Sage Journals: Discover world-class research

Abstract

A diverse range of biochemical and cellular assays are used by medicinal chemists to guide compound optimization. The data collected from these assays influence decisions taken on structure-activity relationship (SAR) campaigns. Therefore, it is paramount that medicinal chemists have a solid understanding of the strengths and limitations of each assay being used to characterize synthesized analogs. For the successful execution of a medicinal chemistry campaign, it is our contention that an early partnership among assay biologists, informaticians, and medicinal chemists must exist. Their combined skill sets are necessary to not only design and develop robust assays but also implement an effective screening cascade in which multiple orthogonal and counter assays are selected to validate the activity and target(s) of the synthesized compounds. We review multiple cases of drug and chemical probe discovery from collaborative National Center for Advancing Translational Sciences/National Institutes of Health projects and published scientific literature in which the evaluation of compounds in secondary or orthogonal assays led to the discovery of unexpected activities, forcing a reconsideration of the original assay design that was used to discover the biological activity of the compound. Using these retrospective case studies, the goal of this Perspective is to hedge toward the development of physiologically relevant assays that are able to capture the true bioactivity of compounds being developed in a medicinal chemistry campaign.

Graphical Abstract

Keywords

assay design drug discovery medicinal chemistry

Introduction

Medicinal chemistry is a core component of research conducted at the Division of Preclinical Innovation (DPI) within the National Center for Advancing Translational Sciences (NCATS) at the National Institutes of Health (NIH). Within the DPI, medicinal chemists have the opportunity to collaborate on multiple projects with assay biologists and informaticians. Most projects originate as collaborations with external investigators who possess deep understandings of specific biological targets, pathways, and disease pathologies. Such collaborations often involve high-throughput screening (HTS) followed by medicinal chemistry optimization of hits to discover small-molecule probes that can be used in proof-of-concept experiments. While domain expertise, especially in synthetic organic chemistry, is a prerequisite for medicinal chemists to conduct translational science, we would like chemists to consider another important characteristic of a translational scientist: boundary crossing, especially toward an understanding of assay biology.¹ This Perspective consists of some high-yield case studies from both NCATS DPI and the literature in which evaluation in orthogonal assays produced unexpected activities of compounds. These examples highlight the importance of assay design and establishing a robust screening tree for medicinal chemists to confirm the bioactivity of their compounds.

Compound Activities Can Depend on Enzyme Source and Substrate Choice

Cell-free enzymatic assays that use recombinant protein and fluorogenic substates remain a mainstay of HTS and medicinal chemistry-based optimization. At NCATS, we have adapted and miniaturized a multitude of such assays to discover small-molecule modulators of lysosomal hydrolases implicated in rare disorders such as Gaucher, Pompei, Wolman, and Fabry disease. Gaucher disease (GD) is a recessive monogenic lysosomal disorder caused by the deficiency of glucocerebrosidase (GCase) that results in excessive accumulation of glucocerebrosides or glycosphingolipids, substrates for GCase in lysosomes ( Fig. 1A ). N370S is the most common missense mutation that occurs in type 1 GD patients, causing misfolding and degradation of the protein via the ER-associated protein pathway. Iminosugars such as the reported GCase inhibitor isofagomine are substrate mimetics that can behave as pharmacologic chaperones; they purportedly bind to the protein, assisting in the folding process and increase trafficking of GCase to the lysosome.² Once in the lysosomes, the inhibitor must be replaced with natural glycosphingolipids so that the mutant enzyme, albeit catalytically less active than wild-type (WT) GCase, can process the stored lipids to ameliorate the disease phenotype. We developed a platform of assays that use artificial GCase substrates (4-methylumbelliferyl-β-D-glucopyranoside, resorufin β-D-glucopyranoside) that liberate a fluorophore in the blue or red region of the ultraviolet (UV) spectrum, respectively ( Fig. 1A ) and discovered non-iminosugar inhibitors as well as compounds that activate the hydrolysis of fluorogenic assay substrates.³

Figure 1.

Compound activity dependence on enzyme source and substrate choice. (A) Chemical structures of a typical glucocerebroside and substrates to monitor the biochemical activity of glucocerebrosidase (GCase). (B) Summary of high-throughput screening hit frequency with recombinant wild-type GCase and splenic N370S GCase. (C) Concentration-dependent activities of representative compounds from four chemotypes. Note differences in activities based on the selection of substrate and GCase source.

Although our initial screening campaign was conducted with WT GCase, we rationalized that it would be more prudent to screen against N370S GCase, the most prevalent mutant found in type 1 GD patients. With access to splenectomy samples from patients, we decided to evaluate spleen homogenates as a natural source of native N370S GCase. To compare screening modalities, we screened a set of 52K compounds versus recombinant WT GCase and splenic N370S GCase using 4-methylumbelliferyl-β-D-glucopyranoside as substrate.⁴ Our immediate observation was that we identified more hits with the WT enzyme, and most of these hits were robust (full inhibition curve with asymptotes) and partial inhibitors ( Fig. 1B ). The N370S spleen homogenate identified fewer hits but more activators. Developing activators presented an attractive therapeutic strategy because GD results from deficient GCase activity. We proposed that additive detergents such as sodium taurocholate that are required to activate recombinant enzyme lead to overactivated GCase in the assays and preclude further activation by small-molecule activators, a phenomenon that is avoided in the presence of natural activators such as saposin C in tissue homogenates.

With several active chemotypes in hand, we compared their activities using recombinant and splenic N370S GCase as well as both fluorogenic substrates and discovered some very unusual trends ( Fig. 1C ). A quinazoline series showed concentration-dependent inhibition with spleen homogenates but not with recombinant N370S GCase; oppositely, a sulfonamide series showed robust inhibition with recombinant N370S GCase and not with splenic GCase.⁵ A thiourea series showed inhibition in both formats.⁶ For the exemplar pyrazolo[1,5-a]pyrimidine activator, we observed activity primarily with the 4-methylumbelliferyl-β-D-glucopyranoside reporter and not the resorufin β-D-glucopyranoside reporter. Although activity for the pyrazolopyrimidine was more efficacious with the recombinant protein, we decided to use the spleen homogenate to optimize the chemical series with the rationale that mutant N370S GCase was more physiologically relevant. The lead compound from the series, ML198 was shown to increase GCase activity and clear glycosphingolipids from patient induced pluripotent stem cells (iPSCs) that were differentiated to macrophages (the primary cell type affected by type 1 GD), underscoring the importance of the screening assay that led to their discovery.^7–10

Although we have yet to unravel the structural basis of such discordant activities between recombinant enzyme and spleen homogenates, we believe more advanced disease-relevant high-throughout cellular assays are required to discover and verify the most relevant activities of each chemical series. This case highlights the importance of testing several sources of biological targets (recombinant, splenic) and target constructs (WT, mutant) in target-based drug discovery. Unexpected discrepancies should not be dismissed but rather thoroughly investigated.

Biochemical and Cellular Assays Can Show Different Compound Activities

Ever since the discovery of the specific Bcr-Abl inhibitor imatinib, scores of kinase inhibitors have been clinically approved for the treatment of cancers, with hundreds more in clinical trials.^11,12 Many kinase inhibitors evolve from HTS and structure-activity relationship (SAR) optimization campaigns that often test for inhibition of recombinant kinases in cell-free formats. However, the following example from NCATS shows how the kinase profiling activity of an inhibitor with cell-free kinases might be in stark contrast with kinase engagement in cells.

Duchenne muscular dystrophy (DMD) is a rare X-linked progressive muscular degenerative disease affecting 1:5000 boys, whereby mutations in the DMD gene lead to complete loss of the dystrophin protein that connects intracellular actin to the extracellular matrix in myofibrils. In one collaboration, NCATS screened for compounds that could upregulate α7β1 integrin, another structural transmembrane protein in myofibers to potentially compensate for the loss of dystrophin.¹³ A pilot screen in a murine myogenic-cultured cell line with Itga7+/LacZ (replacement of one opposing α7 integrin allele with a β-galactosidase reporter) identified SU9516 ( Fig. 2A ), a reported CDK2 inhibitor, that could increase integrin expression. SU9516 was also efficacious in an mdx mouse model of DMD, where it increased muscle function and reduced disease pathology.

Figure 2.

Profiling the kinase activity of SU9516. (A) Reaction mechanism of the KiNativ probe (adenosine triphosphate [ATP] acyl phosphate) with a lysine residue in the ATP binding site of CDK2. (B) Workflow of a KiNativ experiment. The lysate sample mixture is treated with/without inhibitor, followed by labeling with a probe (ATP acyl phosphate). Samples are captured with streptavidin and digested with trypsin, and probe-labeled peptides are eluted and analyzed by mass spectrometry. (C) Summary of SU9516 activities in human DMD myotubules by the KiNativ assay and the cell-free HotSpot assay.

A commercial platform technology (ActiveX Biosciences, San Diego, CA) was used to assess target engagement with the purported target CDK2 and other kinases in myotubes. This particular platform uses a KiNativ probe, a biotinylated adenosine triphosphate (ATP)–mimetic with an acylated biotinylated group at the third phosphate position ( Fig. 2A ), which binds irreversibly to conserved lysine residues at the outer pocket of most kinases. DMD patient myotubes were lysed and incubated with the probe; kinases incorporating the biotinylated substrate were enriched by pull down using streptavidin immobilization (streptavidin is a protein with a very high affinity for biotin with Kd~10^–15 M; it can be immobilized on solid support and used to purify/pull molecules attached to biotin). Subsequent sample proteolysis and liquid chromatography tandem mass spectrometry (LC-MS/MS) identified signature peptide fragments from approximately 300 kinases ( Fig. 2B ). When myotubes were pretreated with compound before lysis, a concentration-dependent reduction of the peptides associated with the kinase that is binding to the compound was observed. At 100 nM SU9516, there was approximately 45% binding to CDK2 in addition to 81% and 79% binding to OSR1 and STLK3, respectively.¹⁴ The authors postulated that inhibition of the stress-sensing SPAK/OSR1 kinase in DMD myocytes by SU9516 was the mechanism by which it downregulates p65-NF-κB activation downstream, reversing the DMD disease phenotype. Incidentally, kinase profiling for SU9516 against 300 recombinant kinases using the HotSpot (Reaction Biology, Malvern, PA) assay platform had already been reported. This assay quantifies the transfer of a radiolabeled γ-phosphate to a substrate.¹⁵ Even with 500 nM SU9516, there was minimal inhibition of OSR1 and STLK3, kinases that were engaged at 100 nM in myotubes in the cellular assay ( Fig. 2C ).

This case reiterates the importance of investigating compound activities in both cell-free and (potentially) the more physiologically relevant cellular assay systems. In certain instances, such as in the case of SU9516, target engagement in cells or tissues might show results discrepant to those obtained under cell-free experimental conditions. Similar observations have been made by others using the KiNativ platform.¹⁶

Multiplex Cellular Kinase Profiling Can Enable Hypotheses of Off-Target Inhibitor Effects

Kinobeads, another chemoproteomic platform to quantify small-molecule kinome engagement in cells, have also exposed disparity in kinome activity between cellular and cell-free assays.¹⁷ Kinobeads work via solid supports that are irreversibly attached to a library of promiscuous kinase inhibitors that bind to kinases present in cell or tissue lysates. After bead washing and proteolytic digestion, bound kinases are quantified based on the release of signature peptide fragments analyzed by LC-MS.¹⁸ In the presence of a competitive free inhibitor, less enzyme binds to the beads, leading to a concentration-dose–dependent reduction of signatures peptides(s) associated with free inhibitor bound kinases ( Fig. 3 ).

Figure 3.

Schematic of Kinobead technology. The assay enabled the characterization of off-target ferrochelatase activity from vemurafenib (test compound) and other kinase inhibitors in clinical development.

Because Kinobeads are affixed with ATP-competitive inhibitors, they also capture other nucleotide-binding proteins such as metabolic kinases and acetyl coenzyme A dehydrogenases via binding to the pyridoxal- or flavin adenine dinucleotide–binding sites, respectively. Using this platform, at least 40 advanced kinase inhibitors (including vermurafenib) were also shown to bind to the heme-synthesizing enzyme ferrochelatase (FECH).¹⁸ Vermurafenib ( Fig. 3 ), a potent and selective azaindole inhibitor of BRAF^V600E, was discovered in a fragment-based drug design SAR campaign that used a biochemical functional assay that monitored inhibition of phosphorylation of a fluorescence resonance energy transfer (FRET)–peptide substrate via a coupled kinase/protease enzymatic assay ( Suppl. Fig. S1 ).¹⁹ Several studies supported the initial safety of vermurafenib and its efficacy versus BRAF^V600E-driven melanomas.^20,21 However, a subset of patients exhibited cutaneous photosensitivity while taking vermurafenib. This had been first ascribed to vermurafenib binding to FECH, a discovery made while characterizing the thermal profiles of 7000 proteins with kinase inhibitors.²² When the same binding was rediscovered with Kinobeads ( Fig. 3C ), a subset of clinically approved kinase inhibitors that bound to FECH was further shown to inhibit FECH biochemical activity, inhibit downstream heme biosynthesis in K562 cells, and block cholate binding to the protoporphyrin binding site.^23,24 This suggested that FECH inhibition likely leads to elevated levels of protoporphyrin, similar to erythropoietic protoporphyria, a disorder of the heme metabolic pathway marked by skin photosensitivity. This undesirable coactivity against FECH in clinically approved kinase inhibitors creates an important metric by which kinase inhibitors should be screened before entering clinical trials.

Proteomics Enables Investigation of Ocular Toxicity in Alzheimer’s Disease

Activity-based protein profiling (ABPP) has been shown to be a versatile approach to gauge small-molecule target engagement, especially with proteins bearing nucleophilic residues needed for catalysis.²⁵ ABPP typically uses chemical probes that have promiscuous electrophilic warheads that covalently bind to the nucleophilic residue (e.g., serine hydroxyl in serine hydrolases or sulfhydryl groups in cysteine side chains). The warhead is attached to a reporter, often a fluorescent group or biotin, that helps in detection and identification of the protein. Such probes can also be designed from active small-molecule leads by decorating them strategically with functional groups such as diazirines, which generate light-activated reactive species for irreversible binding to their target. The following example showcases such a probe that was able to reveal off-target toxicity of several β-secretase 1 (BACE1) inhibitors due to cathepsin D (CatD) inhibition, a target that these compounds had previously been optimized against in biochemical assays with purified protein.

BACE1 (part of the pepsin aspartyl protease super family) has been regarded as a promising target for Alzheimer’s disease (AD) due to its role in catalyzing the breakdown of amyloid precursor protein (APP) to amyloid beta (Aβ), which accumulates as cerebral Aβ plaques in the brains of AD patients.²⁶ Although BACE1 inhibitors are yet to be approved by the U.S. Food and Drug Administration (FDA), several generations have been developed (e.g., LY2811376, AMG-8718, PF-9283; Fig. 4A ; Suppl. Fig. S2 ).^27–29 Both LY2811376 and AMG-8718 were dropped from clinical testing, as these inhibitors caused an accumulation of autofluorescent material in the retinal pigment epithelium (a nonregenerating layer of cells) in rats, eventually leading to severe visual impairment.³⁰ BACE1 knockout mice dosed with LY2811376 and PF-9283 still showed this toxicity, eluding to an off-target mechanism.^27,30,31 A suspected target of this toxicity was CatD, a related protease whose deficiency is correlated with visual impairment.^31–34 During the medicinal chemistry optimization of these compounds, FRET-based assays that analyzed cleavage of peptide substrates with purified aspartyl proteases showed >60-, >1000-, and >200-fold selectivities of BACE1 over CatD and CatE, respectively, for LY2811376, AMG-8718, and PF-9283^27,29,35,36 ( Suppl. Fig. S2 ). A fluorescent polarization (FP) assay using synthetic APP-biotinylated peptide as the substrate had confirmed PF-9283’s BACE activity. Although whole-cell assays had been developed and used to optimize BACE1 activity, such assays were unavailable for CatD.^37,38 This changed when Zuhl and coworkers³⁹ reported a clickable affinity probe that identified CatD as the culprit for the off-target ocular toxicity. They synthesized a closely related derivative of PF-9283, PF-7802 ( Fig. 4A ), which incorporates a benzophenone and alkyne handle for click chemistry. After observing that PF-7802 also inhibited sAPPβ formation (the direct product of BACE1 cleavage of APP) in cells with an IC₅₀ value of 320 nM, they introduced the probe to ARPE-19 cells (a cell line from the retinal pigment epithelium). After cell lysis, UV irradiation led to the covalent binding of the benzophenone to protein targets in its proximity. The alkyne handle was then subjected to a click reaction with TAMRA-azide ( Fig. 4B ), and thus, target proteins were labeled with the TAMRA fluorophore, enabling their visualization when separated by gel electrophoresis. Repeating the same experiment in the presence of PF-9283, they observed a loss in fluorescent signal on a 30 kDa band, indicating that it contained protein(s) being inhibited by PF-9283 (which outcompetes PF-7802 for binding). By replacing TAMRA with biotin, the bound proteins were pulled with streptavidin beads and digested, and using a combination of LC-MS and stable-isotope labeling by amino acids in cell culture,⁴⁰ it was established that CatD was indeed the target of these compounds ( Fig. 4B ). Genetic knockdown of CatD by small interfering RNA in cells also found that this PF-7802-labeled band disappeared, further corroborating this target. Using PF-7802, these authors found that LY2811376, AMG-8718, and PF-9283 all inhibited CatD with IC₅₀ values <300 nM in cells. Chronic dosing of PF-9283 at 80 mpk for 30 d in mice increased cat D levels in the mouse brain (via binding, stabilization, and downregulation of its degradation), demonstrating binding to CatD in vivo.

Figure 4.

Characterization of ocular toxicity of β-secretase (BACE) inhibitors by proteomics. (A) Chemical structures of BACE inhibitors and key assay reagents (TAMRA: tetramethylrhodamine flurophore) (B) Schematic of affinity probe in-gel fluorescence and stable-isotope labeling by amino acids in cell culture experiments.

This case study shows the benefit of cellular on-target validation using a complementary photoaffinity-labeling chemoproteomics approach. Although this case led to drugs being pulled from clinical development, it led to new assay development and toxicology standards that can be applied to future drug candidates operating in the same pathway.

OTS167 and Its Oncologic MELK-Independent Activity

Although not absolutely necessary for FDA approval, target deconvolution and a thorough understanding of a drug candidate mechanism of action can guide therapeutic use and predict potential toxicities.^41,42 In preclinical discovery programs, when a compound shows an expected phenotype in disease-relevant cells, a simple check for target validation would be to eliminate the putative target and check whether the compound still elicits the phenotype. In a retrospective study, Sheltzer et al.⁴³ sought to establish the role of maternal embryonic leucine zipper kinase (MELK) as a prognostic biomarker for cancers addicted to MELK. They chose the triple-negative breast cancer cell lines Cal51 and MDA-MB-231 and the melanoma cell line A375, which had been shown, via RNA interference experiments, to be dependent on MELK for their growth. Using CRISPR-Cas9 technology, they induced frameshift mutations in various regions of MELK to knock out the protein. After verifying by Western blot that MELK expression had been ablated, they were surprised to see that the rate of proliferation in these cells remained unaffected. Furthermore, the MELK inhibitor OST167 ( Suppl. Fig. S3 ), which had been progressed to several clinical trials, was able to inhibit cell proliferation in these cell lines in the absence of MELK. OTS167 had been discovered using an FP assay that measured the inhibition of kinase function using recombinant MELK, and it also inhibited the phosphorylation of MELK-dependent substrates in cells ( Suppl. Fig. S3 ).⁴⁴ A crystal structure of OTS167 within MELK has also been reported.⁴⁵ These facts point out that although OTS167 is a bona fide MELK inhibitor, it could be acting in a MELK-independent manner, especially in cancers such as triple-negative breast cancer, where it was being evaluated in the clinic.

Sheltzer et al.⁴⁶ expanded such mechanistic reexamination to 10 additional cancer drugs targeting six different targets that are in preclinical development and clinical trials. Among key considerations in choosing the targets was that they had no reported mutations that rendered the compounds ineffective. Unfortunately, their observations were the same: the compounds inhibited cancer cell proliferation with the same potencies in the absence of their target. They honed in on a T-LAK cell-originated protein kinase/PDZ binding kinase inhibitor OTS964 ( Suppl. Fig. S3 ). They grew HCT116 colorectal cancer cells, which have an increased mutation frequency rate, at a cytotoxic concentration of OTS964 and isolated 10 clones that were resistant to OTS964. Whole-exome sequencing on these clones revealed that all of them had all missense mutations in cyclin-dependent kinase 11B (CDK11B). Further studies pointed out that the presence of a glycine residue before the conserved DFG kinase motif might play a critical role in the specific binding and inhibition of CDK11B by OTS964.

These case studies stand as cautionary tales for the uncovering of a target and how even when the recognized pathway may prove noncanonical, phenotypic responses may remain. They highlight the importance of target investigation using modern genomic methods such as CRISPR/Cas9 and that RNA interference alone cannot be reliable validate a target. They show that target associations with phenotypes that are consistent with chemical and genetic perturbations are further strengthened by the identification of compound-resistant mutants.

Reporter Assays: PTC124 as a Potent Inhibitor of Firefly Luciferase

Because of tolerable expression and high sensitivity, luciferase and green fluorescent protein reporters are often used in high-throughput assays to tag proteins and measure their concentration, localization, and study signal transduction. Reporter genes that code for luminescent proteins can be inserted after the promoter region (the region of DNA where the transcription machinery binds before start of transcription of a gene) of a protein of interest or next to it. Luciferase enzymes need a substrate that consumes molecular oxygen and ATP to produce a product, adenosine diphosphate, and light emission, which is the output that is measured. For firefly luciferase (FLuc), the substrate and products are D-luciferin and oxyluciferin, respectively ( Fig. 5A ). When small molecules are being evaluated in luciferase reporter assays, there exists the potential for readout interference if such compounds modulate the function of the reporter enzyme (e.g., enzymatic inhibition or inhibitor-induced reporter stabilization). A well-known example is PTC124 (ataluren; Fig. 5A ), which was discovered with the use of a phenotypic FLuc-based assay designed to identify compounds that could promote ribosomal “read-through” of premature nonsense mutations present in a FLuc reporter gene.⁴⁷ In principle, compound-mediated translation read-through would produce the bioluminescent FLuc reporter, whose light output could then be quantified ( Fig. 5A ).⁴⁸ With nanomolar activity in this assay, PTC124 was progressed to clinical trials for DMD and cystic fibrosis (and others as well), diseases in which a group of patients have nonsense mutations in the DMD or CTFR gene that lead to premature termination of translation and degradation of the truncated polypeptides. At NCATS, a screen of 70K compounds in a FLuc biochemical assay revealed 71 FLuc inhibitors that contained a 3,5-diaryl-oxadiazole motif also found in PTC124 ( Fig. 5B ). In cell-based assays with WT luciferase, these compounds showed unexpected activation of the reporter signal, an observation that was explained by binding of the compound to the cellular reporter enzyme, leading to its stabilization and apparent enhancement of activity.

Figure 5.

Compound-mediated interferences in luciferase-based assays. (A) The compound PTC124 was originally reported to promote read-through of stop/nonsense codons during translation. In models likely used in its development, PTC124 and derivatives (PTC124-AMP) have also been shown to bind to firefly luciferase (Fluc) and inhibit its enzymatic activity and stabilize translated FLuc, confounding the interpretation of cell-based FLuc assays. (B) Left: Representative analogs of PTC124 (1a) with a 3,5-diaryl-oxadiazole show an apparent structure-activity relationship (SAR) in a cell-based FLuc assay and SAR in a counterscreen for cell-free FLuc inhibition, which is often referred to as a structure-interference relationship (SIR), as it confounds the cellular SAR on top. Right: Chemical structure of PTC124-AMP, a high-affinity multisubstrate adduct inhibitor (MAI) that was shown to form with PTC124 in FLuc’s substrate pocket.

In orthogonal assays in which the FLuc reporter was replaced with Renilla luciferase (RLuc), PTC124 and analogs failed to inhibit RLuc biochemically and to activate RLuc in cells, further supporting the hypothesis of FLuc inhibition. Introducing PTC124 to purified FLuc or RLuc increased the trypsinization half-life of FLuc but not RLuc, suggesting a stabilized E.I complex for PTC124 and FLuc that was further supported by thermal denaturation assays.⁴⁹ Although counterarguments were made in favor of PTC124 not acting in a FLuc-dependent manner, confirmation of the E.I complex was found through x-ray crystallography.^49,50 LC-MS experiments found that the reversible inhibitor responsible for the complexation was actually PTC124-AMP, a high-affinity multisubstrate inhibitor with a specific meta-carboxylate motif needed to bring about the potent activity ( Fig. 5B ).⁵¹ These later findings have led to considerable debate regarding the mechanism of action driving the readout in the original luciferase-based assay.^51–53 The case of PTC124 and related oxadiazoles illustrates this occasional dilemma in which SAR is confounded by compound-dependent assay interference, leading to so-called “structure-interference relationships,” or SIR^54–60 ( Fig. 5B ).

The search for luciferase reporters with reduced promiscuity toward chemical libraries, smaller size, and brighter signal has led to the discovery of RLuc, NanoLuciferase (NanoLuc), and the turbo luciferase (TurboLuc). Although a class of arylsulfonamides inhibit RLuc,⁵⁵ the most recent 16 kDa TurboLuc shows reduced frequencies of interference by low-molecular-weight compounds.⁵⁶ At NCATS, the NanoLuc reporter is widely used as a split luciferase in high-throughput cellular thermal shift assay (CETSA) experiments to investigate cellular target engagement. A small fragment of NanoLuc (11 amino acid peptide HiBiT, or a 15 amino acid variant 86b) is fused to proteins of interest and allowed to interact with small molecules. This is followed by the addition of the larger fragment of NanoLuc (11S or LgBiT) whose surface complements the surface formed by the smaller fragment; the reconstituted enzyme can then emit light on the addition of the substrate furimazine after heating, to quantify the amount of soluble protein ( Suppl. Fig. S4 ).^57,58 In principle, this format (enzyme complementation, NanoLuc enzyme, luminescence) reduces the likelihood of compound-dependent assay interferences.

G-Protein–Coupled Receptor: Ligand Bias and Activity Dependency on Assay Format

G-protein–coupled receptors (GPCRs) are attractive therapeutic targets for various diseases covering about 30% of all FDA-approved drugs.⁵⁹ These seven-transmembrane receptors bind to extracellular ligands, undergo conformational changes to activate a GTP-to-GDP exchange in an intracellular heterotrimeric G-protein (Gα, Gn, Gγ; Suppl. Fig. S5 ). The β,γ subunit then dissociates to initiate specific intracellular signaling pathways that produce different secondary messengers (cAMP, InsP₃, Ca²⁺); these can be assayed reproducibly in CHO/HEK cells stably transfected with the GPCR. GPCR-regulating kinases can also phosphorylate GPCRs on specific intracellular residues to increase their affinity for β-arrestins that bind to the GPCR initiating their internalization through endocytosis; a process that can also be assayed in stably transfected cells. In SAR campaigns, these assays allow the measurement of “biased agonism” or “functional selectivity” with the notion that selective activity in a specific pathway will lead to a therapeutic effect whereas modulation of another result in adverse side effects.⁶⁰ However, medicinal chemists should be aware that such bias may be implicit in the assay design and the specific cells being used to determine the pharmacology of analogs. The kinetics of ligand-binding and subsequent signaling may have a significant influence on ligand bias. In a study with agonists of the dopamine D2 receptor, the antipsychotic drug aripiprazole was shown to develop a bias toward inhibition of cAMP production over pERK1/2 phosphorylation, as the assay incubation time and temperature were changed from 15 min (25 °C) to 30 min (37 °C); this was attributed to its slow dissociation kinetics.⁶¹ A later study showed that association (k_on) and not dissociation kinetics (k_off), determined by competition assays with a fluorescent ligand, correlated well with extrapyramidal side effects associated with D2-antipsychotic drugs.⁶²

The pain and addiction field has also been influenced by the desire to develop µ-opioid receptor (MOR) agonists that selectively activate the G-protein (analgesia) and not the β-arrestin (abuse, constipation, and respiratory suppression) pathway, a concept that has been challenged by the failure of the G-protein–biased MOR agonists TRV130 to reduce abuse-related effects in rats. Recent work published at the NIH has shown that the galanin receptor forms heteromers with MOR and negatively regulates the adverse dopaminergic effects of opioids.63 These observations hint at the complex pharmacology associated in the discovery of small molecules to modulate GPCR function, the significance of which is further corroborated by the HEAL Initiative (Helping to End Addiction Long-term Initiative), a nationwide effort initiated by NIH to tackle the opioid crisis and develop nonaddictive pain medications.⁶⁴ The paucity of predictive animal models in pain has also thwarted the clinical translation of therapies. However, at times, a simple verification of compound activity in cells that express a rodent receptor may be pragmatic before advancing to a rodent model ( Fig. 6 ). In a campaign that optimized small-molecule agonists of the relaxin hormone receptor by increasing cAMP levels in HEK293 cells stably transfected with human RXFP1 (hRXP1), the lead molecule ML290 proved to be completely inactive in cells expressing mouse RXFP1 (mRXP1), despite 89% sequence homology between mRXFP1 and hRXFP1.⁶⁵ Using site-directed mutagenesis, sequential replacement of four differing amino acid residues in the ECL3 region of the mRXFP1 with those from hRXFP1 furnished a humanized mouse construct mRXFP1-M10M11 that was fully active toward the small-molecule lead agonist. Interestingly, the endogenous ligand relaxin activated all constructs irrespective of ECL3 mutations, which suggested that the small-molecule agonists were noncompetitive at the allosteric ECL3 site. This mutagenesis study helped develop a “humanized transgenic mouse model,” in which the desired hRXFP1 receptor was expressed in the mouse model in lieu of mRXFP1, enabling the testing of the small-molecule agonists in a “human-like” setting. ⁶⁶

Figure 6.

Influence of G-protein–coupled receptors target construct on probe activity. In a project targeting the relaxin hormone receptor, the activity of ML290 is dependent on the sequence of the ECL3 portion of human/mouse RXFP1 receptors. To create a more translationally relevant mouse model, a partially humanized mouse construct (right) was necessary to restore ML290 activity.

These examples highlight the complexities in predicting the therapeutic efficacy of GPCR-bound small molecules a priori based on their pharmacologic activity. Thus, it may be prudent to advance molecules with distinct pharmacologic profiles and adopt assays that use physiologically relevant cells with verified receptor expression and a disease phenotype that can be reversed.

Perspectives

In this Perspective, we summarize several case examples of an assay design that significantly affects medicinal chemistry efforts. There are several important, high-yield take-away messages from these series of vignettes for medicinal chemists and biologists alike. First, seemingly innocuous factors such as the substrate choice, enzyme source, and cell-free versus cellular assays can have significant impacts on the observed biological activity, and these choices can dramatically guide hit selection and SAR. Second, compounds can interfere with the assay readout and can confound data interpretation. This interference is less predictable in more complex assays and can even produce an apparent SAR that reflects compound-dependent interference (SIR). Third, these case studies demonstrate the importance of establishing multiple lines of evidence for informing hit picking and medicinal chemistry optimization. This can include confirmatory, orthogonal, and interference counterscreens to de-risk for interferences, lack of selectivity, and irreproducibility. Furthermore, multiple types of experiments are usually needed to support meaningful (potent, selective) cellular target engagement. Techniques such as proteomics (including ABPP), targeted genetic manipulation (CRISPR), and CETSA can be vital in this regard.

Scientific translation, including preclinical drug and probe discovery, is a team sport that is crucially dependent on crosstalk between multiple disciplines. Results, whether assay testing from biologists or new chemical analogs from chemists, should not be simply passed over the transom. On the contrary, the aforementioned vignettes should encourage medicinal chemists to engage with their biologist colleagues who characterize new analogs. Many of these cases studies are retrospective in nature—reexaminations that ultimately led to the recognition of the unexpected results. This is expected because translation of a basic discovery to a therapeutic is not linear, as often depicted. On the contrary, translation requires constant interactions of neighborhoods of disciplines and expertise, followed by revisitation and reevaluation of the approaches that are being used in a dynamic process.⁶⁷ Indeed, these cases demonstrate that scientific approaches (and assays) evolve as a project progresses.

To better recapitulate disease pathophysiology, in vitro assays are getting more complex. This is evident by the current focus on assays that use multiple types of cells, including patient-derived pluripotent stem cells and complex in vitro models such as organoids and tumor spheroids.^68–72 Advances in rapid high-throughout mass spectrometric analysis are also leading to label-free analysis.^73,74 However, we believe simple biochemical and cellular assays stand their ground in medicinal chemistry. Their reductionist approach is powerful for making decisions about SAR. The goal of this Perspective is to promote partnership and shared understanding about the advantages and limitations of the assays and, more importantly, to encourage scientists to share their learnings about success and failures in their approaches.

Supplemental Material

sj-pdf-1-jbx-10.1177_24725552211026238 – Supplemental material for The Impact of Assay Design on Medicinal Chemistry: Case Studies

Supplemental material, sj-pdf-1-jbx-10.1177_24725552211026238 for The Impact of Assay Design on Medicinal Chemistry: Case Studies by Joshua R. Born, Vinoth Kumar Chenniappan, Danielle P. Davis, Jayme L. Dahlin, Juan J. Marugan and Samarjit Patnaik in SLAS Discovery

Footnotes

Acknowledgements

Authors acknowledge the contributions from the Early Translational Branch of DPI NCATS and extramural collaborators that formed the basis for many of these case studies. Figures were created with BioRender.com and ChemDraw.

Supplemental material is available online with this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Joshua R. Born

References

Gilliland

C. T.

White

Gee

, et al. The Fundamental Characteristics of a Translational Scientist. ACS Pharmacol. Transl. Sci. 2019, 2, 213–216.

Khanna

Benjamin

E. R.

Pellegrino

, et al. The Pharmacological Chaperone Isofagomine Increases the Activity of the Gaucher Disease L444P Mutant form of β-Glucosidase. FEBS J. 2010, 277, 1618–1638.

Zheng

Padia

Urban

D. J.

, et al. Three Classes of Glucocerebrosidase Inhibitors Identified by Quantitative High-Throughput Screening Are Chaperone Leads for Gaucher Disease. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 13192–13197.

Goldin

Zheng

Motabar

, et al. High Throughput Screening for Small Molecule Therapy for Gaucher Disease Using Patient Tissue as the Source of Mutant Glucocerebrosidase. PLoS One 2012, 7, e29861.

Marugan

J. J.

Zheng

Motabar

, et al. Evaluation of Quinazoline Analogues as Glucocerebrosidase Inhibitors with Chaperone Activity. J. Med. Chem. 2011, 54, 1033–1058.

Marugan

J. J.

Huang

Motabar

, et al. Non-Iminosugar Glucocerebrosidase Small Molecule Chaperones. Med. Chem. Comm. 2012, 3, 56–60.

Aflaki

Stubblefield

B. K.

Maniwang

, et al. Macrophage Models of Gaucher Disease for Evaluating Disease Pathogenesis and Candidate Drugs. Sci. Transl. Med. 2014, 6, 240ra73–240ra73.

Aflaki

Borger

D. K.

Moaven

, et al. A New Glucocerebrosidase Chaperone Reduces α-Synuclein and Glycolipid Levels in iPSC-Derived Dopaminergic Neurons from Patients with Gaucher Disease and Parkinsonism. J. Neurosci. 2016, 36, 7441–7452.

Mazzulli

J. R.

Zunke

Tsunemi

, et al. Activation of β-Glucocerebrosidase Reduces Pathological α-Synuclein and Restores Lysosomal Function in Parkinson’s Patient Midbrain Neurons. J. Neurosci. 2016, 36, 7693–7706.

10.

Patnaik

Zheng

Choi

J. H.

, et al. Discovery, Structure–Activity Relationship, and Biological Evaluation of Noninhibitory Small Molecule Chaperones of Glucocerebrosidase. J. Med. Chem. 2012, 55, 5734–5748.

11.

Roskoski

Properties of FDA-Approved Small Molecule Protein Kinase Inhibitors: A 2020 Update. Pharmacol. Res. 2020, 152, 104609.

12.

Bournez

Carles

Peyrat

, et al. Comparative Assessment of Protein Kinase Inhibitors in Public Databases and in PKIDB. Molecules 2020, 25, 3226.

13.

LOPAC®1280—The Library of Pharmacologically Active Compounds. https://www.sigmaaldrich.com/life-science/cell-biology/bioactive-small-molecules/lopac1280-navigator.html (accessed 2021-01-21).

14.

Sarathy

Wuebbles

R. D.

Fontelonga

T. M.

, et al. SU9516 Increases α7β1 Integrin and Ameliorates Disease Progression in the mdx Mouse Model of Duchenne Muscular Dystrophy. Mol. Ther. 2017, 25, 1395–1407.

15.

Anastassiadis

Deacon

S. W.

Devarajan

, et al. Comprehensive Assay of Kinase Catalytic Activity Reveals Features of Kinase Inhibitor Selectivity. Nat. Biotechnol. 2011, 29, 1039–1045.

16.

Patricelli

M. P.

Nomanbhoy

T. K.

, et al. In Situ Kinase Profiling Reveals Functionally Relevant Properties of Native Kinases. Chem. Biol. 2011, 18, 699–710.

17.

Klaeger

Heinzlmeir

Wilhelm

, et al. The Target Landscape of Clinical Kinase Drugs. Science 2017, 358, eaan4368.

18.

Reinecke

Heinzlmeir

Wilhelm

, et al. Kinobeads: A Chemical Proteomic Approach for Kinase Inhibitor Selectivity Profiling and Target Discovery. In Target Discovery and Validation; Plowright AT, Ed.; Wiley, 2019; pp 97–130.

19.

Tsai

Lee

J. T.

Wang

, et al. Discovery of a Selective Inhibitor of Oncogenic B-Raf Kinase with Potent Antimelanoma Activity. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 3041–3046.

20.

Bollag

Hirth

Tsai

, et al. Clinical Efficacy of a RAF Inhibitor Needs Broad Target Blockade in BRAF-Mutant Melanoma. Nature 2010, 467, 596–599.

21.

Yang

Higgins

Kolinsky

, et al. RG7204 (PLX4032), a Selective BRAFV600E Inhibitor, Displays Potent Antitumor Activity in Preclinical Melanoma Models. Cancer Res. 2010, 70, 5518–5527.

22.

Savitski

M. M.

Reinhard

F. B.

Franken

, et al. Tracking Cancer Drugs in Living Cells by Thermal Profiling of the Proteome. Science 2014, 346, 1255784.

23.

C. K.

Dailey

H. A.

Rose

J. P.

, et al. The 2.0 A Structure of Human Ferrochelatase, the Terminal Enzyme of Heme Biosynthesis. Nat. Struct. Biol. 2001, 8, 156–160.

24.

Klaeger

Gohlke

Perrin

, et al. Chemical Proteomics Reveals Ferrochelatase as a Common Off-Target of Kinase Inhibitors. ACS Chem. Biol. 2016, 11, 1245–1254.

25.

Cravatt

B. F.

Wright

A. T.

Kozarich

J. W.

Activity-Based Protein Profiling: From Enzyme Chemistry to Proteomic Chemistry. Annu. Rev. Biochem. 2008, 77, 383–414.

26.

Karran

Mercken

De Strooper

The Amyloid Cascade Hypothesis for Alzheimer’s Disease: An Appraisal for the Development of Therapeutics. Nat. Rev Drug Discov. 2011, 10, 698–712.

27.

May

P. C.

Dean

R. A.

Lowe

S. L.

, et al. Robust Central Reduction of Amyloid-β in Humans with an Orally Available, Non-Peptidic β-Secretase Inhibitor. J. Neurosci. 2011, 31, 16507–16516.

28.

Fielden

M. R.

Werner

Jamison

J. A.

, et al. Retinal Toxicity Induced by a Novel β-secretase Inhibitor in the Sprague-Dawley Rat. Toxicol. Pathol. 2015, 43, 581–592.

29.

Brodney

M. A.

Beck

E. M.

Butler

C. R.

, et al. Utilizing Structures of CYP2D6 and BACE1 Complexes to Reduce Risk of Drug–Drug Interactions with a Novel Series of Centrally Efficacious BACE1 Inhibitors. J. Med. Chem. 2015, 58, 3223–3252.

30.

Mecklenburg

Schraermeyer

An Overview on the Toxic Morphological Changes in the Retinal Pigment Epithelium after Systemic Compound Administration. Toxicol. Pathol. 2007, 35, 252–267.

31.

Roberds

S. L.

Anderson

Basi

, et al. BACE Knockout Mice Are Healthy Despite Lacking the Primary β-Secretase Activity in Brain: Implications for Alzheimer’s Disease Therapeutics. Hum. Mol. Genet. 2001, 10, 1317–1324.

32.

Koike

Nakanishi

Saftig

, et al. Cathepsin D Deficiency Induces Lysosomal Storage with Ceroid Lipofuscin in Mouse CNS Neurons. J. Neurosci. 2000, 20, 6898–6906.

33.

Tyynelä

Sohar

Sleat

D. E.

, et al. A Mutation in the Ovine Cathepsin D Gene Causes a Congenital Lysosomal Storage Disease with Profound Neurodegeneration. EMBO J. 2000, 19, 2786–2792.

34.

Steinfeld

Reinhardt

Schreiber

, et al. Cathepsin D Deficiency Is Associated with a Human Neurodegenerative Disorder. Am. J. Hum. Genet. 2006, 78, 988–998.

35.

Dineen

T. A.

Chen

Cheng

A. C.

, et al. Inhibitors of β-Site Amyloid Precursor Protein Cleaving Enzyme (BACE1): Identification of (S)-7-(2-fluoropyridin-3-yl)-3-((3-methyloxetan-3-yl)ethynyl)-5’H-spiro[chromeno[2,3-b]pyridine-5,4’-oxazol]-2’-amine (AMG-8718). J. Med. Chem. 2014, 57, 9811–9831.

36.

Yang

H. C.

Chai

Mosior

, et al. Biochemical and Kinetic Characterization of BACE1: Investigation into the Putative Species-Specificity for Beta- and Beta’-Cleavage Sites by Human and Murine BACE1. J. Neurochem. 2004, 91, 1249–1259.

37.

Tung

C. H.

Bredow

Mahmood

, et al. Preparation of a cathepsin D Sensitive Near-Infrared Fluorescence Probe for Imaging. Bioconjug. Chem. 1999, 10, 892–896.

38.

Chen

C.-S.

Chen

W.-N.

Zhou

, et al. Probing the Cathepsin D Using a BODIPY FL-Pepstatin A: Applications in Fluorescence Polarization and Microscopy. J. Biochem. Biophys. Methods 2000, 42, 137–151.

39.

Zuhl

A. M.

Nolan

C. E.

Brodney

M. A.

, et al. Chemoproteomic Profiling Reveals That Cathepsin D Off-Target Activity Drives Ocular Toxicity of β-secretase inhibitors. Nat. Commun. 2016, 7, 13042.

40.

Mann

Functional and Quantitative Proteomics Using SILAC. Nat. Rev. Mol. Cell Biol. 2006, 7, 952–958.

41.

Harrison

R. K.

Phase II and Phase III Failures: 2013–2015. Nat. Rev. Drug Discov. 2016, 15, 817–818.

42.

Force

Kolaja

K. L.

Cardiotoxicity of Kinase Inhibitors: The Prediction and Translation of Preclinical Models to Clinical Outcomes. Nat. Rev. Drug Discov. 2011, 10, 111–126.

43.

Lin

Giuliano

C. J.

Sayles

N. M.

, et al. CRISPR/Cas9 Mutagenesis Invalidates a Putative Cancer Dependency Targeted in On-Going Clinical Trials. eLife 2017, 6.

44.

Sportsman

J. R.

Gaudet

E. A.

Boge

Immobilized Metal Ion Affinity-Based Fluorescence Polarization (IMAP): Advances in Kinase Screening. Assay Drug Dev. Technol. 2004, 2, 205–214.

45.

Cho

Y.-S.

Kang

Kim

, et al. The Crystal Structure of MPK38 in Complex with OTSSP167, an Orally Administrative MELK Selective Inhibitor. Biochem. Biophys. Res. Commun. 2014, 447, 7–11.

46.

Lin

Giuliano

C. J.

Palladino

, et al. Off-Target Toxicity Is a Common Mechanism of Action of Cancer Drugs Undergoing Clinical Trials. Sci. Transl. Med. 2019, 11, eaaw8412.

47.

Welch

E. M.

Barton

E. R.

Zhuo

, et al. PTC124 Targets Genetic Disorders Caused by Nonsense Mutations. Nature 2007, 447, 87–91.

48.

Thorne

Inglese

Auld

D. S.

Illuminating Insights into Firefly Luciferase and Other Bioluminescent Reporters Used in Chemical Biology. Chem. Biol. 2010, 17, 646–657.

49.

Auld

D. S.

Thorne

Maguire

W. F.

, et al. Mechanism of PTC124 Activity in Cell-Based Luciferase Assays of Nonsense Codon Suppression. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 3585–3590.

50.

Peltz

S. W.

Welch

E. M.

Jacobson

, et al. Nonsense Suppression Activity of PTC124 (ataluren). Proc. Natl. Acad. Sci. U.S.A. 2009, 106, E64–E65.

51.

Auld

D. S.

Lovell

Thorne

, et al. Molecular Basis for the High-Affinity Binding and Stabilization of Firefly Luciferase by PTC124. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 4878–4883.

52.

Roy

Friesen

W. J.

Tomizawa

, et al. Ataluren Stimulates Ribosomal Selection of Near-Cognate tRNAs to Promote Nonsense Suppression. Proc. Natl. Acad. Sci. U.S.A. 2016, 113, 12508–12513.

53.

McElroy

S. P.

Nomura

Torrie

L. S.

, et al. A Lack of Premature Termination Codon Read-Through Efficacy of PTC124 (Ataluren) in a Diverse Array of Reporter Assays. PLoS Biol. 2013, 11, e1001593.

54.

Dahlin

J. L.

Nissink

J. W. M.

Strasser

J. M.

, et al. PAINS in the Assay: Chemical Mechanisms of Assay Interference and Promiscuous Enzymatic Inhibition Observed during a Sulfhydryl-Scavenging HTS. J. Med. Chem. 2015, 58, 2091–2113.

55.

Cheng

K. C. C.

Inglese

A Coincidence Reporter-Gene System for High-Throughput Screening. Nat. Methods 2012, 9, 937–937.

56.

Auld

D. S.

Narahari

P.-i.

, et al. Characterization and Use of TurboLuc Luciferase as a Reporter for High-Throughput Assays. Biochemistry 2018, 57, 4700–4706.

57.

Martinez

N. J.

Asawa

R. R.

Cyr

M. G.

, et al. A Widely-Applicable High-Throughput Cellular Thermal Shift Assay (CETSA) Using Split Nano Luciferase. Sci. Rep. 2018, 8, 9472.

58.

Dixon

A. S.

Schwinn

M. K.

Hall

M. P.

, et al. NanoLuc Complementation Reporter Optimized for Accurate Measurement of Protein Interactions in Cells. ACS Chem. Biol. 2016, 11, 400–408.

59.

Smith

J. S.

Lefkowitz

R. J.

Rajagopal

Biased Signalling: From Simple Switches to Allosteric Microprocessors. Nat. Rev. Drug Discov. 2018, 17, 243–260.

60.

Franco

Aguinaga

Jiménez

, et al. Biased Receptor Functionality versus Biased Agonism in G-Protein-Coupled Receptors. Biomol. Concepts 2018, 9, 143–154.

61.

Klein Herenbrink

Sykes

D. A.

Donthamsetti

, et al. The Role of Kinetic Context in Apparent Biased Agonism at GPCRs. Nat. Commun. 2016, 7, 10842.

62.

Sykes

D. A.

Moore

Stott

, et al. Extrapyramidal Side Effects of Antipsychotics Are Linked to Their Association Kinetics at Dopamine D2 Receptors. Nat. Commun. 2017, 8, 763.

63.

Cai

N.S.

Quiroz

Bonaventura

, et al. Opioid-Galanin Receptor Heteromers Mediate the Dopaminergic Effects of Opioids. J Clin Invest. 2019, 129, 2730–2744.

64.

National Institutes of Health. The Helping to End Addiction Long-Term Initiative. https://heal.nih.gov/

65.

Xiao

Huang

Chen

C. Z.

, et al. Identification and Optimization of Small-Molecule Agonists of the Human Relaxin Hormone Receptor RXFP1. Nat. Commun. 2013, 4, 1953.

66.

Kaftanovskaya

E. M.

Soula

Myhr

, et al. Human Relaxin Receptor Is Fully Functional in Humanized Mice and Is Activated by Small Molecule Agonist ML290. J. Endocr. Soc. 2017, 1, 712–725.

67.

Wagner

Dahlem

A. M.

Hudson

L. D.

, et al. A Dynamic Map for Learning, Communicating, Navigating and Improving Therapeutic Development. Nat. Rev. Drug Discov. 2018, 17, 150.

68.

Kijanska

Kelm

In Vitro 3D Spheroids and Microtissues: ATP-Based Cell Viability and Toxicity Assays. In Assay Guidance Manual; Markossian

Sittampalam

G. S.

Grossman

, et al., Eds. Eli Lilly and Company, 2004.

69.

Huang

Holtzinger

Jagan

, et al. Ductal Pancreatic Cancer Modeling and Drug Screening Using Human Pluripotent Stem Cell– and Patient-Derived Tumor Organoids. Nat. Med. 2015, 21, 1364–1371.

70.

Dye

B. R.

Hill

D. R.

Ferguson

M. A. H.

, et al. In Vitro Generation of Human Pluripotent Stem Cell Derived Lung Organoids. eLife 2015, 4, e05098.

71.

Castro-Viñuelas

Sanjurjo-Rodríguez

Piñeiro-Ramil

, et al. Generation and Characterization of Human Induced Pluripotent Stem Cells (iPSCs) from Hand Osteoarthritis Patient-Derived Fibroblasts. Sci. Rep. 2020, 10, 4272.

72.

Koga

Chaim

I. A.

Benitez

J. A.

, et al. Longitudinal Assessment of Tumor Development Using Cancer Avatars Derived from Genetically Engineered Pluripotent Stem Cells. Nat. Commun. 2020, 11, 550.

73.

Zhang

Recent Advances in High-Throughput Mass Spectrometry That Accelerates Enzyme Engineering for Biofuel Research. BMC Energy 2020, 2, 1.

74.

McLaren

D. G.

Shah

Wisniewski

, et al. High-Throughput Mass Spectrometry for Hit Identification: Current Landscape and Future Perspectives. SLAS Discov. 2021, 26, 168–191.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.24 MB