Biology subjects

Zhang, M.

Publications and source records attributed to Zhang, M..

At least 631 records · Page 35Linked to original sources

bioRxiv2020

Spatial frequency-based image reconstruction to improve image contrast in multi-offset adaptive optics ophthalmoscopy

Off-axis detection methods in adaptive optics (AO) ophthalmoscopy can enhance image contrast of translucent retinal structures such as cone inner segments and retinal ganglion cells layer neurons. Here, we propose a 2D optical model showing that the phase contrast produced by these methods depends on the offset orientation. While one axis provides an asymmetric light distribution, hence a high phase contrast, the perpendicular axis provides a symmetric one, thus a substantially lower contrast. We support this model with in-vivo human data acquired with a multi-offset AO scanning light ophthalmoscope. Then, using this finding, we provide a post-processing method, named Spatial frequency-based iMAge ReconsTruction (SMART), to optimally combine images from different off-axis detector orientations, significantly increasing the structural cellular contrast of in-vivo human retinal neurons such as conne inner segment, putative rods and retinal ganglion cells.

bioengineering↗

bioRxiv2020

The endosymbiotic origins of the apicoplast link fever-survival and artemisinin-resistance in the malaria parasite

BackgroundThe emergence and spread of Plasmodium falciparum parasites resistant to front-line antimalarial artemisinin-combination therapies (ACT) threatens to erase the considerable gains against the disease of the last decade. We developed a new large-scale phenotypic screening pipeline and used it to carry out the first large-scale forward-genetic phenotype screen in P. falciparum to identify genes that allow parasites to survive febrile temperatures. ResultsScreening identified more than 200 P. falciparum mutants with differential responses to increased temperature. These mutants were more likely to be sensitive to artemisinin derivatives as well as to heightened oxidative stress. Major processes critical for P. falciparum tolerance to febrile temperatures and artemisinin included highly essential, conserved pathways associated with protein-folding, heat-shock and proteasome-mediated degradation, and unexpectedly, isoprenoid biosynthesis, which originated from the ancestral genome of the parasites algal endosymbiont-derived plastid, the apicoplast. Apicoplast-targeted genes in general were up-regulated in response to heat shock, as were other Plasmodium genes with orthologs in plant and algal genomes. ConclusionsPlasmodium falciparum parasites appear to exploit their innate febrile-response mechanisms to mediate resistance to artemisinin. Both responses depend on endosymbiotic cynobacterium-related ancestral genes in the parasites genome, suggesting a link to the evolutionary origins of Plasmodium parasites in free-living ancestors.

genomics↗

bioRxiv2020

Improvements to multi-offset adaptive optics scanning laser ophthalmoscopy for in vivo imaging of individual retinal ganglion cell layer neurons in humans

Previous work has shown that multi-offset detection in adaptive optics scanning laser ophthalmoscopy (AOSLO) can be used to image retinal ganglion cells (RGCs) in monkeys and humans. However, though images of RGCs in anesthetized monkeys with high light levels produced high contrast images of RGCs, images from humans failed to reach the same contrast due to several drawbacks in the previous dual-wavelength multi-offset approach. Our aim here was to design and build a multi-offset detection pattern for humans at safe light levels that could reveal the retinal ganglion cell layer neurons with a contrast, robustness and acquisition time approaching results only previously obtained in monkeys. Here, we present a new imaging system using only one light source, compared to the previous dual-wavelength used on monkeys. Our single-wavelength solution allows for increased light power and eliminates problematic chromatic aberrations. Then, we demonstrate that a radial multi-offset detection pattern with an offset distance of 8-10 Airy Disk Diameter (ADD) is optimal to detect photons multiply scattered in all directions from RGCs thereby enhancing their contrast. This new setup and image processing pipeline led to improved imaging of retinal ganglion cells using multi-offset imaging in AOSLO.

bioengineering↗

bioRxiv2020

Improvements to strip-based digital image registration for robust eye-tracking and to minimize distortions in images from scanned ophthalmic imaging systems

Retinal image-based eye tracking from scanned ophthalmic imaging systems, such as scanning laser ophthalmoscopy, has allowed for precise real-time eye tracking at sub-micron resolution. To achieve real-time processing rates, strip-based image registration methods for real-time applications have several constraints that limit their performance. This trade-off is acceptable for many imaging and psychophysical applications but when the objective is precise eye motion measurement over time, a high error tolerance can be consequential. Dropped strips in these applications can complicate FEMs quantification. Some light starved imaging applications, such as autofluorescence retinal imaging, also require the retention and registration of as much of the data as possible to increase the signal to noise ratio in the final integrated or averaged image. We show here that eye motion can be extracted from image sequences from scanned imaging systems more consistently when the constraints of real-time processing are lifted, and all data is available at the time of registration. This is enabled with additional image processing steps to achieve a more robust solution. Our iterative approach identifies and discards distorted frames, detects coarse motion to generate a synthetic reference frame and then uses it for fine scale motion tracking with improved sensitivity over a larger area. We demonstrate its application here to tracking scanning laser ophthalmoscopy (TSLO) and adaptive optics scanning light ophthalmoscopy (AOSLO). We show that it can successfully capture most of the eye motion across each image sequence, leaving only between 0.04-3.39% of non-blink frames untracked, even with low quality images, while simultaneously minimizing image distortions induced from eye motion. These improvements will facilitate precise FEMs measurement in TSLO and longitudinal tracking of individual cells in AOSLO.

bioengineering↗

bioRxiv2020

Species-resolved sequencing of low-biomass microbiomes by 2bRAD-M

Microbiome samples with low microbial biomass or severe DNA degradation remain challenging for amplicon-based (e.g., 16S/18S-rRNA) or whole-metagenome sequencing (WMS) approaches. Here, we introduce 2bRAD-M, a highly reduced and cost-effective metagenome-sequencing strategy which only sequences ~1% of metagenome and can simultaneously produce species-level bacterial, archaeal, and fungal profiles for low-biomass and highly degraded samples. For mock communities, 2bRAD-M can accurately generate species-level taxonomic profiles for otherwise hard-to-sequence samples with (i) low biomass of merely 1 pg of total DNA, (ii) high host DNA contamination (99%), and (iii) severely fragmented DNA (50-bp) from degraded samples. Tests of 2bRAD-M on stool, skin and environment-surface samples deliver successful reconstruction of comprehensive, high-resolution microbial profiles with agreement across 16S-rRNA, WMS and existing literature. In addition, it enables microbial profiling in formalin-fixed paraffin-embedded (FFPE) cervical tissue samples which were recalcitrant to conventional approaches due to the low amount and heavy degradation of microbial DNA, and discriminated healthy tissue, pre-invasive cancer and invasive cancer via species-level microbial profiles with 91.1% accuracy. Therefore, 2bRAD-M greatly expands the reach of microbiome sequencing.

microbiology↗

bioRxiv2020

3D Genome Contributes to Protein-Protein Interactome

Numerous computational methods have been proposed to predict protein-protein interactions, none of which however, considers the original DNA loci of the interacting proteins in the perspective of 3D genome. Here we retrospect the DNA origins of the interacting proteins in the context of 3D genome and discovered that 1) if a gene pair is more proximate in 3D genome, their corresponding proteins are more likely to interact. 2) signal peptide involvement of PPI affects the corresponding gene-gene proximity in 3D genome space. 3) by incorporating 3D genome information, existing PPI prediction methods can be further improved in terms of accuracy. Combining our previous discoveries, we conjecture the existence of 3D genome driven cellular compartmentalization, meaning the co-localization of DNA elements lead to increased probability of the co-localization of RNA elements and protein elements.

bioinformatics↗

bioRxiv2020

CaMKII activation triggers persistent formation and segregation of postsynaptic liquid phase

Transient information input to brain leads to persistent changes in synaptic circuit, thereby forming memory engrams. Synapse undergoes coordinated functional and structural changes during this process but how such changes are achieved by its component molecules still largely remain enigmatic. We found that activated CaMKII, the central player of synaptic plasticity, undergoes liquid-liquid phase separation (LLPS) with NMDAR subunit GluN2B. Due to CaMKII autophosphorylation, the condensate stably persists even after Ca2+ is removed. The selective binding of activated CaMKII with GluN2B co-segregates AMPAR/neuroligin (NLGN) into a phase-in-phase assembly. Because postsynaptic NLGN clusters presynaptic neurexin and other active zone proteins thereby increasing the release probability of synaptic vesicles, this ensures efficient synaptic transmission. In this way, Ca2+-induced and persistent formation of LLPS by CaMKII serves as molecular basis of memory by functioning as an activity-dependent crosslinker for postsynaptic proteins and segregating trans-synaptic nanocolumns.

neuroscience↗

bioRxiv2020

Age-Associated Insolubility of Parkin in Human Midbrain is Linked to Redox Balance and Sequestration of Reactive Dopamine Metabolites

The mechanisms by which parkin protects the adult human brain from Parkinson disease remain incompletely understood. We hypothesized that parkin cysteines participate in redox reactions, which are reflected in its posttranslational modifications. We found that in human control brain, including the S. nigra, parkin is largely insoluble after age 40 years, which is linked to its oxidation, e.g., at Cys95 and Cys253. In mice, oxidative stress increases posttranslational modifications at parkin cysteines and reduces its solubility. Oxidation of recombinant parkin also promotes insolubility and aggregate formation, but in parallel, lowers hydrogen peroxide (H2O2). This thiol-based redox activity is diminished by parkin point mutants, e.g., p.C431F and p.G328E. Intriguingly, in parkin-deficient human brain H2O2 concentrations are elevated. In prkn-null mice, H2O2 levels are dysregulated under oxidative stress conditions, such as acutely by MPTP-toxin exposure or chronically due to a second genetic hit. In dopamine toxicity studies, wild-type parkin, but not disease-linked mutants, protects human dopaminergic M17 cells, in part through lowering H2O2. Parkin also neutralizes reactive, electrophilic dopamine metabolites via adduct formation, which occurs foremost at primate-specific Cys95. Further, wild-type but not p.C95A-mutant parkin augments melanin formation. In sections of normal, adult human midbrain, parkin specifically co-localizes with neuromelanin pigment, frequently within LAMP-3/CD63+ lysosomes. We conclude that oxidative modifications of parkin cysteines are associated with protective outcomes, which include the reduction of H2O2, conjugation of reactive dopamine metabolites, sequestration of radicals within insoluble aggregates, and increased melanin formation. The loss of these redox effects may augment oxidative stress in dopamine producing neurons of mutant PRKN allele carriers, thereby contributing to neurodegeneration.

neuroscience↗

bioRxiv2020

scPretrain: Multi-task self-supervised learning for cell type classification

Rapidly generated scRNA-seq datasets enable us to understand cellular differences and the function of each individual cell at single-cell resolution. Cell type classification, which aims at characterizing and labeling groups of cells according to their gene expression, is one of the most important steps for single-cell analysis. To facilitate the manual curation process, supervised learning methods have been used to automatically classify cells. Most of the existing supervised learning approaches only utilize annotated cells in the training step while ignoring the more abundant unannotated cells. In this paper, we proposed scPretrain, a multi-task self-supervised learning approach that jointly considers annotated and unannotated cells for cell type classification. scPretrain consists of a pre-training step and a fine-tuning step. In the pre-training step, scPretrain uses a multi-task learning framework to train a feature extraction encoder based on each datasets pseudo-labels, where only unannotated cells are used. In the fine-tuning step, scPretrain fine-tunes this feature extraction encoder using the limited annotated cells in a new dataset. We evaluated scPretrain on 60 diverse datasets from different technologies, species and organs, and obtained a significant improvement on both cell type classification and cell clustering. Moreover, the representations obtained by scPretrain in the pre-training step also enhanced the performance of conventional classifiers such as random forest, logistic regression and support vector machines. scPretrain is able to effectively utilize the massive amount of unlabelled data and be applied to annotating increasingly generated scRNA-seq datasets. Availabilityhttps://github.com/ruiyi-zhang/scPretrain\

bioinformatics↗

bioRxiv2020

The multi-scale architecture of mammalian sperm flagella and implications for ciliary motility

Motile cilia are molecular machines used by a myriad of eukaryotic cells to swim through fluid environments. However, available molecular structures represent only a handful of cell types, limiting our understanding of how cilia are modified to support motility in diverse media. Here, we use cryo-focused ion beam milling-enabled cryo-electron tomography to image sperm flagella from three mammalian species. We resolve in-cell structures of centrioles, axonemal doublets, central pair apparatus, and endpiece singlets, revealing novel protofilament-bridging microtubule inner proteins throughout the flagellum. We present native structures of the flagellar base, which is crucial for shaping the flagellar beat. We show that outer dense fibers are directly coupled to microtubule doublets in the principal piece but not in the midpiece. Thus, mammalian sperm flagella are ornamented across scales, from protofilament-bracing structures rein-forcing microtubules at the nano-scale to accessory structures that impose micron-scale asymmetries on the entire assembly. Our structures provide vital foundations for linking molecular structure to ciliary motility and evolution.

cell biology↗

bioRxiv2020

Challenges in Benchmarking Metagenomic Profilers

Accurate microbial identification and abundance estimation are crucial for metagenomics analysis. Various methods for classifying metagenomic data and estimating taxonomic profiles, broadly referred to as metagenomic profilers, have been developed. Yet, benchmarking metagenomic profilers remains challenging because some tools are designed to report relative sequence abundance while others report relative taxonomic abundance. Here, we show how misleading conclusions can be drawn by neglecting this distinction between relative abundance types when benchmarking metagenomic profilers. Moreover, we show compelling evidence that interchanging sequence abundance and taxonomic abundance will influence both per-sample summary statistics and cross-sample comparisons. We suggest that the microbiome research community should pay attention to potentially misleading biological conclusions arising from this issue when benchmarking metagenomic profilers, by carefully considering the type of abundance data that was analyzed and interpreted, and clearly stating the strategy used for metagenomic profiling.

bioinformatics↗

bioRxiv2020

Performance of deep learning restoration methods for the extraction of particle dynamics in noisy microscopy image sequences

Image-based particle tracking is an essential tool to answer research questions in cell biology and beyond. A major challenge of particle tracking in living systems is that low light exposure is required to avoid phototoxicity and photobleaching. In addition, high-speed imaging used to fully capture particle motion dictates fast image acquisition rates. Short exposure times come at the expense of tracking accuracy. This is generally true for quantitative microscopy approaches and particularly relevant to single molecule tracking where the number of photons emitted from a single chromophore is limited. Image restoration methods based on deep learning dramatically improve the signal-to-noise ratio in low-exposure datasets. However, it is not clear whether images generated by these methods yield accurate quantitative measurements such as diffusion parameters in (single) particle tracking experiments. Here, we evaluate the performance of two popular deep learning denoising software packages for particle tracking, using synthetic datasets and movies of diffusing chromatin as biological examples. With synthetic data, both supervised and unsupervised deep learning restored particle motions with high accuracy in two-dimensional datasets, whereas artifacts were introduced by the denoisers in 3D datasets. Experimentally, we found that, while both supervised and unsupervised approaches improved the number of trackable particles and tracking accuracy, supervised learning generally outperformed the unsupervised approach, as expected. We also highlight that with extremely noisy image sequences, deep learning algorithms produce deceiving artifacts, which underscores the need to carefully evaluate the results. Finally, we address the challenge of selecting hyper-parameters to train convolutional neural networks by implementing a frugal Bayesian optimizer that rapidly explores multidimensional parameter spaces, identifying networks yielding optional particle tracking accuracy. Our study provides quantitative outcome measures of image restoration using deep learning. We anticipate broad application of the approaches presented here to critically evaluate artificial intelligence solutions for quantitative microscopy.

biophysics↗

bioRxiv2020

An improved comprehensive strategy for deep and quantitative N-glycomics based on optimization of sample preparation, isotope-based data quality control and quantification, new N-glycan libraries and new algorithms

Withdrawal statement"The authors have withdrawn their manuscript because the authors need to re-organize the data and writing, meanwhile more experimental evidence from human cellular N-glycome will be added to support their viewpoints. In this preprint, only one human cell line, Acute Promyelocytic Leukemia (APL) cells, was deeply investigated to show the N-glycome landscape. To provide the shared features of human N-glycome, more human cells should be investigated. Therefore, the authors do not wish this work to be cited as reference for the project. If you have any questions, please contact the corresponding author."

biochemistry↗

bioRxiv2020

Integrated characterization of SARS-CoV-2 genome, microbiome, antibiotic resistance and host response from single throat swabs

The ongoing coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, poses a severe threat to humanity. Rapid and comprehensive analysis of both pathogen and host sequencing data is critical to track infection and inform therapies. In this study, we performed unbiased metatranscriptomic analysis of clinical samples from COVID-19 patients using a newly-developed RNA-seq library construction method (TRACE-seq), which utilizes tagmentation activity of Tn5 on RNA/DNA hybrids. This approach avoids the laborious and time-consuming steps in traditional RNA-seq procedure, and hence is fast, sensitive and convenient. We demonstrated that TRACE-seq allowed integrated characterization of full genome information of SARS-CoV-2, putative pathogens causing coinfection, antibiotic resistance and host response from single throat swabs. We believe that the integrated information will deepen our understanding of pathogenesis and improve diagnostic accuracy for infectious diseases.

microbiology↗

bioRxiv2020

Identification of a peroxidase inhibitor that enhances kanamycin activity

BackgroundThe threat of antimicrobial resistance calls for more efforts in basic science, drug discovery, and clinical development, particularly gram-negative carbapenem-resistant pathogens. Objectives and methodsWhole-cell-based screening was performed to identify novel antibacterial agents against Acinetobacter baumannii ATCC19606. Spontaneously resistant mutant selection, whole-genome sequencing, and surface plasmon resonance were used for target identification and confirmation. Checkerboard titration assay was used for drug combination analysis. ResultsA small molecule named 6D1 with the chemical structure of 6-fluorobenzo[d]isothiazol-3(2H)-one was identified and exhibited activity against A. baumannii ATCC19606 strain (minimal inhibitory concentration, MIC = 1 mg/L). The mutation in the plasmid-derived ohrB gene that encodes a peroxidase was identified in spontaneously resistant mutants. Treatment of the bacteria with 6D1 resulted in increased sensitivity to peroxide such as tert-butyl hydroperoxide. The binding of 6D1 and OhrB was confirmed by surface plasmon resonance. Interestingly, the MIC of kanamycin against spontaneously resistant mutants decreased. Finally, we identified the effect of 6D1 on enhancing the antibacterial activity of kanamycin, including New Delhi metallo-{beta}-lactamase (NDM-1)-producing carbapenem-resistant Klebsiella pneumoniae, but not in strains carrying kanamycin resistance genes. ConclusionsIn this study, we identified a peroxidase inhibitor that suppresses the growth of A. baumannii and enhances the antibacterial activity of kanamycin. We propose that peroxidase may be potentially used as a target for kanamycin adjuvant development.

microbiology↗

bioRxiv2020

CIGAR-seq, a CRISPR/Cas-based method for unbiased screening of novel mRNA modification regulators

Cellular RNA is decorated with over 170 types of chemical modifications. Many modifications in mRNA, including m6A and m5C, have been associated with critical cellular functions under physiological and/or pathological conditions. To understand the biological functions of these modifications, it is vital to identify the regulators that modulate the modification rate. However, a high-throughput method for unbiased screening of these regulators is so far lacking. Here, we report such a method combining pooled CRISPR screen and reporters with RNA modification readout, termed CRISPR integrated gRNA and reporter sequencing (CIGAR-seq). Using CIGAR-seq, we discovered NSUN6 as a novel mRNA m5C methyltransferase. Subsequent mRNA bisulfite sequencing in HAP1 cells without or with NSUN6 and/or NSUN2 knockout showed that NSUN6 and NSUN2 worked on non-overlapping subsets of mRNA m5C sites, and together contributed to almost all the m5C modification in mRNA. Finally, using m1A as an example, we demonstrated that CIGAR-seq can be easily adapted for identifying regulators of other mRNA modification.

molecular biology↗

bioRxiv2020

Distinct default mode network subsystems show similarities and differences in the effect of task focus across reading and autobiographical memory

While reading, the mind can wander to unrelated autobiographical information, creating a perceptually-decoupled state detrimental to narrative comprehension. To understand how this mind-wandering state emerges, we asked whether retrieving autobiographical content necessitates functional disengagement from visual input. In Experiment 1, brain activity was recorded using functional magnetic resonance imaging (fMRI) in an experimental situation mimicking naturally occurring mind-wandering, allowing us to precisely delineate neural regions involved in memory and reading. Individuals read expository texts and ignored personally relevant autobiographical memories, as well as the opposite situation. Medial regions of the default mode network (DMN) were recruited during memory retrieval. In contrast, left temporal and lateral prefrontal regions of the DMN, as well as ventral visual cortex, were recruited when reading for comprehension. Experiment 2 used functional connectivity at rest to establish that (i) DMN regions linked to memory are more functionally decoupled from regions of ventral visual cortex than regions in the same network engaged when reading, and (ii) individuals reporting more mind-wandering and worse comprehension, while reading in the lab, showed increased functional decoupling between visually-connected DMN sites important for reading and a region of dorsal occipital cortex linked to autobiographical memory in Experiment 1. These data suggest we lose track of the narrative when our mind wanders because the generation of autobiographical mental content relies on cortical regions within the DMN which are functionally decoupled from ventral visual regions engaged during reading. Significance statementWhen the mind wanders during reading, we lose track of information from the narrative. We hypothesised that poor comprehension occurs because retrieving autobiographical memories reduces the perceptual coupling necessary to understand written words. We show that default mode network (DMN) areas involved in reading are functionally more connected to ventral visual regions than DMN regions important for autobiographical memory. Furthermore, individuals who mind-wander more, and comprehend less, have weaker connectivity between visually-coupled DMN regions linked to reading and dorsal occipital areas linked to autobiographical memory. These data suggest that when our minds wander during reading, retrieval of personally-relevant information activates DMN regions that are functionally disconnected from visual input, creating a perceptually decoupled state detrimental to comprehension.

neuroscience↗

bioRxiv2020

Regulation of lipid accumulation-induced ROS in myeloid-derived suppressor cells via targeting fatty-acid transport protein 2 enhanced anti-PD-L1 tumor immunotherapy

Despite the remarkable success and efficacy of immune checkpoint blockade (ICB) therapy against the PD-1/PD-L1 axis, it induces sustained responses in a sizeable minority of cancer patients due to the activation of immunosuppressive factors such as myeloid-derived suppressor cells (MDSCs). Inhibiting the immunosuppressive function of MDSCs is critical for successful cancer ICB therapy. Interestingly, lipid metabolism is a crucial factor in modulating MDSCs function. Fatty acid transport protein 2 (FATP2) conferred the function of PMN-MDSCs in cancer via the upregulation of arachidonic acid metabolism. However, whether regulating lipid accumulation in MDSCs by targeting FATP2 could block MDSCs reactive oxygen species (ROS) production and enhance PD-L1 blockade-mediated tumor immunotherapy remains unexplored. Here we report that FATP2 regulated lipid accumulation, ROS, and immunosuppressive function of MDSCs in tumor-bearing mice. Tumor cells-derived granulocyte macrophage-colony stimulating factor (GM-CSF) induced FATP2 expression in MDSCs by activation of STAT3 signaling pathway. Pharmaceutical blockade of FATP2 expression in MDSCs by lipofermata decreased lipid accumulation, reduced ROS, blocked immunosuppressive activity, and consequently inhibited tumor growth. More importantly, lipofermata inhibition of FATP2 in MDSCs enhanced anti-PD-L1 tumor immunotherapy via the upregulation of CD107a and reduced PD-L1 expression on tumor-infiltrating CD8+ T-cells. Furthermore, the combination therapy blocked MDSCs suppressive role on T-cells thereby enhanced T-cells ability for the production of IFN-{gamma}. These findings indicate that FATP2 plays a key role in modulating lipid-induced ROS in MDSCs and targeting FATP2 in MDSCs provides a novel therapeutic approach to enhance anti-PD-L1 cancer immunotherapy. Research highlightsO_LIFATP2 inhibition in MDSCs blocked ROS-mediated immunosuppressive function and promoted MDSCs differentiation to immune-stimulatory phenotype. C_LIO_LIGM-CSF mediated FATP2 expression via activation of STAT3 signaling leading to lipid accumulation-induced ROS in MDSCs. C_LIO_LIFATP2 inhibition enhanced anti-PD-L1 tumor immunotherapy by reducing MDSCs immunosuppressive activity in blocking T-cells ability to produce IFN-{gamma}. C_LIO_LICombination of anti-PD-L1 antibody and FATP2 inhibitor decreased MDSCs accumulation and PD-L1 surface expression in the spleens of LLC-bearing mice. C_LIO_LICombined treatment of FATP2 inhibitor and PD-L1 blockade-mediated tumor immunotherapy enhanced tumor-infiltrating CD8+ T cells activation via increased CD107a and reduced PD-L1 surface expression. C_LI

immunology↗