994. Action of FR901228, a novel antitumor bicyclic depsipeptide produced by Chromobacterium violaceum no. 968, on Ha-ras transformed NIH3T3 cells. Biosci. Biotechnol. Biochem. 58:1579–83
137. Valenzano KJ, Khanna R, Powe AC, Boyd R, Lee G, et al. 2011. Identification and characterization of pharmacological chaperones to correct enzyme deficiencies in lysosomal storage disorders. Assay Drug Dev. Technol. 9:213–35
138. Van Goor F, Hadida S, Grootenhuis PD, Burton B, Cao D, et al. 2009. Rescue of CF airway epithelial cell function in vitro by a CFTR potentiator, VX-770. PNAS 106:18825–30
139. Van Goor F, Hadida S, Grootenhuis PD, Burton B, Stack JH, et al. 2011. Correction of the F508del- CFTR protein processing defect in vitro by the investigational drug VX-809. PNAS 108:18843–48
140. Van Goor F, Straley KS, Cao D, Gonzalez J, Hadida S, et al. 2006. Rescue of ∆F508-CFTR trafficking
and gating in human cystic fibrosis airway primary cultures by small molecules. Am. J. Physiol. Lung Cell Mol. Physiol. 290:L1117–30
141. Van Rossum A, Holsopple M. 2016. Enzyme replacement or substrate reduction? A review of Gaucher disease treatment options. Hosp. Pharm. 51:553–63
142. Varma H, Voisine C, DeMarco CT, Cattaneo E, Lo DC, et al. 2007. Selective inhibitors of death in mutant huntingtin cells. Nat. Chem. Biol. 3:99–100
143. Wang J, Gines S, MacDonald ME, Gusella JF. 2005. Reversal of a full-length mutant huntingtin neuronal cell phenotype by chemical inhibitors of polyglutamine-mediated aggregation. BMC Neurosci. 6:1
144. Wang W, Duan W, Igarashi S, Morita H, Nakamura M, Ross CA. 2005. Compounds blocking mutant huntingtin toxicity identified using a Huntington’s disease neuronal cell model. Neurobiol. Dis. 20:500–8
145. Welch EM, Barton ER, Zhuo J, Tomizawa Y, Friesen WJ, et al. 2007. PTC124 targets genetic disorders caused by nonsense mutations. Nature 447:87–91

www.annualreviews.org • Small-Molecule Drug Discovery 15.25

146. Woll MG, Qi H, Turpoff A, Zhang N, Zhang X, et al. 2016. Discovery and optimization of small molecule splicing modifiers of survival motor neuron 2 as a treatment for spinal muscular atrophy. J. Med. Chem. 59:6070–85
147. Xiao J, Marugan JJ, Zheng W, Titus S, Southall N, et al. 2011. Discovery, synthesis, and biological
evaluation of novel SMN protein modulators. J. Med. Chem. 54:6215–33
148. Yagoda N, von Rechenberg M, Zaganjor E, Bauer AJ, Yang WS, et al. 2007. RAS-RAF-MEK-dependent oxidative cell death involving voltage-dependent anion channels. Nature 447:864–68
149. Zhang C, Bollag G. 2010. Scaffold-based design of kinase inhibitors for cancer therapy. Curr. Opin.
Genet. Dev. 20:79–86
150. Zhang JH, Chung TD, Oldenburg KR. 1999. A simple statistical parameter for use in evaluation and validation of high throughput screening assays. J. Biomol. Screen. 4:67–73
151. Zhang N, Bailus BJ, Ring KL, Ellerby LM. 2016. iPSC-based drug screening for Huntington’s disease.
Brain Res. 1638:42–56
152. Zhang X, Smith DL, Meriin AB, Engemann S, Russel DE, et al. 2005. A potent small molecule inhibits polyglutamine aggregation in Huntington’s disease neurons and suppresses neurodegeneration in vivo. PNAS 102:892–97
153. Zimmermann J, Caravatti G, Mett H, Meyer T, Muller M, et al. 1996. Phenylamino-pyrimidine (PAP)
derivatives: a new class of potent and selective inhibitors of protein kinase C (PKC). Arch. Pharm.
329:371–76

15.26 Markossian et al.