Convolutional neural network transformer (CNNT) for fluorescence microscopy image denoising with improved generalization and fast adaptation.

Deep neural networks can improve the quality of fluorescence microscopy images. Previous methods, based on Convolutional Neural Networks (CNNs), require time-consuming training of individual models for each experiment, impairing their applicability and generalization. In this study, we propose a novel imaging-transformer based model, Convolutional Neural Network Transformer (CNNT), that outperforms CNN based networks for image denoising. We train a general CNNT based backbone model from pairwise high-low Signal-to-Noise Ratio (SNR) image volumes, gathered from a single type of fluorescence microscope, an instant Structured Illumination Microscope. Fast adaptation to new microscopes is achieved by fine-tuning the backbone on only 5-10 image volume pairs per new experiment. Results show that the CNNT backbone and fine-tuning scheme significantly reduces training time and improves image quality, outperforming models trained using only CNNs such as 3D-RCAN and Noise2Fast. We show three examples of efficacy of this approach in wide-field, two-photon, and confocal fluorescence microscopy.

The AKT2/SIRT5/TFEB pathway as a potential therapeutic target in non-neovascular AMD.

Non-neovascular or dry age-related macular degeneration (AMD) is a multi-factorial disease with degeneration of the aging retinal-pigmented epithelium (RPE). Lysosomes play a crucial role in RPE health via phagocytosis and autophagy, which are regulated by transcription factor EB/E3 (TFEB/E3). Here, we find that increased AKT2 inhibits PGC-1α to downregulate SIRT5, which we identify as an AKT2 binding partner. Crosstalk between SIRT5 and AKT2 facilitates TFEB-dependent lysosomal function in the RPE. AKT2/SIRT5/TFEB pathway inhibition in the RPE induced lysosome/autophagy signaling abnormalities, disrupted mitochondrial function and induced release of debris contributing to drusen. Accordingly, AKT2 overexpression in the RPE caused a dry AMD-like phenotype in aging Akt2 KI mice, as evident from decline in retinal function. Importantly, we show that induced pluripotent stem cell-derived RPE encoding the major risk variant associated with AMD (complement factor H; CFH Y402H) express increased AKT2, impairing TFEB/TFE3-dependent lysosomal function. Collectively, these findings suggest that targeting the AKT2/SIRT5/TFEB pathway may be an effective therapy to delay the progression of dry AMD.

Regulation of Caenorhabditis elegans HLH-30 subcellular localization dynamics: Evidence for a redox-dependent mechanism.

Basic Helix-Loop-Helix (bHLH) transcription factors TFEB/TFE3 and HLH-30 are key regulators of autophagy induction and lysosomal biogenesis in mammals and C. elegans, respectively. While much is known about the regulation of TFEB/TFE3, how HLH-30 subcellular dynamics and transactivation are modulated are yet poorly understood. Thus, elucidating the regulation of C. elegans HLH-30 will provide evolutionary insight into the mechanisms governing the function of bHLH transcription factor family. We report here that HLH-30 is retained in the cytoplasm mainly through its conserved Ser201 residue and that HLH-30 physically interacts with the 14-3-3 protein FTT-2 in this location. The FoxO transcription factor DAF-16 is not required for HLH-30 nuclear translocation upon stress, despite that both proteins partner to form a complex that coordinately regulates several organismal responses. Similar as described for DAF-16, the importin IMB-2 assists HLH-30 nuclear translocation, but constitutive HLH-30 nuclear localization is not sufficient to trigger its distinctive transcriptional response. Furthermore, we identify FTT-2 as the target of diethyl maleate (DEM), a GSH depletor that causes a transient nuclear translocation of HLH-30. Together, our work demonstrates that the regulation of TFEB/TFE3 and HLH-30 family members is evolutionarily conserved and that, in addition to a direct redox regulation through its conserved single cysteine residue, HLH-30 can also be indirectly regulated by a redox-dependent mechanism, probably through FTT-2 oxidation.

Convolutional Neural Network Transformer (CNNT) for Fluorescence Microscopy image Denoising with Improved Generalization and Fast Adaptation.

Deep neural networks have been applied to improve the image quality of fluorescence microscopy imaging. Previous methods are based on convolutional neural networks (CNNs) which generally require more time-consuming training of separate models for each new imaging experiment, impairing the applicability and generalization. Once the model is trained (typically with tens to hundreds of image pairs) it can then be used to enhance new images that are like the training data. In this study, we proposed a novel imaging-transformer based model, Convolutional Neural Network Transformer (CNNT), to outperform the CNN networks for image denoising. In our scheme we have trained a single CNNT based "backbone model" from pairwise high-low SNR images for one type of fluorescence microscope (instance structured illumination, iSim). Fast adaption to new applications was achieved by fine-tuning the backbone on only 5-10 sample pairs per new experiment. Results show the CNNT backbone and fine-tuning scheme significantly reduces the training time and improves the image quality, outperformed training separate models using CNN approaches such as - RCAN and Noise2Fast. Here we show three examples of the efficacy of this approach on denoising wide-field, two-photon and confocal fluorescence data. In the confocal experiment, which is a 5×5 tiled acquisition, the fine-tuned CNNT model reduces the scan time form one hour to eight minutes, with improved quality.

TMEM55B links autophagy flux, lysosomal repair, and TFE3 activation in response to oxidative stress.

Lysosomes have emerged as critical regulators of cellular homeostasis. Here we show that the lysosomal protein TMEM55B contributes to restore cellular homeostasis in response to oxidative stress by three different mechanisms: (1) TMEM55B mediates NEDD4-dependent PLEKHM1 ubiquitination, causing PLEKHM1 proteasomal degradation and halting autophagosome/lysosome fusion; (2) TMEM55B promotes recruitment of components of the ESCRT machinery to lysosomal membranes to stimulate lysosomal repair; and (3) TMEM55B sequesters the FLCN/FNIP complex to facilitate translocation of the transcription factor TFE3 to the nucleus, allowing expression of transcriptional programs that enable cellular adaptation to stress. Knockout of tmem55 genes in zebrafish embryos increases their susceptibility to oxidative stress, causing early death of tmem55-KO animals in response to arsenite toxicity. Altogether, our work identifies a role for TMEM55B as a molecular sensor that coordinates autophagosome degradation, lysosomal repair, and activation of stress responses.

Cholesteryl Hemiazelate Present in Cardiovascular Disease Patients Causes Lysosome Dysfunction in Murine Fibroblasts.

There is growing evidence supporting the role of fibroblasts in all stages of atherosclerosis, from the initial phase to fibrous cap and plaque formation. In the arterial wall, as with macrophages and vascular smooth muscle cells, fibroblasts are exposed to a myriad of LDL lipids, including the lipid species formed during the oxidation of their polyunsaturated fatty acids of cholesteryl esters (PUFA-CEs). Recently, our group identified the final oxidation products of the PUFA-CEs, cholesteryl hemiesters (ChE), in tissues from cardiovascular disease patients. Cholesteryl hemiazelate (ChA), the most prevalent lipid of this family, is sufficient to impact lysosome function in macrophages and vascular smooth muscle cells, with consequences for their homeostasis. Here, we show that the lysosomal compartment of ChA-treated fibroblasts also becomes dysfunctional. Indeed, fibroblasts exposed to ChA exhibited a perinuclear accumulation of enlarged lysosomes full of neutral lipids. However, this outcome did not trigger de novo lysosome biogenesis, and only the lysosomal transcription factor E3 (TFE3) was slightly transcriptionally upregulated. As a consequence, autophagy was inhibited, probably via mTORC1 activation, culminating in fibroblasts' apoptosis. Our findings suggest that the impairment of lysosome function and autophagy and the induction of apoptosis in fibroblasts may represent an additional mechanism by which ChA can contribute to the progression of atherosclerosis.

TFEB.

Contreras and Puertollano introduce TFEB, a transcription factor that orchestrates cellular responses to stress via mechanisms including upregulation of lysosome biogenesis and autophagy.

The AKT2/SIRT5/TFEB pathway as a potential therapeutic target in atrophic AMD.

Age-related macular degeneration (AMD), the leading cause of geriatric blindness, is a multi-factorial disease with retinal-pigmented epithelial (RPE) cell dysfunction as a central pathogenic driver. With RPE degeneration, lysosomal function is a core process that is disrupted. Transcription factors EB/E3 (TFEB/E3) tightly control lysosomal function; their disruption can cause aging disorders, such as AMD. Here, we show that induced pluripotent stem cells (iPSC)-derived RPE cells with the complement factor H variant [ CFH (Y402H)] have increased AKT2, which impairs TFEB/TFE3 nuclear translocation and lysosomal function. Increased AKT2 can inhibit PGC1α, which downregulates SIRT5, an AKT2 binding partner. SIRT5 and AKT2 co-regulate each other, thereby modulating TFEB-dependent lysosomal function in the RPE. Failure of the AKT2/SIRT5/TFEB pathway in the RPE induced abnormalities in the autophagy-lysosome cellular axis by upregulating secretory autophagy, thereby releasing a plethora of factors that likely contribute to drusen formation, a hallmark of AMD. Finally, overexpressing AKT2 in RPE cells in mice led to an AMD-like phenotype. Thus, targeting the AKT2/SIRT5/TFEB pathway could be a potential therapy for atrophic AMD.

AAV-mediated delivery of secreted acid α-glucosidase with enhanced uptake corrects neuromuscular pathology in Pompe mice.

Gene therapy is under advanced clinical development for several lysosomal storage disorders. Pompe disease, a debilitating neuromuscular illness affecting infants, children, and adults with different severity, is caused by a deficiency of lysosomal glycogen-degrading enzyme acid α-glucosidase (GAA). Here, we demonstrated that adeno-associated virus-mediated (AAV-mediated) systemic gene transfer reversed glycogen storage in all key therapeutic targets - skeletal and cardiac muscles, the diaphragm, and the central nervous system - in both young and severely affected old Gaa-knockout mice. Furthermore, the therapy reversed secondary cellular abnormalities in skeletal muscle, such as those in autophagy and mTORC1/AMPK signaling. We used an AAV9 vector encoding a chimeric human GAA protein with enhanced uptake and secretion to facilitate efficient spread of the expressed protein among multiple target tissues. These results lay the groundwork for a future clinical development strategy in Pompe disease.

Oxidized cholesteryl ester induces exocytosis of dysfunctional lysosomes in lipidotic macrophages.

A key event in atherogenesis is the formation of lipid-loaded macrophages, lipidotic cells, which exhibit irreversible accumulation of undigested modified low-density lipoproteins (LDL) in lysosomes. This event culminates in the loss of cell homeostasis, inflammation, and cell death. Nevertheless, the exact chemical etiology of atherogenesis and the molecular and cellular mechanisms responsible for the impairment of lysosome function in plaque macrophages are still unknown. Here, we demonstrate that macrophages exposed to cholesteryl hemiazelate (ChA), one of the most prevalent products of LDL-derived cholesteryl ester oxidation, exhibit enlarged peripheral dysfunctional lysosomes full of undigested ChA and neutral lipids. Both lysosome area and accumulation of neutral lipids are partially irreversible. Interestingly, the dysfunctional peripheral lysosomes are more prone to fuse with the plasma membrane, secreting their undigested luminal content into the extracellular milieu with potential consequences for the pathology. We further demonstrate that this phenotype is mechanistically linked to the nuclear translocation of the MiT/TFE family of transcription factors. The induction of lysosome biogenesis by ChA appears to partially protect macrophages from lipid-induced cytotoxicity. In sum, our data show that ChA is involved in the etiology of lysosome dysfunction and promotes the exocytosis of these organelles. This latter event is a new mechanism that may be important in the pathogenesis of atherosclerosis.

Human variation impacting MCOLN2 restricts Salmonella Typhi replication by magnesium deprivation.

Human genetic diversity can reveal critical factors in host-pathogen interactions. This is especially useful for human-restricted pathogens like Salmonella enterica serovar Typhi (S. Typhi), the cause of typhoid fever. One key defense during bacterial infection is nutritional immunity: host cells attempt to restrict bacterial replication by denying bacteria access to key nutrients or supplying toxic metabolites. Here, a cellular genome-wide association study of intracellular replication by S. Typhi in nearly a thousand cell lines from around the world-and extensive follow-up using intracellular S. Typhi transcriptomics and manipulation of magnesium availability-demonstrates that the divalent cation channel mucolipin-2 (MCOLN2 or TRPML2) restricts S. Typhi intracellular replication through magnesium deprivation. Mg2+ currents, conducted through MCOLN2 and out of endolysosomes, were measured directly using patch-clamping of the endolysosomal membrane. Our results reveal Mg2+ limitation as a key component of nutritional immunity against S. Typhi and as a source of variable host resistance.

Beta-coronaviruses exploit cellular stress responses by modulating TFEB and TFE3 activity.

Beta-coronaviruses have emerged as a severe threat to global health. Undercovering the interplay between host and beta-coronaviruses is essential for understanding disease pathogenesis and developing efficient treatments. Here we report that the transcription factors TFEB and TFE3 translocate from the cytosol to the nucleus in response to beta-coronavirus infection by a mechanism that requires activation of calcineurin phosphatase. In the nucleus, TFEB and TFE3 bind to the promoter of multiple lysosomal and immune genes. Accordingly, MHV-induced upregulation of immune regulators is significantly decreased in TFEB/TFE3-depleted cells. Conversely, over-expression of either TFEB or TFE3 is sufficient to increase expression of several cytokines and chemokines. The reduced immune response observed in the absence of TFEB and TFE3 results in increased cellular survival of infected cells but also in reduced lysosomal exocytosis and decreased viral infectivity. These results suggest a central role of TFEB and TFE3 in cellular response to beta-coronavirus infection.

p38 MAPK-dependent phosphorylation of TFEB promotes monocyte-to-macrophage differentiation.

The transcription factor EB (TFEB) regulates energy homeostasis and cellular response to a wide variety of stress conditions, including nutrient deprivation, oxidative stress, organelle damage, and pathogens. Here we identify S401 as a novel phosphorylation site within the TFEB proline-rich domain. Phosphorylation of S401 increases significantly in response to oxidative stress, UVC light, growth factors, and LPS, whereas this increase is prevented by p38 MAPK inhibition or depletion, revealing a new role for p38 MAPK in TFEB regulation. Mutation of S401 in THP1 cells demonstrates that the p38 MAPK/TFEB pathway plays a particularly relevant role during monocyte differentiation into macrophages. TFEB-S401A monocytes fail to upregulate the expression of multiple immune genes in response to PMA-induced differentiation, including critical cytokines, chemokines, and growth factors. Polarization of M0 macrophages into M1 inflammatory macrophages is also aberrant in TFEB-S401A cells. These results indicate that TFEB-S401 phosphorylation links differentiation signals to the transcriptional control of monocyte differentiation.

The IRG1/itaconate/TFEB axis: A new weapon in macrophage antibacterial defense.

Zhang et al. (2022) report that itaconate, a mitochondrial metabolite produced by macrophages upon inflammatory stimuli, activates the master regulator of lysosomal biogenesis TFEB to facilitate clearance of invading bacteria and efficient immune response.

The FACT complex facilitates expression of lysosomal and antioxidant genes through binding to TFEB and TFE3.

TFEB (transcription factor EB) and TFE3 (transcription factor binding to IGHM enhancer 3) orchestrate the cellular response to a variety of stressors, including nutrient deprivation, oxidative stress and pathogens. Here we describe a novel interaction of TFEB and TFE3 with the FAcilitates Chromatin Transcription (FACT) complex, a heterodimeric histone chaperone consisting of SSRP1 and SUPT16H that mediates nucleosome disassembly and assembly, thus facilitating transcription. Extracellular stimuli, such as nutrient deprivation or oxidative stress, induce nuclear translocation and activation of TFEB and TFE3, which then associate with the FACT complex to regulate stress-induced gene transcription. Depletion of FACT does not affect TFEB activation, stability, or binding to the promoter of target genes. In contrast, reduction of FACT levels by siRNA or treatment with the FACT inhibitor curaxin, severely impairs induction of numerous antioxidant and lysosomal genes, revealing a crucial role of FACT as a regulator of cellular homeostasis. Furthermore, upregulation of antioxidant genes induced by TFEB over-expression is significantly reduced by curaxin, consistent with a role of FACT as a TFEB transcriptional activator. Together, our data show that chromatin remodeling at the promoter of stress-responsive genes by FACT is important for efficient expression of TFEB and TFE3 targets, thus providing a link between environmental changes, chromatin modifications and transcriptional regulation.Abbreviations: ADNP2, ADNP homeobox 2; ATP6V0D1, ATPase H+ transporting V0 subunit d1; ATP6V1A, ATPase H+ transporting V1 subunit A; ATP6V1C1, ATPase H+ transporting V1 subunit C1; CSNK2/CK2, casein kinase 2; CLCN7, chloride voltage-gated channel 7; CTSD, cathepsin D; CTSZ, cathepsin Z; EBSS, earle's balanced salt solution; FACT complex, facilitates chromatin transcription complex; FOXO3, forkhead box O3; HEXA, hexosaminidase subunit alpha; HIF1A, hypoxia inducible factor 1 subunit alpha; HMOX1, heme oxygenase 1; LAMP1, lysosomal associated membrane protein 1; MAFF, MAF bZIP transcription factor F; MAFG, MAF bZIP transcription factor G; MCOLN1, mucolipin TRP cation channel 1; MTORC1, mechanistic target of rapamycin kinase complex 1; NaAsO2, sodium arsenite; POLR2, RNA polymerase II; PPARGC1A, PPARG coactivator 1 alpha; PYROXD1, pyridine nucleotide-disulfide oxidoreductase domain 1; RRAGC, Ras related GTP binding C; SEC13, SEC13 homolog, nuclear pore and COPII coat complex component; SLC38A9, solute carrier family 38 member 9; SSRP1, structure specific recognition protein 1; SUPT16H, SPT16 homolog, facilitates chromatin remodeling subunit; TFEB, transcription factor EB; TFE3, transcription factor binding to IGHM enhancer 3; TXNRD1, thioredoxin reductase 1; UVRAG, UV radiation resistance associated; WDR59, WD repeat domain 59.

HSP90 inhibitors induce GPNMB cell-surface expression by modulating lysosomal positioning and sensitize breast cancer cells to glembatumumab vedotin.

Transmembrane glycoprotein NMB (GPNMB) is a prognostic marker of poor outcome in patients with triple-negative breast cancer (TNBC). Glembatumumab Vedotin, an antibody drug conjugate targeting GPNMB, exhibits variable efficacy against GPNMB-positive metastatic TNBC as a single agent. We show that GPNMB levels increase in response to standard-of-care and experimental therapies for multiple breast cancer subtypes. While these therapeutic stressors induce GPNMB expression through differential engagement of the MiTF family of transcription factors, not all are capable of increasing GPNMB cell-surface localization required for Glembatumumab Vedotin inhibition. Using a FACS-based genetic screen, we discovered that suppression of heat shock protein 90 (HSP90) concomitantly increases GPNMB expression and cell-surface localization. Mechanistically, HSP90 inhibition resulted in lysosomal dispersion towards the cell periphery and fusion with the plasma membrane, which delivers GPNMB to the cell surface. Finally, treatment with HSP90 inhibitors sensitizes breast cancers to Glembatumumab Vedotin in vivo, suggesting that combination of HSP90 inhibitors and Glembatumumab Vedotin may be a viable treatment strategy for patients with metastatic TNBC.

Chemoenzymatic glycan-selective remodeling of a therapeutic lysosomal enzyme with high-affinity M6P-glycan ligands. Enzyme substrate specificity is the name of the game.

Functionalization of therapeutic lysosomal enzymes with mannose-6-phosphate (M6P) glycan ligands represents a major strategy for enhancing the cation-independent M6P receptor (CI-MPR)-mediated cellular uptake, thus improving the overall therapeutic efficacy of the enzymes. However, the minimal high-affinity M6P-containing N-glycan ligands remain to be identified and their efficient and site-selective conjugation to therapeutic lysosomal enzymes is a challenging task. We report here the chemical synthesis of truncated M6P-glycan oxazolines and their use for enzymatic glycan remodeling of recombinant human acid α-glucosidase (rhGAA), an enzyme used for treatment of Pompe disease which is a disorder caused by a deficiency of the glycogen-degrading lysosomal enzyme. Structure-activity relationship studies identified M6P tetrasaccharide oxazoline as the minimal substrate for enzymatic transglycosylation yielding high-affinity M6P glycan ligands for the CI-MPR. Taking advantage of the substrate specificity of endoglycosidases Endo-A and Endo-F3, we found that Endo-A and Endo-F3 could efficiently deglycosylate the respective high-mannose and complex type N-glycans in rhGAA and site-selectively transfer the synthetic M6P N-glycan to the deglycosylated rhGAA without product hydrolysis. This discovery enabled a highly efficient one-pot deglycosylation/transglycosylation strategy for site-selective M6P-glycan remodeling of rhGAA to obtain a more homogeneous product. The Endo-A and Endo-F3 remodeled rhGAAs maintained full enzyme activity and demonstrated 6- and 20-fold enhanced binding affinities for CI-MPR receptor, respectively. Using an in vitro cell model system for Pompe disease, we demonstrated that the M6P-glycan remodeled rhGAA greatly outperformed the commercial rhGAA (Lumizyme) and resulted in the reversal of cellular pathology. This study provides a general and efficient method for site-selective M6P-glycan remodeling of recombinant lysosomal enzymes to achieve enhanced M6P receptor binding and cellular uptake, which could lead to improved overall therapeutic efficacy of enzyme replacement therapy.

How Lysosomes Sense, Integrate, and Cope with Stress.

Lysosomes are in the center of the cellular control of catabolic and anabolic processes. These membrane-surrounded acidic organelles contain around 70 hydrolases, 200 membrane proteins, and numerous accessory proteins associated with the cytosolic surface of lysosomes. Accessory and transmembrane proteins assemble in signaling complexes that sense and integrate multiple signals and transmit the information to the nucleus. This communication allows cells to respond to changes in multiple environmental conditions, including nutrient levels, pathogens, energy availability, and lysosomal damage, with the goal of restoring cellular homeostasis. This review summarizes our current understanding of the major molecular players and known pathways that are involved in control of metabolic and stress responses that either originate from lysosomes or regulate lysosomal functions.