Open Source Bayesian Models. 1. Application to ADME/Tox and Drug Discovery Datasets.

On the order of hundreds of absorption, distribution, metabolism, excretion, and toxicity (ADME/Tox) models have been described in the literature in the past decade which are more often than not inaccessible to anyone but their authors. Public accessibility is also an issue with computational models for bioactivity, and the ability to share such models still remains a major challenge limiting drug discovery. We describe the creation of a reference implementation of a Bayesian model-building software module, which we have released as an open source component that is now included in the Chemistry Development Kit (CDK) project, as well as implemented in the CDD Vault and in several mobile apps. We use this implementation to build an array of Bayesian models for ADME/Tox, in vitro and in vivo bioactivity, and other physicochemical properties. We show that these models possess cross-validation receiver operator curve values comparable to those generated previously in prior publications using alternative tools. We have now described how the implementation of Bayesian models with FCFP6 descriptors generated in the CDD Vault enables the rapid production of robust machine learning models from public data or the user's own datasets. The current study sets the stage for generating models in proprietary software (such as CDD) and exporting these models in a format that could be run in open source software using CDK components. This work also demonstrates that we can enable biocomputation across distributed private or public datasets to enhance drug discovery.

Immunohistochemical FAP Expression Reflects 68Ga-FAPI PET Imaging Properties of Low- and High-Grade Intraductal Papillary Mucinous Neoplasms and Pancreatic Ductal Adenocarcinoma.

Pancreatic intraductal papillary mucinous neoplasms (IPMNs) are grossly visible (typically > 5 mm) intraductal epithelial neoplasms of mucin-producing cells, arising in the main pancreatic duct or its branches. According to the current 2-tiered grading scheme, these lesions are categorized as having either low-grade (LG) dysplasia, which has a benign prognosis, or high-grade (HG) dysplasia, which formally represents a carcinoma in situ and thus can transform to pancreatic ductal adenocarcinoma (PDAC). Because both entities require different treatments according to their risk of becoming malignant, a precise pretherapeutic diagnostic differentiation is inevitable for adequate patient management. Recently, our group has demonstrated that 68Ga-fibroblast activation protein (FAP) inhibitor (FAPI) PET/CT shows great potential for the differentiation of LG IPMNs, HG IPMNs, and PDAC according to marked differences in signal intensity and tracer dynamics. The purpose of this study was to biologically validate FAP as a target for PET imaging by analyzing immunohistochemical FAP expression in LG IPMNs, HG IPMNs, and PDAC and comparing with SUV and time to peak (TTP) measured in our prior study. Methods: To evaluate the correlation of the expression level of FAP and α-smooth muscle actin (αSMA) in neoplasm-associated stroma depending on the degree of dysplasia in IPMNs, 98 patients with a diagnosis of LG IPMN, HG IPMN, PDAC with associated HG IPMN, or PDAC who underwent pancreatic surgery at the University Hospital Heidelberg between 2017 and 2023 were identified using the database of the Institute of Pathology, University Hospital Heidelberg. In a reevaluation of hematoxylin- and eosin-stained tissue sections of formalin-fixed and paraffin-embedded resection material from the archive, which was originally generated for histopathologic routine diagnostics, a regrading of IPMNs was performed by a pathologist according to the current 2-tiered grading scheme, consequently eliminating the former diagnosis of "IPMN with intermediate-grade dysplasia." For each case, semithin tissue sections of 3 paraffin blocks containing neoplasm were immunohistologically stained with antibodies directed against FAP and αSMA. In a masked approach, a semiquantitative analysis of the immunohistochemically stained slides was finally performed by a pathologist by adapting the immunoreactive score (IRS) and human epidermal growth factor receptor 2 (Her2)/neu score to determine the intensity and percentage of FAP- and αSMA-positive cells. Afterward, the IRS of 14 patients who underwent 68Ga-FAPI-74 PET/CT in our previous study was compared with their SUVmax, SUVmean, and TTP for result validation. Results: From 98 patients, 294 specimens (3 replicates per patient) were immunohistochemically stained for FAP and αSMA. Twenty-three patients had LG IPMNs, 11 had HG IPMNs, 10 had HG IPMNs plus PDAC, and 54 had PDAC. The tumor stroma was in all cases variably positive for FAP. The staining intensity, percentage of FAP-positive stroma, IRS, and Her2/neu score increased with higher malignancy. αSMA expression could be shown in normal pancreatic stroma as well as within peri- and intraneoplastic desmoplastic reaction. No homogeneous increase in intensity, percentage, IRS, and Her2/neu score with higher malignancy was observed for αSMA. The comparison of the mean IRS of FAP with the mean SUVmax, SUVmean, and TTP of 68Ga-GAPI-74 PET/CT showed a matching value increasing with higher malignancy in 68Ga-FAPI-74 PET imaging and immunohistochemical FAP expression. Conclusion: The immunohistochemical staining of IPMNs and PDAC validates FAP as a biology-based stromal target for in vivo imaging. Increasing expression of FAP in lesions with a higher degree of malignancy matches the expectation of a stronger FAP expression in PDAC and HG IPMNs than in LG IPMNs and corroborates our previous findings of higher SUVs and a longer TTP in PDAC and HG IPMNs than in LG IPMNs.

Diagnostic Potential of Supplemental Static and Dynamic 68Ga-FAPI-46 PET for Primary 18F-FDG-Negative Pulmonary Lesions.

PET using 68Ga-labeled fibroblast activation protein (FAP) inhibitors (FAPIs) holds high potential for diagnostic imaging of various malignancies, including lung cancer (LC). However, 18F-FDG PET is still the clinical gold standard for LC imaging. Several subtypes of LC, especially lepidic LC, are frequently 18F-FDG PET-negative, which markedly hampers the assessment of single pulmonary lesions suggestive of LC. Here, we evaluated the diagnostic potential of static and dynamic 68Ga-FAPI-46 PET in the 18F-FDG-negative pulmonary lesions of 19 patients who underwent surgery or biopsy for histologic diagnosis after PET imaging. For target validation, FAP expression in lepidic LC was confirmed by FAP immunohistochemistry. Methods: Hematoxylin and eosin staining and FAP immunohistochemistry of 24 tissue sections of lepidic LC from the local tissue bank were performed and analyzed visually. Clinically, 19 patients underwent static and dynamic 68Ga-FAPI-46 PET in addition to 18F-FDG PET based on individual clinical indications. Static PET data of both examinations were analyzed by determining SUVmax, SUVmean, and tumor-to-background ratio (TBR) against the blood pool, as well as relative parameters (68Ga-FAPI-46 in relation to18F-FDG), of histologically confirmed LC and benign lesions. Time-activity curves and dynamic parameters (time to peak, slope, k 1, k 2, k 3, and k 4) were extracted from dynamic 68Ga-FAPI-46 PET data. The sensitivity and specificity of all parameters were analyzed by calculating receiver-operating-characteristic curves. Results: FAP immunohistochemistry confirmed the presence of strongly FAP-positive cancer-associated fibroblasts in lepidic LC. LC showed markedly elevated 68Ga-FAPI-46 uptake, higher TBRs, and higher 68Ga-FAPI-46-to-18F-FDG ratios for all parameters than did benign pulmonary lesions. Dynamic imaging analysis revealed differential time-activity curves for LC and benign pulmonary lesions: initially increasing time-activity curves with a decent slope were typical of LC, and steadily decreasing time-activity curve indicated benign pulmonary lesions, as was reflected by a significantly increased time to peak and significantly smaller absolute values of the slope for LC. Relative 68Ga-FAPI-46-to-18F-FDG ratios regarding SUVmax and TBR showed the highest sensitivity and specificity for the discrimination of LC from benign pulmonary lesions. Conclusion: 68Ga-FAPI-46 PET is a powerful new tool for the assessment of single 18F-FDG-negative pulmonary lesions and may optimize patient stratification in this clinical setting.

Static and Dynamic 68Ga-FAPI PET/CT for the Detection of Malignant Transformation of Intraductal Papillary Mucinous Neoplasia of the Pancreas.

Pancreatic ductal adenocarcinoma (PDAC) may arise from intraductal papillary mucinous neoplasms (IPMN) with malignant transformation, but a significant portion of IPMN remains to show benign behavior. Therefore, it is important to differentiate between benign IPMN and IPMN lesions undergoing malignant transformation. However, nonoperative differentiation by ultrasound, CT, MRI, and carbohydrate antigen 19-9 (CA19-9) is still unsatisfactory. Here, we assessed the clinical feasibility of additional assessment of malignancy by PET using 68Ga-labeled fibroblast activation protein inhibitors (68Ga-FAPI PET) in 25 patients with MRI- or CT-proven cystic pancreatic lesions. Methods: Twenty-five patients with cystic pancreatic lesions who were followed up in the European Pancreas Center of Heidelberg University hospital and who were led to surgical resection or fine-needle aspiration due to suspicious clinical, laboratory chemistry, or radiologic findings were examined by static (all patients) and dynamic (20 patients) 68Ga-FAPI PET. Cystic pancreatic lesions were delineated and SUVmax and SUVmean were determined. Time-activity curves and dynamic parameters (time to peak, K 1, k 2, K3, k 4) were extracted from dynamic PET data. Receiver-operating curves of static and dynamic PET parameters were calculated. Results: Eleven of the patients had menacing IPMN (high-grade IPMN with [6 cases] or without [5 cases] progression into PDAC) and 11 low-grade IPMN; 3 patients had other benign entities. Menacing IMPN showed significantly elevated 68Ga-FAPI uptake compared with low-grade IPMN and other benign cystic lesions. In dynamic imaging, menacing IPMN showed increasing time-activity curves followed by slow decrease afterward; time-activity curves of low-grade IPMN showed an immediate peak followed by rapid decrease for about 10 min and slower decrease for the rest of the time. Receiver-operating curves showed high sensitivity and specificity (area under the curve greater than 80%) of static and dynamic PET parameters for the differentiation of IPMN subtypes. Conclusion: 68Ga-FAPI PET is a helpful new tool for the differentiation of menacing and low-grade IPMN and shows the potential to avoid unnecessary surgery for nonmalignant pancreatic IPMN.

Impact of 68Ga-FAPI PET/CT on Staging and Oncologic Management in a Cohort of 226 Patients with Various Cancers.

Since the development of fibroblast activation protein-targeted radiopharmaceuticals, 68Ga-fibroblast activation protein inhibitor (FAPI) PET/CT has been found to be suitable for detecting primary and metastatic lesions in many types of tumors. However, there is currently a lack of reliable data regarding the clinical impact of this family of probes. To address this gap, the present study aimed to analyze the clinical impact of 68Ga-FAPI PET/CT by examining a large cohort of patients with various tumors. Methods: In total, 226 patients (137 male and 89 female) were included in this retrospective analysis. Pancreatic cancer and head and neck cancers were the most common tumor types in this cohort. TNM stage and oncologic management were initially determined with gold standard imaging, and these results were compared with 68Ga-FAPI PET/CT. Changes were classified as major and minor. Results: For 42% of all patients, TNM stage was changed by 68Ga-FAPI PET/CT results. Most of these changes resulted in upstaging. A change in clinical management occurred in 117 of 226 patients. Although a major change in management occurred in only 12% of patients, there was a significant improvement in the ability to accurately plan radiation therapy. In general, the highest clinical impact of 68Ga-FAPI PET/CT imaging was found in patients with lung cancer, pancreatic cancer, and head and neck tumors. Conclusion: 68Ga-FAPI PET/CT is a promising imaging probe that has a significant impact on TNM stage and clinical management. 68Ga-FAPI PET/CT promises to be a crucial new technology that will improve on conventional radiologic imaging methods such as contrast-enhanced CT and contrast-enhanced MRI typically acquired for cancer staging.

Comparison of early and late 68Ga-FAPI-46-PET in 33 patients with possible recurrence of pancreatic ductal adenocarcinomas.

Positron emission tomography with 68Gallium (68Ga) labeled inhibitors of fibroblast activation protein (68Ga-FAPI-PET) is a promising imaging technique for patients with recurrent pancreatic ductal adenocarcinomas (PDAC). To date, it is not clear if different acquisition timepoints for 68Ga-FAPI-PET may result in comparable imaging information and if repetitive 68Ga-FAPI-PET imaging may add diagnostic value to single timepoint acquisition for recurrent PDAC. Here we analyzed retrospectively early (20 min p.i.) and late (60 min p.i.) 68Ga-FAPI-PET imaging using FAPI-46 of 33 patients with possible recurrence of PDAC concerning detection rates and uptake over time of local recurrences, metastases, inflammatory lesions of the pancreas, cholestatic lesions of the liver and reactive tissue. 33 patients with histologically confirmed PDAC after complete or partial resection of the pancreas and possible recurrence were examined by 68Ga-FAPI-46-PET acquired 20- and 60-min post injection (p.i.) of the radiotracer. FAPI-positive lesions were classified as local recurrences, metastases, inflammatory lesions of the pancreas (ILP), cholestatic lesions of the liver and reactive tissue based on histology, PET- and CT-morphology and clinical information. Lesions were contoured, and standardized uptake values (SUVmax and SUVmean) and target-to-background ratios (TBR) were analyzed for both acquisition timepoints. In total, 152 FAPI-positive lesions (22 local relapses, 47 metastases, 26 inflammatory lesions of the pancreas, 28 reactive tissues, and 29 cholestatic lesions) were detected. Detection rates for the early and late acquisition of 68Ga-FAPI-46-PET were almost identical except cholestatic lesions, which showed a higher detection rate at early imaging. SUV parameters and TBRs of ILP significantly decreased over time. Cholestatic lesions showed a tendency towards decreasing uptake. All other types of lesions showed relatively stable uptake over time. Early and late acquisition of 68Ga-FAPI-PET results in comparable imaging information in patients with possible recurrence of PDAC. Two timepoint imaging offers additional diagnostic potential concerning differential diagnoses.

Repetitive Early 68Ga-FAPI PET Acquisition Comparing 68Ga-FAPI-02, 68Ga-FAPI-46, and 68Ga-FAPI-74: Methodologic and Diagnostic Implications for Malignant, Inflammatory/Reactive, and Degenerative Lesions.

68Ga-labeled fibroblast activation protein (FAP) inhibitor (68Ga-FAPI) PET targets 68Ga-FAPI-positive activated fibroblasts and is a promising imaging technique for various types of cancer and nonmalignant pathologies. However, discrimination between malignant and nonmalignant 68Ga-FAPI-positive lesions based on static PET with a single acquisition time point can be challenging. Additionally, the optimal imaging time point for 68Ga-FAPI PET has not been identified yet, and different 68Ga-FAPI tracer variants are currently used. In this retrospective analysis, we evaluate the diagnostic value of repetitive early 68Ga-FAPI PET with 68Ga-FAPI-02, 68Ga-FAPI-46, and 68Ga-FAPI-74 for malignant, inflammatory/reactive, and degenerative lesions and describe the implications for future 68Ga-FAPI imaging protocols. Methods: Whole-body PET scans of 24 cancer patients were acquired at 10, 22, 34, 46, and 58 min after the administration of 150-250 MBq of 68Ga-FAPI tracer molecules (8 patients each for 68Ga-FAPI-02, 68Ga-FAPI-46, and 68Ga-FAPI-74). Detection rates and SUVs (SUVmax and SUVmean) for healthy tissues, cancer manifestations, and nonmalignant lesions were measured, and target-to-background ratios (TBR) versus blood and fat were calculated for all acquisition time points. Results: For most healthy tissues except fat and spinal canal, biodistribution analysis showed decreasing uptake over time. We analyzed 134 malignant, inflammatory/reactive, and degenerative lesions. Detection rates were minimally reduced for the first 2 acquisition time points and remained at a constant high level from 34 to 58 min after injection. The uptake of all 3 variants was higher in malignant and inflammatory/reactive lesions than in degenerative lesions. 68Ga-FAPI-46 showed the highest uptake and TBRs in all pathologies. For all variants, TBRs versus blood constantly increased over time for all pathologies, and TBRs versus fat were constant or decreased slightly. Conclusion: 68Ga-FAPI PET/CT is a promising imaging modality for malignancies and benign lesions. Repetitive early PET acquisition added diagnostic value for the discrimination of malignant from nonmalignant 68Ga-FAPI-positive lesions. High detection rates and TBRs over time confirmed that PET acquisition earlier than 60 min after injection delivers high-contrast images. Additionally, considering clinical feasibility, acquisition at 30-40 min after injection might be a reasonable compromise. Different 68Ga-FAPI variants show significant differences in time-dependent biodistributional behavior and should be selected carefully depending on the clinical setting.

Data Mining and Computational Modeling of High-Throughput Screening Datasets.

We are now seeing the benefit of investments made over the last decade in high-throughput screening (HTS) that is resulting in large structure activity datasets entering public and open databases such as ChEMBL and PubChem. The growth of academic HTS screening centers and the increasing move to academia for early stage drug discovery suggests a great need for the informatics tools and methods to mine such data and learn from it. Collaborative Drug Discovery, Inc. (CDD) has developed a number of tools for storing, mining, securely and selectively sharing, as well as learning from such HTS data. We present a new web based data mining and visualization module directly within the CDD Vault platform for high-throughput drug discovery data that makes use of a novel technology stack following modern reactive design principles. We also describe CDD Models within the CDD Vault platform that enables researchers to share models, share predictions from models, and create models from distributed, heterogeneous data. Our system is built on top of the Collaborative Drug Discovery Vault Activity and Registration data repository ecosystem which allows users to manipulate and visualize thousands of molecules in real time. This can be performed in any browser on any platform. In this chapter we present examples of its use with public datasets in CDD Vault. Such approaches can complement other cheminformatics tools, whether open source or commercial, in providing approaches for data mining and modeling of HTS data.

Calibration of the continuous glucose monitoring system for transient glucose monitoring.

BACKGROUND: Oral glucose tolerance testing (OGTT) remains the gold standard for diagnosis of diabetes and is used commonly in the research laboratory. The Medtronic MiniMed (North-ridge, CA) Continuous Glucose Monitoring System (CGMS Gold), developed for long-term monitoring of glycemic levels, could provide a convenient means for tracking OGTT results during research protocols; however, the sensor demonstrates glucose and time dependencies that preclude direct application of the company-provided conversion algorithm in the first 12-24 h after sensor insertion. Here, we propose an alternative conversion algorithm that permits utilization of the CGMS monitor for glucose monitoring during this time. METHODS: Healthy female participants underwent CGMS monitoring during OGTT or fasting sessions in combination with finger-stick blood glucose measurements. Logarithmic transformations and multiple regression analysis were used to quantify the time and glucose dependence of the sensors. RESULTS: Sensor performance, as characterized by sensitivity (ratio of sensor current to capillary blood glucose levels), was shown to vary logarithmically with glucose levels as well as time after sensor insertion. A conversion algorithm developed on the basis of these observations was tested on 17 subjects during OGTT. A mean absolute relative difference between capillary blood glucose and CGMS values of 11.6 +/- 6.5% for the new algorithm was seen, compared to 20.6 +/- 5.9% with the Medtronic Solutions version 3.0c algorithm. CONCLUSIONS: Incorporation of the glucose and time dependence in CGMS sensor data yields an improved mean absolute difference between actual and estimated blood glucose values compared to the Medtronic-supplied algorithm in the immediate time period following sensor insertion.

Collaborative drug discovery for More Medicines for Tuberculosis (MM4TB).

Neglected disease drug discovery is generally poorly funded compared with major diseases and hence there is an increasing focus on collaboration and precompetitive efforts such as public-private partnerships (PPPs). The More Medicines for Tuberculosis (MM4TB) project is one such collaboration funded by the EU with the goal of discovering new drugs for tuberculosis. Collaborative Drug Discovery has provided a commercial web-based platform called CDD Vault which is a hosted collaborative solution for securely sharing diverse chemistry and biology data. Using CDD Vault alongside other commercial and free cheminformatics tools has enabled support of this and other large collaborative projects, aiding drug discovery efforts and fostering collaboration. We will describe CDD's efforts in assisting with the MM4TB project.

A collaborative database and computational models for tuberculosis drug discovery.

The search for molecules with activity against Mycobacterium tuberculosis (Mtb) is employing many approaches in parallel including high throughput screening and computational methods. We have developed a database (CDD TB) to capture public and private Mtb data while enabling data mining and collaborations with other researchers. We have used the public data along with several cheminformatics approaches to produce models that describe active and inactive compounds. We have compared these datasets to those for known FDA approved drugs and between Mtb active and inactive compounds. The distribution of polar surface area and pK(a) of active compounds was found to be a statistically significant determinant of activity against Mtb. Hydrophobicity was not always statistically significant. Bayesian classification models for 220, 463 molecules were generated and tested with external molecules, and enabled the discrimination of active or inactive substructures from other datasets in the CDD TB. Computational pharmacophores based on known Mtb drugs were able to map to and retrieve a small subset of some of the Mtb datasets, including a high percentage of Mtb actives. The combination of the database, dataset analysis, Bayesian and pharmacophore models provides new insights into molecular properties and features that are determinants of activity in whole cells. This study provides novel insights into the key 1D molecular descriptors, 2D chemical substructures and 3D pharmacophores which can be used to mine the chemistry space, prioritizing those molecules with a higher probability of activity against Mtb.

Analysis and hit filtering of a very large library of compounds screened against Mycobacterium tuberculosis.

There is an urgent need for new drugs against tuberculosis which annually claims 1.7-1.8 million lives. One approach to identify potential leads is to screen in vitro small molecules against Mycobacterium tuberculosis (Mtb). Until recently there was no central repository to collect information on compounds screened. Consequently, it has been difficult to analyze molecular properties of compounds that inhibit the growth of Mtb in vitro. We have collected data from publically available sources on over 300 000 small molecules deposited in the Collaborative Drug Discovery TB Database. A cheminformatics analysis on these compounds indicates that inhibitors of the growth of Mtb have statistically higher mean logP, rule of 5 alerts, while also having lower HBD count, atom count and lower PSA (ChemAxon descriptors), compared to compounds that are classed as inactive. Additionally, Bayesian models for selecting Mtb active compounds were evaluated with over 100 000 compounds and, they demonstrated 10 fold enrichment over random for the top ranked 600 compounds. This represents a promising approach for finding compounds active against Mtb in whole cells screened under the same in vitro conditions. Various sets of Mtb hit molecules were also examined by various filtering rules used widely in the pharmaceutical industry to identify compounds with potentially reactive moieties. We found differences between the number of compounds flagged by these rules in Mtb datasets, malaria hits, FDA approved drugs and antibiotics. Combining these approaches may enable selection of compounds with increased probability of inhibition of whole cell Mtb activity.

Mts1 regulates the assembly of nonmuscle myosin-IIA.

The formation of myosin-II filaments is fundamental to contractile and motile processes in nonmuscle cells, and elucidating the mechanisms controlling filament assembly is essential for understanding how myosin-II rapidly responds to changing conditions within the cell. Several proteins including KRP and a novel 38 kDa protein (1, 2) have been shown to modulate filament assembly through the stabilization of myosin-II assemblies. In contrast, we demonstrate that mts1, a member of the Ca(2+)-regulated S100 family of proteins, may regulate the monomeric, unassembled state in an isoform-specific manner. Biochemical analyses demonstrate that mts1 has a 9-fold higher affinity for myosin-IIA filaments than for myosin-IIB filaments. At stoichiometric levels, mts1 inhibits the assembly of myosin-IIA monomers into filaments and promotes the disassembly of myosin-IIA filaments into monomers; however, mts1 has little effect on the assembly properties of myosin-IIB. Using a solution based-assay, we have demonstrated that mts1 binds to residues 1909-1924 of the myosin-IIA heavy chain, which is near the C-terminal tip of the alpha-helical coiled-coil. The observation that mts1 binds a linear sequence of approximately 16 amino acids is consistent with other S100 family members, which bind linear sequences of 13-22 residues in their protein targets. In addition, mts1 increases the critical monomer concentration for myosin-IIA filament assembly by approximately 11-fold. Kinetic assembly assays indicate that the elongation rate and the extent of polymerization depend on the initial myosin-IIA concentration; however, mts1 had only a small affect on the half-time for assembly and predominately affected the extent of myosin IIA polymerization. Altogether, these observations are consistent with mts1 regulating myosin IIA assembly by monomer sequestration and suggest that mts1 regulates cell shape and motility through the modulation of myosin-IIA function.