Application of artificial intelligence and machine learning techniques to the analysis of dynamic protein sequences.

We apply methods of Artificial Intelligence and Machine Learning to protein dynamic bioinformatics. We rewrite the sequences of a large protein data set, containing both folded and intrinsically disordered molecules, using a representation developed previously, which encodes the intrinsic dynamic properties of the naturally occurring amino acids. We Fourier analyze the resulting sequences. It is demonstrated that classification models built using several different supervised learning methods are able to successfully distinguish folded from intrinsically disordered proteins from sequence alone. It is further shown that the most important sequence property for this discrimination is the sequence mobility, which is the sequence averaged value of the residue-specific average alpha carbon B factor. This is in agreement with previous work, in which we have demonstrated the central role played by the sequence mobility in protein dynamic bioinformatics and biophysics. This finding opens a path to the application of dynamic bioinformatics, in combination with machine learning algorithms, to a range of significant biomedical problems.

Allosteric inhibition of SHP2 phosphatase inhibits cancers driven by receptor tyrosine kinases.

The non-receptor protein tyrosine phosphatase SHP2, encoded by PTPN11, has an important role in signal transduction downstream of growth factor receptor signalling and was the first reported oncogenic tyrosine phosphatase. Activating mutations of SHP2 have been associated with developmental pathologies such as Noonan syndrome and are found in multiple cancer types, including leukaemia, lung and breast cancer and neuroblastoma. SHP2 is ubiquitously expressed and regulates cell survival and proliferation primarily through activation of the RAS­ERK signalling pathway. It is also a key mediator of the programmed cell death 1 (PD-1) and B- and T-lymphocyte attenuator (BTLA) immune checkpoint pathways. Reduction of SHP2 activity suppresses tumour cell growth and is a potential target of cancer therapy. Here we report the discovery of a highly potent (IC50 = 0.071 µM), selective and orally bioavailable small-molecule SHP2 inhibitor, SHP099, that stabilizes SHP2 in an auto-inhibited conformation. SHP099 concurrently binds to the interface of the N-terminal SH2, C-terminal SH2, and protein tyrosine phosphatase domains, thus inhibiting SHP2 activity through an allosteric mechanism. SHP099 suppresses RAS­ERK signalling to inhibit the proliferation of receptor-tyrosine-kinase-driven human cancer cells in vitro and is efficacious in mouse tumour xenograft models. Together, these data demonstrate that pharmacological inhibition of SHP2 is a valid therapeutic approach for the treatment of cancers.

Identification of an allosteric benzothiazolopyrimidone inhibitor of the oncogenic protein tyrosine phosphatase SHP2.

The PTPN11 oncogene encodes the cytoplasmic protein tyrosine phosphatase SHP2, which, through its role in multiple signaling pathways, promotes the progression of hematological malignancies and other cancers. Here, we employ high-throughput screening to discover a lead chemical scaffold, the benzothiazolopyrimidones, that allosterically inhibits this oncogenic phosphatase by simultaneously engaging the C-SH2 and PTP domains. We improved our lead to generate an analogue that better suppresses SHP2 activity in vitro. Suppression of Erk phopsphorylation by the lead compound is also consistent with SHP2 inhibition in AML cells. Our findings provide an alternative starting point for therapeutic intervention and will catalyze investigations into the relationship between SHP2 conformational regulation, activity, and disease progression.

Structural and Functional Consequences of Three Cancer-Associated Mutations of the Oncogenic Phosphatase SHP2.

The proto-oncogene PTPN11 encodes a cytoplasmic protein tyrosine phosphatase, SHP2, which is required for normal development and sustained activation of the Ras-MAPK signaling pathway. Germline mutations in SHP2 cause developmental disorders, and somatic mutations have been identified in childhood and adult cancers and drive leukemia in mice. Despite our knowledge of the PTPN11 variations associated with pathology, the structural and functional consequences of many disease-associated mutants remain poorly understood. Here, we combine X-ray crystallography, small-angle X-ray scattering, and biochemistry to elucidate structural and mechanistic features of three cancer-associated SHP2 variants harboring single point mutations within the N-SH2:PTP interdomain autoinhibitory interface. Our findings directly compare the impact of each mutation on autoinhibition of the phosphatase and advance the development of structure-guided and mutation-specific SHP2 therapies.

Synthesis of ciprofloxacin dimers for evaluation of bacterial permeability in atypical chemical space.

We describe the synthesis and evaluation of a library of variably-linked ciprofloxacin dimers. These structures unify and expand on the use of fluoroquinolones as probes throughout the antibiotic literature. A dimeric analog (19) showed enhanced inhibition of its intracellular target (DNA gyrase), and translation to antibacterial activity in whole cells was demonstrated. Overall, cell permeation was governed by physicochemical properties and bacterial type. A principal component analysis demonstrated that the dimers occupy a unique and privileged region of chemical space most similar to the macrolide class of antibiotics.

2-Alkyloxazoles as potent and selective PI4KIIIß inhibitors demonstrating inhibition of HCV replication.

Synthesis and SAR of 2-alkyloxazoles as class III phosphatidylinositol-4-kinase beta (PI4KIIIß) inhibitors is described. These compounds demonstrate that inhibition of PI4KIIIß leads to potent inhibition of HCV replication as observed in genotype (GT) 1a and 1b replicon and GT2a JFH1 virus assays in vitro.

Predictions of Colloidal Molecular Aggregation Using AI/ML Models.

To facilitate the triage of hits from small molecule screens, we have used various AI/ML techniques and experimentally observed data sets to build models aimed at predicting colloidal aggregation of small organic molecules in aqueous solution. We have found that Naïve Bayesian and deep neural networks outperform logistic regression, recursive partitioning tree, support vector machine, and random forest techniques by having the lowest balanced error rate (BER) for the test set. Derived predictive classification models consistently and successfully discriminated aggregator molecules from nonaggregator hits. An analysis of molecular descriptors in favor of colloidal aggregation confirms previous observations (hydrophobicity, molecular weight, and solubility) in addition to undescribed molecular descriptors such as the fraction of sp3 carbon atoms (Fsp3), and electrotopological state of hydroxyl groups (ES_Sum_sOH). Naïve Bayesian modeling and scaffold tree analysis have revealed chemical features/scaffolds contributing the most to colloidal aggregation and nonaggregation, respectively. These results highlight the importance of scaffolds with high Fsp3 values in promoting nonaggregation. Matched molecular pair analysis (MMPA) has also deciphered context-dependent substitutions, which can be used to design nonaggregator molecules. We found that most matched molecular pairs have a neutral effect on aggregation propensity. We have prospectively applied our predictive models to assist in chemical library triage for optimal plate selection diversity and purchase for high throughput screening (HTS) in drug discovery projects.

Discovery of Darovasertib (NVP-LXS196), a Pan-PKC Inhibitor for the Treatment of Metastatic Uveal Melanoma.

Uveal melanoma (UM) is the most common primary intraocular malignancy in the adult eye. Despite the aggressive local management of primary UM, the development of metastases is common with no effective treatment options for metastatic disease. Genetic analysis of UM samples reveals the presence of mutually exclusive activating mutations in the Gq alpha subunits GNAQ and GNA11. One of the key downstream targets of the constitutively active Gq alpha subunits is the protein kinase C (PKC) signaling pathway. Herein, we describe the discovery of darovasertib (NVP-LXS196), a potent pan-PKC inhibitor with high whole kinome selectivity. The lead series was optimized for kinase and off target selectivity to afford a compound that is rapidly absorbed and well tolerated in preclinical species. LXS196 is being investigated in the clinic as a monotherapy and in combination with other agents for the treatment of uveal melanoma (UM), including primary UM and metastatic uveal melanoma (MUM).

Efficacy of LFF571 in a hamster model of Clostridium difficile infection.

LFF571 is a novel semisynthetic thiopeptide antibiotic with potent activity against a variety of Gram-positive pathogens, including Clostridium difficile. In vivo efficacy of LFF571 was compared to vancomycin in a hamster model of C. difficile infection (CDI). Infection was induced in Golden Syrian hamsters using a toxigenic strain of C. difficile. Treatment started 24 h postinfection and consisted of saline, vancomycin, or LFF571. Cox regression was used to analyze survival data from a cohort of animals evaluated across seven serial experimental groups treated with vancomycin at 20 mg/kg, LFF571 at 5 mg/kg, or vehicle alone. Survival was right censored; animals were not observed beyond day 21. At death or end of study, cecal contents were tested for C. difficile toxins A and B. In summary, the data showed that 5 mg/kg LFF571 decreased the risk of death by 79% (P < 0.0001) and 69% (P = 0.0022) compared with saline and 20 mg/kg vancomycin, respectively. Further analysis of the pooled data indicated that the survival benefit of LFF571 treatment at 5 mg/kg compared to vancomycin at 20 mg/kg was due primarily to a decrease in the risk of recurrence after end of treatment. Animals successfully treated with LFF571 or vancomycin had no detectable C. difficile toxin. Overall, LFF571 was more efficacious at the end of the study, at a lower dose, and with fewer recurrences, than vancomycin in the hamster model of CDI. LFF571 is being assessed in humans for safety and efficacy in the treatment of C. difficile infections.

Think Biologically, Act Chemically.

A drug is a sophisticated molecule, purposely evolved, resulting from the accumulation of knowledge learned and exploited from simpler molecules over time. Advanced molecules with increased sophistication and capability are derived from simpler, less sophisticated structures with less capabilities. Medicinal chemists do not find, stumble upon, accidentally discover, screen for, or construct drugs. We purposefully evolve molecules through the use of feedback cycles; we emphasize efficiency and simplicity in pursuit of multiproperty homeostasis; and we design and learn from molecular outliers. This Miniperspective illustrates inspirational themes from nature including evolution, feedback cycles, homeostasis, efficiency, and mutation. These biological themes are then exemplified in modern medicinal chemistry practices, such as design-make-test-analyze cycles (feedback), balancing molecular properties (homeostasis), defining the minimum pharmacophore (simplicity, efficiency), understanding molecular outliers (mutants), and the unifying concept of molecular evolution.

4-Aminothiazolyl analogs of GE2270 A: design, synthesis and evaluation of imidazole analogs.

Imidazole analogs of the antibiotic natural product GE2270 A (1) were designed, synthesized, and evaluated for gram positive bacteria growth inhibition. A recently reported, copper-mediated synthesis was exploited to prepare 4-thiazolyl imidazole analogs of 1. The synthesis described represents a structurally complex, natural product-based application of this recently reported synthetic methodology. In addition, the biological evaluation of the imidazole-based analogs further define the SAR of the 4-aminothiazolyl-based antibacterial template.

Time-resolved phosphoproteomics reveals scaffolding and catalysis-responsive patterns of SHP2-dependent signaling.

SHP2 is a protein tyrosine phosphatase that normally potentiates intracellular signaling by growth factors, antigen receptors, and some cytokines, yet is frequently mutated in human cancer. Here, we examine the role of SHP2 in the responses of breast cancer cells to EGF by monitoring phosphoproteome dynamics when SHP2 is allosterically inhibited by SHP099. The dynamics of phosphotyrosine abundance at more than 400 tyrosine residues reveal six distinct response signatures following SHP099 treatment and washout. Remarkably, in addition to newly identified substrate sites on proteins such as occludin, ARHGAP35, and PLCγ2, another class of sites shows reduced phosphotyrosine abundance upon SHP2 inhibition. Sites of decreased phospho-abundance are enriched on proteins with two nearby phosphotyrosine residues, which can be directly protected from dephosphorylation by the paired SH2 domains of SHP2 itself. These findings highlight the distinct roles of the scaffolding and catalytic activities of SHP2 in effecting a transmembrane signaling response.

Combinations with Allosteric SHP2 Inhibitor TNO155 to Block Receptor Tyrosine Kinase Signaling.

PURPOSE: SHP2 inhibitors offer an appealing and novel approach to inhibit receptor tyrosine kinase (RTK) signaling, which is the oncogenic driver in many tumors or is frequently feedback activated in response to targeted therapies including RTK inhibitors and MAPK inhibitors. We seek to evaluate the efficacy and synergistic mechanisms of combinations with a novel SHP2 inhibitor, TNO155, to inform their clinical development. EXPERIMENTAL DESIGN: The combinations of TNO155 with EGFR inhibitors (EGFRi), BRAFi, KRASG12Ci, CDK4/6i, and anti-programmed cell death-1 (PD-1) antibody were tested in appropriate cancer models in vitro and in vivo, and their effects on downstream signaling were examined. RESULTS: In EGFR-mutant lung cancer models, combination benefit of TNO155 and the EGFRi nazartinib was observed, coincident with sustained ERK inhibition. In BRAFV600E colorectal cancer models, TNO155 synergized with BRAF plus MEK inhibitors by blocking ERK feedback activation by different RTKs. In KRASG12C cancer cells, TNO155 effectively blocked the feedback activation of wild-type KRAS or other RAS isoforms induced by KRASG12Ci and greatly enhanced efficacy. In addition, TNO155 and the CDK4/6 inhibitor ribociclib showed combination benefit in a large panel of lung and colorectal cancer patient-derived xenografts, including those with KRAS mutations. Finally, TNO155 effectively inhibited RAS activation by colony-stimulating factor 1 receptor, which is critical for the maturation of immunosuppressive tumor-associated macrophages, and showed combination activity with anti-PD-1 antibody. CONCLUSIONS: Our findings suggest TNO155 is an effective agent for blocking both tumor-promoting and immune-suppressive RTK signaling in RTK- and MAPK-driven cancers and their tumor microenvironment. Our data provide the rationale for evaluating these combinations clinically.

SHP2 blockade enhances anti-tumor immunity via tumor cell intrinsic and extrinsic mechanisms.

SHP2 is a ubiquitous tyrosine phosphatase involved in regulating both tumor and immune cell signaling. In this study, we discovered a novel immune modulatory function of SHP2. Targeting this protein with allosteric SHP2 inhibitors promoted anti-tumor immunity, including enhancing T cell cytotoxic function and immune-mediated tumor regression. Knockout of SHP2 using CRISPR/Cas9 gene editing showed that targeting SHP2 in cancer cells contributes to this immune response. Inhibition of SHP2 activity augmented tumor intrinsic IFNγ signaling resulting in enhanced chemoattractant cytokine release and cytotoxic T cell recruitment, as well as increased expression of MHC Class I and PD-L1 on the cancer cell surface. Furthermore, SHP2 inhibition diminished the differentiation and inhibitory function of immune suppressive myeloid cells in the tumor microenvironment. SHP2 inhibition enhanced responses to anti-PD-1 blockade in syngeneic mouse models. Overall, our study reveals novel functions of SHP2 in tumor immunity and proposes that targeting SHP2 is a promising strategy for cancer immunotherapy.

Resistance to allosteric SHP2 inhibition in FGFR-driven cancers through rapid feedback activation of FGFR.

SHP2 mediates RAS activation downstream of multiple receptor tyrosine kinases (RTKs) and cancer cell lines dependent on RTKs are in general dependent on SHP2. Profiling of the allosteric SHP2 inhibitor SHP099 across cancer cell lines harboring various RTK dependencies reveals that FGFR-dependent cells are often insensitive to SHP099 when compared to EGFR-dependent cells. We find that FGFR-driven cells depend on SHP2 but exhibit resistance to SHP2 inhibitors in vitro and in vivo. Treatment of such models with SHP2 inhibitors results in an initial decrease in phosphorylated ERK1/2 (p-ERK) levels, however p-ERK levels rapidly rebound within two hours. This p-ERK rebound is blocked by FGFR inhibitors or high doses of SHP2 inhibitors. Mechanistically, compared with EGFR-driven cells, FGFR-driven cells tend to express high levels of RTK negative regulators such as the SPRY family proteins, which are rapidly downregulated upon ERK inhibition. Moreover, over-expression of SPRY4 in FGFR-driven cells prevents MAPK pathway reactivation and sensitizes them to SHP2 inhibitors. We also identified two novel combination approaches to enhance the efficacy of SHP2 inhibitors, either with a distinct site 2 allosteric SHP2 inhibitor or with a RAS-SOS1 interaction inhibitor. Our findings suggest the rapid FGFR feedback activation following initial pathway inhibition by SHP2 inhibitors may promote the open conformation of SHP2 and lead to resistance to SHP2 inhibitors. These findings may assist to refine patient selection and predict resistance mechanisms in the clinical development of SHP2 inhibitors and to suggest strategies for discovering SHP2 inhibitors that are more effective against upstream feedback activation.

Identification of TNO155, an Allosteric SHP2 Inhibitor for the Treatment of Cancer.

SHP2 is a nonreceptor protein tyrosine phosphatase encoded by the PTPN11 gene and is involved in cell growth and differentiation via the MAPK signaling pathway. SHP2 also plays an important role in the programed cell death pathway (PD-1/PD-L1). As an oncoprotein as well as a potential immunomodulator, controlling SHP2 activity is of high therapeutic interest. As part of our comprehensive program targeting SHP2, we identified multiple allosteric binding modes of inhibition and optimized numerous chemical scaffolds in parallel. In this drug annotation report, we detail the identification and optimization of the pyrazine class of allosteric SHP2 inhibitors. Structure and property based drug design enabled the identification of protein-ligand interactions, potent cellular inhibition, control of physicochemical, pharmaceutical and selectivity properties, and potent in vivo antitumor activity. These studies culminated in the discovery of TNO155, (3S,4S)-8-(6-amino-5-((2-amino-3-chloropyridin-4-yl)thio)pyrazin-2-yl)-3-methyl-2-oxa-8-azaspiro[4.5]decan-4-amine (1), a highly potent, selective, orally efficacious, and first-in-class SHP2 inhibitor currently in clinical trials for cancer.

Ribosomally synthesized thiopeptide antibiotics targeting elongation factor Tu.

We identified the thiomuracins, a novel family of thiopeptides produced by a rare-actinomycete bacterium typed as a Nonomuraea species, via a screen for inhibition of growth of the bacterial pathogen Staphylococcus aureus. Thiopeptides are a class of macrocyclic, highly modified peptides that are decorated by thiazoles and defined by a central six-membered heterocyclic ring system. Mining the genomes of thiopeptide-producing strains revealed the elusive biosynthetic route for this class of antibiotics. The thiopeptides are chromosomally encoded, ribosomally synthesized proteins, and isolation of gene clusters for production of thiomuracin and the related thiopeptide GE2270A revealed the post-translational machinery required for maturation. The target of the thiomuracins was identified as bacterial Elongation Factor Tu (EF-Tu). In addition to potently inhibiting a target that is unexploited by marketed human therapeutics, the thiomuracins have a low propensity for selecting for antibiotic resistance and confer no measurable cross-resistance to antibiotics in clinical use.

Tumor Intrinsic Efficacy by SHP2 and RTK Inhibitors in KRAS-Mutant Cancers.

KRAS, an oncogene mutated in nearly one third of human cancers, remains a pharmacologic challenge for direct inhibition except for recent advances in selective inhibitors targeting the G12C variant. Here, we report that selective inhibition of the protein tyrosine phosphatase, SHP2, can impair the proliferation of KRAS-mutant cancer cells in vitro and in vivo using cell line xenografts and primary human tumors. In vitro, sensitivity of KRAS-mutant cells toward the allosteric SHP2 inhibitor, SHP099, is not apparent when cells are grown on plastic in 2D monolayer, but is revealed when cells are grown as 3D multicellular spheroids. This antitumor activity is also observed in vivo in mouse models. Interrogation of the MAPK pathway in SHP099-treated KRAS-mutant cancer models demonstrated similar modulation of p-ERK and DUSP6 transcripts in 2D, 3D, and in vivo, suggesting a MAPK pathway-dependent mechanism and possible non-MAPK pathway-dependent mechanisms in tumor cells or tumor microenvironment for the in vivo efficacy. For the KRASG12C MIAPaCa-2 model, we demonstrate that the efficacy is cancer cell intrinsic as there is minimal antiangiogenic activity by SHP099, and the effects of SHP099 is recapitulated by genetic depletion of SHP2 in cancer cells. Furthermore, we demonstrate that SHP099 efficacy in KRAS-mutant models can be recapitulated with RTK inhibitors, suggesting RTK activity is responsible for the SHP2 activation. Taken together, these data reveal that many KRAS-mutant cancers depend on upstream signaling from RTK and SHP2, and provide a new therapeutic framework for treating KRAS-mutant cancers with SHP2 inhibitors.

6-Amino-3-methylpyrimidinones as Potent, Selective, and Orally Efficacious SHP2 Inhibitors.

Protein tyrosine phosphatase SHP2 is an oncoprotein associated with cancer as well as a potential immune modulator because of its role in the programmed cell death PD-L1/PD-1 pathway. In the preceding manuscript, we described the optimization of a fused, bicyclic screening hit for potency, selectivity, and physicochemical properties in order to further expand the chemical diversity of allosteric SHP2 inhibitors. In this manuscript, we describe the further expansion of our approach, morphing the fused, bicyclic system into a novel monocyclic pyrimidinone scaffold through our understanding of SAR and use of structure-based design. These studies led to the identification of SHP394 (1), an orally efficacious inhibitor of SHP2, with high lipophilic efficiency, improved potency, and enhanced pharmacokinetic properties. We also report other pyrimidinone analogues with favorable pharmacokinetic and potency profiles. Overall, this work improves upon our previously described allosteric inhibitors and exemplifies and extends the range of permissible chemical templates that inhibit SHP2 via the allosteric mechanism.

Optimization of Fused Bicyclic Allosteric SHP2 Inhibitors.

SHP2 is a nonreceptor protein tyrosine phosphatase within the mitogen-activated protein kinase (MAPK) pathway controlling cell growth, differentiation, and oncogenic transformation. SHP2 also participates in the programed cell death pathway (PD-1/PD-L1) governing immune surveillance. Small-molecule inhibition of SHP2 has been widely investigated, including in our previous reports describing SHP099 (2), which binds to a tunnel-like allosteric binding site. To broaden our approach to allosteric inhibition of SHP2, we conducted additional hit finding, evaluation, and structure-based scaffold morphing. These studies, reported here in the first of two papers, led to the identification of multiple 5,6-fused bicyclic scaffolds that bind to the same allosteric tunnel as 2. We demonstrate the structural diversity permitted by the tunnel pharmacophore and culminated in the identification of pyrazolopyrimidinones (e.g., SHP389, 1) that modulate MAPK signaling in vivo. These studies also served as the basis for further scaffold morphing and optimization, detailed in the following manuscript.