Background: It is a critical challenge to diagnose leptomeningeal metastasis (LM), given its technical difficulty and the lack of typical symptoms. The existing gold standard of diagnosing LM is to use positive cerebrospinal fluid (CSF) cytology, which consumes significantly more time to classify cells under a microscope. Objective: This study aims to establish a deep learning model to classify cancer cells in CSF, thus facilitating doctors to achieve an accurate and fast diagnosis of LM in an early stage. Method: The cerebrospinal fluid laboratory of Xijing Hospital provides 53,255 cells from 90 LM patients in the research. We used two deep convolutional neural networks (CNN) models to classify cells in the CSF. A five-way cell classification model (CNN1) consists of lymphocytes, monocytes, neutrophils, erythrocytes, and cancer cells. A four-way cancer cell classification model (CNN2) consists of lung cancer cells, gastric cancer cells, breast cancer cells, and pancreatic cancer cells. Here, the CNN models were constructed by Resnet-inception-V2. We evaluated the performance of the proposed models on two external datasets and compared them with the results from 42 doctors of various levels of experience in the human-machine tests. Furthermore, we develop a computer-aided diagnosis (CAD) software to generate cytology diagnosis reports in the research rapidly. Results: With respect to the validation set, the mean average precision (mAP) of CNN1 is over 95% and that of CNN2 is close to 80%. Hence, the proposed deep learning model effectively classifies cells in CSF to facilitate the screening of cancer cells. In the human-machine tests, the accuracy of CNN1 is similar to the results from experts, with higher accuracy than doctors in other levels. Moreover, the overall accuracy of CNN2 is 10% higher than that of experts, with a time consumption of only one-third of that consumed by an expert. Using the CAD software saves 90% working time of cytologists. Conclusion: A deep learning method has been developed to assist the LM diagnosis with high accuracy and low time consumption effectively. Thanks to labeled data and step-by-step training, our proposed method can successfully classify cancer cells in the CSF to assist LM diagnosis early. In addition, this unique research can predict cancer’s primary source of LM, which relies on cytomorphologic features without immunohistochemistry. Our results show that deep learning can be widely used in medical images to classify cerebrospinal fluid cells. For complex cancer classification tasks, the accuracy of the proposed method is significantly higher than that of specialist doctors, and its performance is better than that of junior doctors and interns. The application of CNNs and CAD software may ultimately aid in expediting the diagnosis and overcoming the shortage of experienced cytologists, thereby facilitating earlier treatment and improving the prognosis of LM.

Retinitis Pigmentosa represents a group of genetic disorders that cause progressive vision loss via degeneration of photoreceptors, but there is in principle no treatment available. For any therapy development, a deeper comprehension of the disease-leading mechanism(s) at the molecular level is needed. Here we focused on the cGMP-PKG system, which has been suggested to be a driver in several models of the disease. To gain insights in its downstream signaling we manipulated the cGMP-PKG system with the aid of organotypic retinal explant cultures from either a mouse-based disease model, i.e. the rd1 mouse, or its healthy wild-type counterpart (wt), by adding different types of cGMP analogues to either inhibit or activate PKG in retinal explants from rd1 and wt, respectively. An RNA sequencing was then performed to study the cGMP-PKG dependent transcriptome. Expression changes of gene sets related to specific pathways or functions, that fulfilled criteria involving that the changes should match PKG activation and inhibition, were determined via bioinformatics. The analyses highlighted that several gene sets linked to oxidative phosphorylation and mitochondrial pathways were regulated by this enzyme system. Specifically, the expression of such pathway components was upregulated in the rd1 treated with PKG inhibitor and downregulated in the wt with PKG activator treatment, suggesting that cGMP-PKG act as a negative regulator in this context. Downregulation of energy production pathways may thus play an integral part in the mechanism behind the degeneration for at least several RP mutations.

Retinitis pigmentosa (RP) is an inherited disorder that results in vision impairment but general and mutation-independent therapeutic strategies are not available. However, it is widely regarded that the cGMP system, including cGMP and its interactor cGMP-dependent protein kinase (PKG), acts as a crucial effector during retinal degeneration. We have previously identified a list of cGMP-PKG-dependent genes in the context of RP, and in this study, we further validated one of these, namely pyruvate kinase 2 (PKM2), and investigated the potential role of PKM2 for the photoreceptors' well-being during RP. With the aid of organotypic retinal explant cultures, we pharmacologically manipulated the PKM2 activities in two different RP mouse models (rd2 and rd10) via the addition of TEPP-46 (a PKM2 activator) and found that activation of PKM2 alleviates the progress of photoreceptor death in the rd10 mouse model. We also noted that the expression of both PKM2 and one of its targets, glucose transporter-1 (Glut1), showed alterations depending on the degeneration state. The observations provide supportive evidence that PKM2 may serve as a novel potential molecular target in RP.

The disease retinitis pigmentosa (RP) leads to photoreceptor degeneration by a yet undefined mechanism(s). In several RP mouse models (i.e., rd mice), a high cyclic GMP (cGMP) level within photoreceptors is detected, suggesting that cGMP plays a role in degeneration. The rap guanine exchange factor 4 (EPAC2) is activated by cyclic AMP (cAMP) and is an accepted cGMP-interacting protein. It is unclear whether and how cGMP interacts with EPAC2 in degenerating photoreceptors; we therefore investigated EPAC2 expression and interactions with cGMP and cAMP in retinas of the rd1 and rd10 models for retinal degeneration. EPAC2 expression in the photoreceptor layer increased significantly during rd1 and rd10 degeneration, and an increase in EPAC2 interactions with cGMP but not cAMP in the rd1 was also seen via a proximity ligation assay on histological sections. Retinal explant cultures revealed that pharmacological inhibition of the EPAC2 activity reduced the photoreceptor layer thickness in the rd10 retina, suggesting that EPAC2 inhibition promotes degeneration. Taken together, our results support the hypothesis that high degeneration-related cGMP leads to increased EPAC2 and cGMP interactions, inhibiting EPAC2. By inference, EPAC2 could have neuroprotective capacities that may be exploited in the future.

Cyclin dependent kinase 1 (CDK1) has long been known to drive the cell cycle and to regulate the division and differentiation of cells. Apart from its role in mitosis regulation, it also exerts multiple functions as a protein kinase, including engagement in cell death, possibly via a cell cycle-independent mechanism. The latter is suggested, since CDK1 re-expression can be found in non-dividing and terminally differentiated neurons in several neurodegeneration models. However, the details of if and how CDK1 might be involved in the neurodegenerative condition, retinitis pigmentosa (RP), which displays progressive vision loss, are unclear. In the present study, we investigated CDK1 in degenerating RP photoreceptors of the rd1 RP model, including whether there is a link between this kinase and the cGMP-PKG system, which is regarded as a disease driver. With experiments performed using either in vivo retinal tissue or in vitro material, via organotypic retinal explants, our results showed that CDK1 appears in the photoreceptors at a late stage of their degeneration, and in such a position, it may be associated with the cGMP-PKG network.