Refine
Has Fulltext
- yes (13)
Is part of the Bibliography
- yes (13)
Document Type
- Journal article (12)
- Doctoral Thesis (1)
Keywords
- artificial intelligence (5)
- deep learning (3)
- machine learning (3)
- prediction (3)
- cancer (2)
- classification (2)
- ANP (1)
- ATM (1)
- Angiotensin II (1)
- Atriales natriuretisches Hormon (1)
Institute
- Klinik und Poliklinik für Mund-, Kiefer- und Plastische Gesichtschirurgie (6)
- Pathologisches Institut (4)
- Medizinische Klinik und Poliklinik II (2)
- Theodor-Boveri-Institut für Biowissenschaften (2)
- Comprehensive Cancer Center Mainfranken (1)
- Institut für Virologie und Immunbiologie (1)
- Klinik und Poliklinik für Dermatologie, Venerologie und Allergologie (1)
- Klinik und Poliklinik für Strahlentherapie (1)
- Physiologisches Institut (1)
Das atriale natriuretische Peptid (ANP) wird infolge einer Zunahme des atrialen Drucks aus den Myozyten des Atriums sezerniert. Es spielt lokal eine bedeutende, protektive Rolle und wirkt der Entstehung von Herzhypertrophie und Fibrose entgegen. Darüber hinaus kommt ANP vor allem eine wichtige Rolle als endokrines Hormon zu, das den arteriellen Blutdruck und das Blutvolumen regelt. Diese physiologischen Effekte vermittelt das Herzhormon durch seinen Rezeptor, das Transmembranprotein Guanylatzyklase A (GC-A). Durch Bindung von ANP an die extrazelluläre Domäne der GC-A wird intrazellulär, durch die katalytische Domäne des Rezeptors, der sekundäre Botenstoff cGMP gebildet. Patienten mit einer, durch Bluthochdruck verursachten Herzhypertrophie und Herzinsuffizienz weisen erhöhte ANP-Konzentrationen im Plasma auf. Die durch ANP vermittelten, protektiven Effekte sind allerdings vermindert. Zahlreiche Studien haben in vitro gezeigt, dass die chronische Inkubation der GC-A mit ihrem Liganden, sowie die Behandlung von GC-A exprimierenden Zellen mit Hormonen wie Angiotensin II, zur Desensitisierung des Rezeptors führen. Der Verlust der Funktionsfähigkeit geht einher mit der Dephosphorylierung des Rezeptors an spezifischen, intrazellulär lokalisierten Aminosäuren. Durch die Erforschung dieses Mechanismus und Identifizierung möglicher Interaktionspartner in vivo könnte der Grundstein für neue oder verbesserte Therapieformen gelegt werden.
Im ersten Teil der vorliegenden Arbeit wurde eine kürzlich identifizierte Isoform des GC-A-Rezeptors identifiziert, die durch alternatives Spleißen des Exons 4 entsteht und in einer Vielzahl untersuchter Gewebe der Maus vorkommt. Die Deletion umfasst 51 Basenpaare und resultiert in einem um 17 Aminosäuren verkürzten GC-A-Rezeptor (GC-AΔLys314-Gln330). Molekulare Modellierungen der extrazellulären Domänen des wildtypischen GC-A-Rezeptors und der Isoform zeigten, dass sich die Deletion im membrannahen Bereich der extrazellulären Domäne und damit deutlich entfernt von der ANP-Bindungsdomäne befindet. Oberflächenbiotinylierungs- und Zellfraktionierungsversuche zeigten, dass die Isoform des GC-A-Rezeptors an der Oberfläche von Zellmembranen transient transfizierter HEK 293-Zellen präsentiert wird. Jedoch zeigten die ANP-Stimulationsexperimente unter Anwendung von cGMP-Radioimmunassay (cGMP-RIA) und Förster-Resonanzenergietransfer (FRET)-Messungen, dass die Isoform nicht zur ANP-vermittelten intrazellulären cGMP-Bildung stimuliert werden kann. Im Rahmen von ANP-Bindungsstudien mit 125I-ANP wurde gezeigt, dass GC-AΔLys314-Gln330 die Fähigkeit zur Bindung des Liganden ANP verloren hat. Jedoch zeigten die Koimmunpräzipitationsversuche, dass die Isoform des GC-A-Rezeptors Heterodimere mit dem wildtypischen GC-A-Rezeptor bilden und dadurch die ligandeninduzierte Bildung von cGMP reduzieren kann. In vivo konnte gezeigt werden, dass unter Angiotensin II-induzierter Hypertonie die mRNA-Expression für GC-AΔLys314-Gln330 in der Lunge gesteigert, und gleichzeitig die ANP-vermittelte cGMP-Bildung deutlich reduziert ist. Daher kann davon ausgegangen werden, dass das alternative Spleißen ein regulierender Mechanismus ist, der auf den ANP/GC-A-Signalweg Einfluss nimmt. Angiotensin II-induziertes alternatives Spleißen des GC-A-Gens kann daher einen neuen Mechanismus für die Verringerung der Sensitivität des GC-A-Rezeptors gegenüber ANP darstellen.
Im zweiten Teil der vorliegenden Arbeit wurden transgene Tiere mit kardiomyozytenspezifischer Überexpression eines Epitop-getaggten GC-A-Rezeptors generiert. Durch dieses Modell sollte es ermöglicht werden, den Rezeptor aus murinem Gewebe anreichern und aufreinigen zu können um danach Analysen zu posttranslationalen Veränderungen und möglichen Interaktionspartnern durchzuführen. Zunächst wurde in eine FLAG-Epitop-getaggte GC-A zusätzlich ein HA-tag, sowie eine Erkennungssequenz für die Protease des tobacco etch virus (TEV) eingefügt. Die Expression und Funktionsfähigkeit des modifizierten Rezeptors wurde durch ANP-Stimulationsexperimente unter Anwendung von cGMP-RIA und FRET-Messungen verifiziert. Die Funktionsfähigkeit der TEV-Erkennungssequenz wurde durch die Elution mittels TEV-Protease nach Immunpräzipitation (IP) nachgewiesen. In vivo wurde an Mäusen die Expression und Lokalisation der GC-A auf Proteinebene, unter Anwendung von Zellfraktionierungsexperimenten und Immunpräzipitationen, überprüft. Die entstandenen transgenen Tiere zeigten eine deutliche, in den Zellmembranen von Kardiomyozyten lokalisierte, Überexpression des Rezeptors. Dieser konnte über das HA-tag angereichert und aufgereinigt werden. Um die Funktionsfähigkeit des modifizierten Rezeptors in vivo nachzuweisen, wurde in zwei Versuchsreihen kardiale Hypertrophie durch chronische Applikation von Angiotensin II induziert. Es wurde postuliert, dass die Überexpression funktionsfähiger GC-A im Herzen die Tiere vor Herzhypertrophie schützt. Die Ergebnisse der Studien zeigen allerdings, dass die generierten transgene Tiere trotz kardiomyozytenspezifischer Überexpression des Rezeptors nicht den erwarteten Schutz vor Herzhypertrophie aufwiesen, sondern ähnlich wie ihre wildtypischen Geschwistertiere reagieren. Jedoch gelang es mit Hilfe des Überexpressionsmodells zusammen mit anderen Mitarbeitern der AG Kuhn eine zuvor in vitro beschriebene Interaktion des GC-A-Rezeptors mit den Kationenkanälen TRPC3 und TRPC6 in vivo nachzuweisen. Somit besteht die Möglichkeit die Epitope und das murine Überexpressionsmodell auch zukünftig zu nutzen, um Interaktionspartner der GC-A zu identifizieren.
Extramedullary disease (EMD) represents a high-risk state of multiple myeloma (MM) associated with poor prognosis. While most anti-myeloma therapeutics demonstrate limited efficacy in this setting, some studies exploring the utility of chimeric antigen receptor (CAR)-modified T cells reported promising results. We have recently designed SLAMF7-directed CAR T cells for the treatment of MM. SLAMF7 is a transmembrane receptor expressed on myeloma cells that plays a role in myeloma cell homing to the bone marrow. Currently, the only approved anti-SLAMF7 therapeutic is the monoclonal antibody elotuzumab, but its efficacy in EMD has not been investigated thoroughly. Thus, we retrospectively analyzed the efficacy of elotuzumab-based combination therapy in a cohort of 15 patients with EMD. Moreover, since the presence of the target antigen is an indispensable prerequisite for effective targeted therapy, we investigated the SLAMF7 expression on extramedullary located tumor cells before and after treatment. We observed limited efficacy of elotuzumab-based combination therapies, with an overall response rate of 40% and a progression-free and overall survival of 3.8 and 12.9 months, respectively. Before treatment initiation, all available EMD tissue specimens (n = 3) demonstrated a strong and consistent SLAMF7 surface expression by immunohistochemistry. Furthermore, to investigate a potential antigen reduction under therapeutic selection pressure, we analyzed samples of de novo EMD (n = 3) outgrown during elotuzumab treatment. Again, immunohistochemistry documented strong and consistent SLAMF7 expression in all samples. In aggregate, our data point towards a retained expression of SLAMF7 in EMD and encourage the development of more potent SLAMF7-directed immunotherapies, such as CAR T cells.
Emerging data demonstrate that the activity of immune cells can be modulated by microbial molecules. Here, we show that the short-chain fatty acids (SCFAs) pentanoate and butyrate enhance the anti-tumor activity of cytotoxic T lymphocytes (CTLs) and chimeric antigen receptor (CAR) T cells through metabolic and epigenetic reprograming. We show that in vitro treatment of CTLs and CAR T cells with pentanoate and butyrate increases the function of mTOR as a central cellular metabolic sensor, and inhibits class I histone deacetylase activity. This reprogramming results in elevated production of effector molecules such as CD25, IFN-γ and TNF-α, and significantly enhances the anti-tumor activity of antigen-specific CTLs and ROR1-targeting CAR T cells in syngeneic murine melanoma and pancreatic cancer models. Our data shed light onto microbial molecules that may be used for enhancing cellular anti-tumor immunity. Collectively, we identify pentanoate and butyrate as two SCFAs with therapeutic utility in the context of cellular cancer immunotherapy.
Oroantral communication (OAC) is a common complication after tooth extraction of upper molars. Profound preoperative panoramic radiography analysis might potentially help predict OAC following tooth extraction. In this exploratory study, we evaluated n = 300 consecutive cases (100 OAC and 200 controls) and trained five machine learning algorithms (VGG16, InceptionV3, MobileNetV2, EfficientNet, and ResNet50) to predict OAC versus non-OAC (binary classification task) from the input images. Further, four oral and maxillofacial experts evaluated the respective panoramic radiography and determined performance metrics (accuracy, area under the curve (AUC), precision, recall, F1-score, and receiver operating characteristics curve) of all diagnostic approaches. Cohen's kappa was used to evaluate the agreement between expert evaluations. The deep learning algorithms reached high specificity (highest specificity 100% for InceptionV3) but low sensitivity (highest sensitivity 42.86% for MobileNetV2). The AUCs from VGG16, InceptionV3, MobileNetV2, EfficientNet, and ResNet50 were 0.53, 0.60, 0.67, 0.51, and 0.56, respectively. Expert 1–4 reached an AUC of 0.550, 0.629, 0.500, and 0.579, respectively. The specificity of the expert evaluations ranged from 51.74% to 95.02%, whereas sensitivity ranged from 14.14% to 59.60%. Cohen's kappa revealed a poor agreement for the oral and maxillofacial expert evaluations (Cohen's kappa: 0.1285). Overall, present data indicate that OAC cannot be sufficiently predicted from preoperative panoramic radiography. The false-negative rate, i.e., the rate of positive cases (OAC) missed by the deep learning algorithms, ranged from 57.14% to 95.24%. Surgeons should not solely rely on panoramic radiography when evaluating the probability of OAC occurrence. Clinical testing of OAC is warranted after each upper-molar tooth extraction.
Associations between periodontitis and COPD: An artificial intelligence-based analysis of NHANES III
(2022)
A number of cross-sectional epidemiological studies suggest that poor oral health is associated with respiratory diseases. However, the number of cases within the studies was limited, and the studies had different measurement conditions. By analyzing data from the National Health and Nutrition Examination Survey III (NHANES III), this study aimed to investigate possible associations between chronic obstructive pulmonary disease (COPD) and periodontitis in the general population. COPD was diagnosed in cases where FEV (1)/FVC ratio was below 70% (non-COPD versus COPD; binary classification task). We used unsupervised learning utilizing k-means clustering to identify clusters in the data. COPD classes were predicted with logistic regression, a random forest classifier, a stochastic gradient descent (SGD) classifier, k-nearest neighbors, a decision tree classifier, Gaussian naive Bayes (GaussianNB), support vector machines (SVM), a custom-made convolutional neural network (CNN), a multilayer perceptron artificial neural network (MLP), and a radial basis function neural network (RBNN) in Python. We calculated the accuracy of the prediction and the area under the curve (AUC). The most important predictors were determined using feature importance analysis. Results: Overall, 15,868 participants and 19 feature variables were included. Based on k-means clustering, the data were separated into two clusters that identified two risk characteristic groups of patients. The algorithms reached AUCs between 0.608 (DTC) and 0.953% (CNN) for the classification of COPD classes. Feature importance analysis of deep learning algorithms indicated that age and mean attachment loss were the most important features in predicting COPD. Conclusions: Data analysis of a large population showed that machine learning and deep learning algorithms could predict COPD cases based on demographics and oral health feature variables. This study indicates that periodontitis might be an important predictor of COPD. Further prospective studies examining the association between periodontitis and COPD are warranted to validate the present results.
High programmed cell death 1 ligand 1 (PD-L1) protein expression and copy number alterations (CNAs) of the corresponding genomic locus 9p24.1 in Hodgkin- and Reed–Sternberg cells (HRSC) have been shown to be associated with favourable response to anti-PD-1 checkpoint inhibition in relapsed/refractory (r/r) classical Hodgkin lymphoma (cHL). In the present study, we investigated baseline 9p24.1 status as well as PD-L1 and major histocompatibility complex (MHC) class I and II protein expression in 82 biopsies from patients with early stage unfavourable cHL treated with anti-PD-1-based first-line treatment in the German Hodgkin Study Group (GHSG) NIVAHL trial (ClinicalTrials.gov Identifier: NCT03004833). All evaluated specimens showed 9p24.1 CNA in HRSC to some extent, but with high intratumoral heterogeneity and an overall smaller range of alterations than reported in advanced-stage or r/r cHL. All but two cases (97%) showed PD-L1 expression by the tumour cells in variable amounts. While MHC-I was rarely expressed in >50% of HRSC, MHC-II expression in >50% of HRSC was found more frequently. No obvious impact of 9p24.1 CNA or PD-L1 and MHC-I/II expression on early response to the highly effective anti-PD-1-based NIVAHL first-line treatment was observed. Further studies evaluating an expanded panel of potential biomarkers are needed to optimally stratify anti-PD-1 first-line cHL treatment.
Background
Despite advances in treatment of patients with non-small cell lung cancer, carriers of certain genetic alterations are prone to failure. One such factor frequently mutated, is the tumor suppressor PTEN. These tumors are supposed to be more resistant to radiation, chemo- and immunotherapy.
Results
We demonstrate that loss of PTEN led to altered expression of transcriptional programs which directly regulate therapy resistance, resulting in establishment of radiation resistance. While PTEN-deficient tumor cells were not dependent on DNA-PK for IR resistance nor activated ATR during IR, they showed a significant dependence for the DNA damage kinase ATM. Pharmacologic inhibition of ATM, via KU-60019 and AZD1390 at non-toxic doses, restored and even synergized with IR in PTEN-deficient human and murine NSCLC cells as well in a multicellular organotypic ex vivo tumor model.
Conclusion
PTEN tumors are addicted to ATM to detect and repair radiation induced DNA damage. This creates an exploitable bottleneck. At least in cellulo and ex vivo we show that low concentration of ATM inhibitor is able to synergise with IR to treat PTEN-deficient tumors in genetically well-defined IR resistant lung cancer models.
The development of two conventional dendritic cells (DC) subsets (cDC1 and cDC2) and the plasmacytoid DC (pDC) in vivo and in cultures of bone marrow (BM) cells is mediated by the growth factor Flt3L. However, little is known about the factors that direct the development of the individual DC subsets. Here, we describe the selective in vitro generation of murine ESAM\(^{low}\) CD103\(^{-}\) XCR1\(^{-}\) CD172a\(^{+}\) CD11b\(^{+}\) cDC2 from BM by treatment with a combination of Flt3L, LIF, and IL‐10 (collectively named as FL10). FL10 promotes common dendritic cell progenitors (CDP) proliferation in the cultures, similar to Flt3L and CDP sorted and cultured in FL10 generate exclusively cDC2. These cDC2 express the transcription factors Irf4, Klf4, and Notch2, and their growth is reduced using BM from Irf4\(^{-/-}\) mice, but the expression of Batf3 and Tcf4 is low. Functionally they respond to TLR3, TLR4, and TLR9 signals by upregulation of the surface maturation markers MHC II, CD80, CD86, and CD40, while they poorly secrete proinflammatory cytokines. Peptide presentation to TCR transgenic OT‐II cells induced proliferation and IFN‐γ production that was similar to GM‐CSF‐generated BM‐DC and higher than Flt3L‐generated DC. Together, our data support that FL10 culture of BM cells selectively promotes CDP‐derived ESAM\(^{low}\) cDC2 (cDC2B) development and survival in vitro.
Background: Oro-antral communication (OAC) is a common complication following the extraction of upper molar teeth. The Archer and the Root Sinus (RS) systems can be used to classify impacted teeth in panoramic radiographs. The Archer classes B-D and the Root Sinus classes III, IV have been associated with an increased risk of OAC following tooth extraction in the upper molar region. In our previous study, we found that panoramic radiographs are not reliable for predicting OAC. This study aimed to (1) determine the feasibility of automating the classification (Archer/RS classes) of impacted teeth from panoramic radiographs, (2) determine the distribution of OAC stratified by classification system classes for the purposes of decision tree construction, and (3) determine the feasibility of automating the prediction of OAC utilizing the mentioned classification systems. Methods: We utilized multiple supervised pre-trained machine learning models (VGG16, ResNet50, Inceptionv3, EfficientNet, MobileNetV2), one custom-made convolutional neural network (CNN) model, and a Bag of Visual Words (BoVW) technique to evaluate the performance to predict the clinical classification systems RS and Archer from panoramic radiographs (Aim 1). We then used Chi-square Automatic Interaction Detectors (CHAID) to determine the distribution of OAC stratified by the Archer/RS classes to introduce a decision tree for simple use in clinics (Aim 2). Lastly, we tested the ability of a multilayer perceptron artificial neural network (MLP) and a radial basis function neural network (RBNN) to predict OAC based on the high-risk classes RS III, IV, and Archer B-D (Aim 3). Results: We achieved accuracies of up to 0.771 for EfficientNet and MobileNetV2 when examining the Archer classification. For the AUC, we obtained values of up to 0.902 for our custom-made CNN. In comparison, the detection of the RS classification achieved accuracies of up to 0.792 for the BoVW and an AUC of up to 0.716 for our custom-made CNN. Overall, the Archer classification was detected more reliably than the RS classification when considering all algorithms. CHAID predicted 77.4% correctness for the Archer classification and 81.4% for the RS classification. MLP (AUC: 0.590) and RBNN (AUC: 0.590) for the Archer classification as well as MLP 0.638) and RBNN (0.630) for the RS classification did not show sufficient predictive capability for OAC. Conclusions: The results reveal that impacted teeth can be classified using panoramic radiographs (best AUC: 0.902), and the classification systems can be stratified according to their relationship to OAC (81.4% correct for RS classification). However, the Archer and RS classes did not achieve satisfactory AUCs for predicting OAC (best AUC: 0.638). Additional research is needed to validate the results externally and to develop a reliable risk stratification tool based on the present findings.
Periodontitis is one of the most prevalent diseases worldwide. The degree of radiographic bone loss can be used to assess the course of therapy or the severity of the disease. Since automated bone loss detection has many benefits, our goal was to develop a multi-object detection algorithm based on artificial intelligence that would be able to detect and quantify radiographic bone loss using standard two-dimensional radiographic images in the maxillary posterior region. This study was conducted by combining three recent online databases and validating the results using an external validation dataset from our organization. There were 1414 images for training and testing and 341 for external validation in the final dataset. We applied a Keypoint RCNN with a ResNet-50-FPN backbone network for both boundary box and keypoint detection. The intersection over union (IoU) and the object keypoint similarity (OKS) were used for model evaluation. The evaluation of the boundary box metrics showed a moderate overlapping with the ground truth, revealing an average precision of up to 0.758. The average precision and recall over all five folds were 0.694 and 0.611, respectively. Mean average precision and recall for the keypoint detection were 0.632 and 0.579, respectively. Despite only using a small and heterogeneous set of images for training, our results indicate that the algorithm is able to learn the objects of interest, although without sufficient accuracy due to the limited number of images and a large amount of information available in panoramic radiographs. Considering the widespread availability of panoramic radiographs as well as the increasing use of online databases, the presented model can be further improved in the future to facilitate its implementation in clinics.