Application of Artificial Intelligence/Machine Vision & Learning for the Development of a Live Single-cell Phenotypic Biomarker Test to Predict Prostate Cancer Tumor Aggressiveness

To assess the usefulness and applications of machine vision (MV) and machine learning (ML) techniques that have been used to develop a single cell-based phenotypic (live and fixed biomarkers) platform that correlates with tumor biological aggressiveness and risk stratification, 100 fresh prostate samples were acquired, and areas of prostate cancer were determined by post-surgery pathology reports logged by an independent pathologist. The prostate samples were dissociated into single-cell suspensions in the presence of an extracellular matrix formulation. These samples were analyzed via live-cell microscopy. Dynamic and fixed phenotypic biomarkers per cell were quantified using objective MV software and ML algorithms. The predictive nature of the ML algorithms was developed in two stages. First, random forest (RF) algorithms were developed using 70% of the samples. The developed algorithms were then tested for their predictive performance using the blinded test dataset that contained 30% of the samples in the second stage. Based on the ROC (receiver operating characteristic) curve analysis, thresholds were set to maximize both sensitivity and specificity. We determined the sensitivity and specificity of the assay by comparing the algorithm-generated predictions with adverse pathologic features in the radical prostatectomy (RP) specimens. Using MV and ML algorithms, the biomarkers predictive of adverse pathology at RP were ranked and a prostate cancer patient risk stratification test was developed that distinguishes patients based on surgical adverse pathology features. The ability to identify and track large numbers of individual cells over the length of the microscopy experimental monitoring cycles, in an automated way, created a large biomarker dataset of primary biomarkers. This biomarker dataset was then interrogated with ML algorithms used to correlate with post-surgical adverse pathology findings. Algorithms were generated that predicted adverse pathology with >0.85 sensitivity and specificity and an AUC (area under the curve) of >0.85. Phenotypic biomarkers provide cellular and molecular details that are informative for predicting post-surgical adverse pathologies when considering tumor biopsy samples. Artificial intelligence ML-based approaches for cancer risk stratification are emerging as important and powerful tools to compliment current measures of risk stratification. These techniques have capabilities to address tumor heterogeneity and the molecular complexity of prostate cancer. Specifically, the phenotypic test is a novel example of leveraging biomarkers and advances in MV and ML for developing a powerful prognostic and risk-stratification tool for prostate cancer patients.     

A new approach in the design of electronic portal imaging devices for portal dosimetry in radiotherapy

A CCD-based EPID using new crystal-assembly X-ray (CAX) converters is investigated for radiotherapy dosimetry. The proposed EPID design consists in replacing the common phosphor X-ray converters of current CCD-based EPIDs with high-stopping-power CAX converters. A Test Imaging Device (TID), consisting of a 30-mm-thick CAX converter made of Bismuth Germanate (BGO), coupled to a highly sensitive CCD camera, was used to evaluate the accessible imaging and dosimetric performance of the proposed design. The system response to dose and its dependence on photon beam energy were investigated. The effects of ghosting, dose rate, field size and phantom thickness were evaluated as well. The same measurements were also performed with our clinically used aSi-EPID so that comparisons of performance could be directly inferred. The TID displayed no detectable ghosting or sensitivity to dose rate. Its response to MU exposure was found to be linear within about ±1%. The level of glare induced in the TID and the aSi-EPID were equivalent. The TID resolution was higher than that of the aSi-EPID on the axis, but was found to decrease with off-axis distance. Finally, the image quality, assessed on the basis of signal-to-noise ratio in low dose radiographs of the larynx of a patient, was higher for the TID. The imaging performance accessible with the TID proved to be satisfying and its dosimetric capability was found to be superior to that of the current aSi-EPID.     

An osteoclast-targeting agent for imaging and therapy of bone metastasis

A hybrid compound (DO3A-BP) featuring a radiometal bifunctional chelator (1,4,7,10-tetraazacyclotetradecane-N,N′,N″,N-tetraacetic acid, DOTA) and an osteoclast-targeting moiety (bisphosphonate) was designed and synthesized. The ¹¹¹In-labeled complex of DO3A-BP showed significantly elevated uptake in osteoclasts compared to the undifferentiated adherent bone marrow derived cells. Biodistribution studies revealed a favorable tissue distribution profile in normal mice with high bone uptake and long retention, and low or negligible accumulation in non-target organs.     

Design and synthesis of MMP inhibitors with appended fluorescent tags for imaging and visualization of matrix metalloproteinase enzymes

We describe the synthesis, MMP-2 and 9 potency, and in vitro evaluation of a series of α-sulfone hydroxmate MMP inhibitors conjugated to a series of dyes with different absorption/emission lamina maxima’s that can be used to visualize tumors.     

A mathematical feasibility argument for the use of aptamers in chemotherapy and imaging

A challenge for drug design is to create molecules with optimal functions that also partition efficiently into the appropriate in vivo compartment(s). This is particularly true in cancer treatments because cancer cells upregulate their expression of multidrug resistance transporters, which necessitates a higher concentration of extracellular drug to promote sufficiently high intracellular concentrations for cell killing. Pharmacokinetics can be improved by ancillary molecules, such as cyclodextrins, that increase the effective concentrations of hydrophobic drugs in the blood by providing hydrophobic binding pockets. However, the extent to which the extracellular concentration of drug can be increased is limited. A second approach, different from the ‘push’ mechanism just discussed, is a ‘pull’ mechanism by which the effective intracellular concentrations of a drug is increased by a molecule with an affinity for the drug that is located inside the cell. Here we propose and give a proof in principle that intracellular RNA aptamers might perform this function.The mathematical model considers the following: Suppose I denotes a drug (inhibitor) that must be distributed spatially throughout a cell, but that tends to remain outside the cell due the transport properties of the cell membrane. Suppose that E, an enzyme that binds to I, is expressed by the cell and remains in the cell. It may be that the equilibrium is not sufficiently far enough to the right to drive enough free inhibitor into the cell to completely inhibit the enzyme.Here we evaluate the use of an intracellular aptamer with affinity for the inhibitor (I) to increase the efficiency of inhibitor transport across the cell membrane and thus drive the above equilibrium further to the right than would ordinarily be the case. We show that this outcome will occur if: (1) the aptamer neither binds too tightly nor too weakly to the inhibitor than the enzyme and (2) the aptamer is much more diffusible in the cell cytoplasm than the enzyme. Thus, we propose and show by simulation that an intracellular aptamer can be enlisted for an integrated approach to increasing inhibitor effectiveness and imaging aptamer-expressing cells.     

Syntheses and evaluation of a homologous series of aza-vesamicol as improved radioiodine-labeled probes for sigma-1 receptor imaging

Sigma-1 receptor imaging probes for determining the expression levels are desirable for diagnoses of various diseases and companion diagnoses of therapeutic agents targeting the sigma-1 receptor. In this study, we aimed to develop probes with higher affinity for the sigma-1 receptor. For this purpose, we synthesized and evaluated compounds, namely, vesamicol derivatives, in which alkyl chains of varying chain length were introduced between a piperazine ring and a benzene ring. The binding affinity of the vesamicol derivatives for the sigma-1 receptor tended to increase depending on the length of the alkyl chain between the benzene ring and the piperazine ring. The sigma-1 receptor of 2-(4-(3-phenylpropyl)piperazin-1-yl)cyclohexan-1-ol (5) (K_i?=?5.8?nM) exhibited the highest binding affinity; therefore, we introduced radioiodine into the benzene ring in 5. The radioiodine labeled probe [¹²⁵I]2-(4-(3-(4-iodophenyl)propyl)piperazin-1-yl)cyclohexan-1-ol ([¹²⁵I]10) showed high accumulation in the sigma-1 receptor expressing DU-145 cells both in vitro and in vivo. Co-injection of [¹²⁵I]10 with an excess level of a sigma receptor ligand, haloperidol, resulted in a significant decrease in the tumor accumulation in vitro and in vivo, indicating sigma receptor-mediated tumor uptake. These results provide useful information for developing sigma-1 receptor imaging probes.     

Comparison of the biological properties of two anti-mucin-1 antibodies prepared for imaging and therapy

 A comparison was made between the murine anti-MUC1 antibody BC2 (which reacts with the peptide epitope APDTR) and the “humanised” antibody hCTMO1 from CellTech, which reacts with the MUC1 epitope RPAP. Preliminary studies demonstrated that hCTMO1 was a “good” antibody whereas BC2 was not. Various parameters were determined and conclusions reached. (a) Affinity: the affinity of hCTMO1 was 2.60×10⁷M^–1 and that of BC2 was 1.36×10⁷M^–1; we did not consider these numbers to be substantially different, although hCTMO1 was clearly of higher affinity than BC2. (b) On/off rate at 4°C: both antibodies bound effectively to the MUC-1 transfectant MOR5-CF2; the association rate for hCTMO1 was 3.8 times that of BC2 and the dissociation rate for BC2 was twice as fast as that of hCTMO1. (c) On/off rates at 37°C: at 37°C the association rate for hCTMO1 was greater than that of BC2. (d) Internalization: hCTMO1 was also more efficient at internalising bound antibody; 70% of bound hCTMO1 was internalised, whilst 6% of bound BC2 was internalised. From these studies it was clear that, while hCTMO1 was of similar affinity to BC2, the faster uptake and internalisation and lower off rate indicated that it was likely to be a superior antibody; this was proven in vivo. (e) Localisation: hCTMO1 bound much better in vivo than BC2 (68% compared to 28%). (f) Therapeutic experiments: BC2-idarubicin conjugates were essentially ineffective in eradicating tumours in mice whereas hCTMO1-idarubicin had a dramatic effect on breast cancer tumour cells growing in mice. We conclude that the simple measurements on/off rates and internalisation at 37°C are the most important parameters to use to determine antibody effectiveness, prior to embarking on clinical studies. Received: 29 January 1997 / Accepted: 13 May 1997     

Evaluation of 99mTc-probestin for imaging APN expressing tumors by SPECT

Aminopeptidase N (APN) is known to play important roles in tumor angiogenesis, tumor cell invasion, and metastasis. Thus, APN is an attractive biomarker for imaging tumor angiogenesis. Here we report results obtained from biodistribution and single photon emission computed tomography (SPECT) imaging studies of a technetium-99m labeled probestin (a potent APN inhibitor) conjugate containing a tripeptide, Asp-DAP-Cys (DAP = 2,3-diaminopropionic acid), chelator and a 8-amino-3,6-dioxaoctanoic acid (PEG₂) linker conducted in nude mice xenografted with HT-1080 human fibrosarcoma tumors (APN-positive tumors). These results collectively demonstrate that ^99mTc-probestin uptake by tumors and other APN expressing tissues in vivo is specific and validate the use of probestin as a vector for targeting APN in vivo.     

一种根癌农杆菌介导的简单、快捷转化雪柑的新方法

以雪柑种子为外植体,以pC1301-PMI-LFY重组载体为转化载体,将蘸有农杆菌侵染液的接种针在种子胚的顶端分生组织部位刺2-3次,深度1 mm左右,当实生苗成活后移栽.通过PCR检测和PCR-Southern杂交检测得到了1株转基因植株,转化率为4.35％.针刺法简便快速,在雪柑上能够成功应用,为木本植物的遗传转化提供了新的思路.     

Clinical relevance of novel imaging technologies for sentinel lymph node identification and staging

The sentinel lymph node (SLN) concept has become a standard of care for patients with breast cancer and melanoma, yet its clinical application to other cancer types has been somewhat limited. This is mainly due to the reduced accuracy of conventional SLN mapping techniques (using blue dye and/or radiocolloids as lymphatic tracers) in cancer types where lymphatic drainage is more complex, and SLNs are within close proximity to other nodes or the tumour site. In recent years, many novel techniques for SLN mapping have been developed including fluorescence, x-ray, and magnetic resonant detection. Whilst each technique has its own advantages/disadvantages, the role of targeted contrast agents (for enhanced retention in the SLN, or for immunostaging) is increasing, and may represent the new standard for mapping the SLN in many solid organ tumours. This review article discusses current limitations of conventional techniques, limiting factors of nanoparticulate based contrast agents, and efforts to circumvent these limitations with modern tracer architecture.     

Paramagnetic liposomes containing amphiphilic bisamide derivatives of Gd-DTPA with aromatic side chain groups as possible contrast agents for magnetic resonance imaging

Three amphiphilic DTPA bisamide derivatives containing long-chain phenylalanine esters (with 14, 16 and 18 carbon atoms in the alkyl chain) were synthesized and their corresponding gadolinium(III) complexes were prepared. The attempts to form paramagnetic micelles carrying the gadolinium(III) complexes yielded unstable or polydisperse micelles implying that the presence of the bulky aromatic side groups in the amphiphilic Gd-DTPA bisamide complexes results in an inefficient packing of the paramagnetic complex into micelles. All complexes were efficiently incorporated into liposomes consisting of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), yielding stable and monodisperse paramagnetic liposomes. All liposomes had a comparable size, typically between 120 and 160 nm. As a result of the reduced mobility of the gadolinium(III) complexes, solutions of these supramolecular structures show a higher relaxivity than solutions of Gd-DTPA. However, the relaxivity gain is lower compared to compounds consisting of purely aliphatic chains of the same length, most likely due to the less efficient packing or increased local mobility of the gadolinium(III) complex. In the case of the Gd-DTPA bisamide complex with 18 carbon atoms, the immobilization inside the liposomal structure is less effective, probably because the aliphatic chains of the complex are longer than the alkyl chains of the DPPC host, resulting in a relatively high local mobility. The paramagnetic liposomes containing the Gd-DTPA bisamide complexes with 14 carbon atoms showed the highest relaxivity because the optimal length match between the hydrophobic chains of the DPPC and the ligand allowed very efficient packing of the paramagnetic complex into the liposome.     

Engineered annexin A5 variants have impaired cell entry for molecular imaging of apoptosis using pretargeting strategies

Phosphatidylserine (PS) on apoptotic cells is a target for diagnosis and therapy using annexin A5 (anxA5). Pretargeting is a strategy developed to improve signal to background ratio for molecular imaging and to minimize undesired side effects of pharmacological and radiotherapy. Pretargeting relies on accessibility of the target finder on the surface of the target cell. anxA5 binds PS and crystallizes in a two-dimensional network covering the PS-expressing cell surface. Two-dimensional crystallization is the driving force for anxA5 internalization by PS-expressing cells. Here, we report structure/function analysis of anxA5 internalization. Guided by structural bioinformatics including protein-protein docking, we revealed that the amino acids Arg(63), Lys(70), Lys(101), Glu(138), Asp(139), and Asn(160) engage in intermolecular salt bridges within the anxA5 trimer, which is the basic building block of the two-dimensional network. Disruption of the salt bridges by site-directed mutagenesis does not affect PS binding but inhibits trimer formation and cell entry of surface-bound anxA5. The anxA5 variants with impaired internalization are superior molecular imaging agents in pretargeting strategies as compared with wild-type anxA5.     

Facile radiosynthesis of new carbon-11-labeled propanamide derivatives as selective androgen receptor modulator (SARM) radioligands for prostate cancer imaging

The androgen receptor (AR) is an attractive target for the treatment and molecular imaging of prostate cancer. New carbon-11-labeled propanamide derivatives were first designed and synthesized as selective androgen receptor modulator (SARM) radioligands for prostate cancer imaging using the biomedical imaging technique positron emission tomography (PET). The target tracers, (S)-N-(4-cyano-3-(trifluoromethyl)phenyl)-2-hydroxy-3-(2-[¹¹C]methoxyphenoxy)-2-methylpropanamide ([¹¹C]8a), (S)-2-hydroxy-3-(2-[¹¹C]methoxyphenoxy)-2-methyl-N-(4-nitro-3-(trifluoromethyl)phenyl)propanamide ([¹¹C]8e), (S)-N-(4-cyano-3-(trifluoromethyl)phenyl)-2-hydroxy-3-(4-[¹¹C]methoxyphenoxy)-2-methylpropanamide ([¹¹C]8c) and (S)-2-hydroxy-3-(4-[¹¹C]methoxyphenoxy)-2-methyl-N-(4-nitro-3-(trifluoromethyl)phenyl)propanamide ([¹¹C]8g), were prepared by O-[¹¹C]methylation of their corresponding precursors, (S)-N-(4-cyano-3-(trifluoromethyl)phenyl)-2-hydroxy-3-(2-hydroxyphenoxy)-2-methylpropanamide (9a), (S)-2-hydroxy-3-(2-hydroxyphenoxy)-2-methyl-N-(4-nitro-3-(trifluoromethyl)phenyl)propanamide (9b), (S)-N-(4-cyano-3-(trifluoromethyl)phenyl)-2-hydroxy-3-(4-hydroxyphenoxy)-2-methylpropanamide (9c) and (S)-2-hydroxy-3-(4-hydroxyphenoxy)-2-methyl-N-(4-nitro-3-(trifluoromethyl)phenyl)propanamide (9d), with [¹¹C]CH₃OTf under basic conditions and isolated by a simplified C-18 solid-phase extraction (SPE) method in 55 ± 5% (n = 5) radiochemical yields based on [¹¹C]CO₂ and decay corrected to end of bombardment (EOB). The overall synthesis time from EOB was 23 min, the radiochemical purity was >99%, and the specific activity at end of synthesis (EOS) was 277.5 ± 92.5 GBq/μmol (n = 5).     

In vivo Macrophage Imaging Using MR Targeted Contrast Agent for Longitudinal Evaluation of Septic Arthritis

Macrophages are key-cells in the initiation, the development and the regulation of the inflammatory response to bacterial infection. Macrophages are intensively and increasingly recruited in septic joints from the early phases of infection and the infiltration is supposed to regress once efficient removal of the pathogens is obtained. The ability to identify in vivo macrophage activity in an infected joint can therefore provide two main applications: early detection of acute synovitis and monitoring of therapy.In vivo noninvasive detection of macrophages can be performed with magnetic resonance imaging using iron nanoparticles such as ultrasmall superparamagnetic iron oxide (USPIO). After intravascular or intraarticular administration, USPIO are specifically phagocytized by activated macrophages, and, due to their magnetic properties, induce signal changes in tissues presenting macrophage infiltration. A quantitative evaluation of the infiltrate is feasible, as the area with signal loss (number of dark pixels) observed on gradient echo MR images after particles injection is correlated with the amount of iron within the tissue and therefore reflects the number of USPIO-loaded cells.We present here a protocol to perform macrophage imaging using USPIO-enhanced MR imaging in an animal model of septic arthritis, allowing an initial and longitudinal in vivo noninvasive evaluation of macrophages infiltration and an assessment of therapy action.     

Improved decision making for prioritizing tumor targeting antibodies in human xenografts: Utility of fluorescence imaging to verify tumor target expression,antibody binding and optimization of dosage and application schedule

Preclinical efficacy studies of antibodies targeting a tumor-associated antigen are only justified when the expression of the relevant antigen has been demonstrated. Conventionally, antigen expression level is examined by immunohistochemistry of formalin-fixed paraffin-embedded tumor tissue section. This method represents the diagnostic “gold standard” for tumor target evaluation, but is affected by a number of factors, such as epitope masking and insufficient antigen retrieval. As a consequence, variances and discrepancies in histological staining results can occur, which may influence decision-making and therapeutic outcome. To overcome these problems, we have used different fluorescence-labeled therapeutic antibodies targeting human epidermal growth factor receptor (HER) family members and insulin-like growth factor-1 receptor (IGF1R) in combination with fluorescence imaging modalities to determine tumor antigen expression, drug-target interaction, and biodistribution and tumor saturation kinetics in non-small cell lung cancer xenografts. For this, whole-body fluorescence intensities of labeled antibodies, applied as a single compound or antibody mixture, were measured in Calu-1 and Calu-3 tumor-bearing mice, then ex vivo multispectral tumor tissue analysis at microscopic resolution was performed. With the aid of this simple and fast imaging method, we were able to analyze the tumor cell receptor status of HER1–3 and IGF1R, monitor the antibody-target interaction and evaluate the receptor binding sites of anti-HER2-targeting antibodies. Based on this, the most suitable tumor model, best therapeutic antibody, and optimal treatment dosage and application schedule was selected. Predictions drawn from obtained imaging data were in excellent concordance with outcome of conducted preclinical efficacy studies. Our results clearly demonstrate the great potential of combined in vivo and ex vivo fluorescence imaging for the preclinical development and characterization of monoclonal antibodies.     

Construction of a Preclinical Multimodality Phantom Using Tissue-mimicking Materials for Quality Assurance in Tumor Size Measurement

World Health Organization (WHO) and the Response Evaluation Criteria in Solid Tumors (RECIST) working groups advocated standardized criteria for radiologic assessment of solid tumors in response to anti-tumor drug therapy in the 1980s and 1990s, respectively. WHO criteria measure solid tumors in two-dimensions, whereas RECIST measurements use only one-dimension which is considered to be more reproducible ^{1, 2, 3,4,5}. These criteria have been widely used as the only imaging biomarker approved by the United States Food and Drug Administration (FDA) ⁶. In order to measure tumor response to anti-tumor drugs on images with accuracy, therefore, a robust quality assurance (QA) procedures and corresponding QA phantom are needed.To address this need, the authors constructed a preclinical multimodality (for ultrasound (US), computed tomography (CT) and magnetic resonance imaging (MRI)) phantom using tissue-mimicking (TM) materials based on the limited number of target lesions required by RECIST by revising a Gammex US commercial phantom ⁷. The Appendix in Lee et al. demonstrates the procedures of phantom fabrication ⁷. In this article, all protocols are introduced in a step-by-step fashion beginning with procedures for preparing the silicone molds for casting tumor-simulating test objects in the phantom, followed by preparation of TM materials for multimodality imaging, and finally construction of the preclinical multimodality QA phantom. The primary purpose of this paper is to provide the protocols to allow anyone interested in independently constructing a phantom for their own projects. QA procedures for tumor size measurement, and RECIST, WHO and volume measurement results of test objects made at multiple institutions using this QA phantom are shown in detail in Lee et al. ⁸.     

CRISPys: Optimal sgRNA Design for Editing Multiple Members of a Gene Family Using the CRISPR System

The development of the CRISPR–Cas9 system in recent years has made eukaryotic genome editing, and specifically gene knockout for reverse genetics, a simple and effective task. The system is directed to a genomic target site by a programmed single-guide RNA (sgRNA) that base-pairs with it, subsequently leading to site-specific modifications. However, many gene families in eukaryotic genomes exhibit partially overlapping functions, and thus, the knockout of one gene might be concealed by the function of the other. In such cases, the reduced specificity of the CRISPR–Cas9 system, which may lead to the modification of genomic sites that are not identical to the sgRNA, can be harnessed for the simultaneous knockout of multiple homologous genes. We introduce CRISPys, an algorithm for the optimal design of sgRNAs that would potentially target multiple members of a given gene family. CRISPys first clusters all the potential targets in the input sequences into a hierarchical tree structure that specifies the similarity among them. Then, sgRNAs are proposed in the internal nodes of the tree by embedding mismatches where needed, such that the efficiency to edit the induced targets is maximized. We suggest several approaches for designing the optimal individual sgRNA and an approach to compute the optimal set of sgRNAs for cases when the experimental platform allows for more than one. The latter may optionally account for the homologous relationships among gene-family members. We further show that CRISPys outperforms simpler alignment-based techniques by in silico examination over all gene families in the Solanum lycopersicum genome.