0594 Can a Deep Convolutional Neural Network Extract Diagnostic Information on Obstructive Sleep Apnea from Images?

Introduction

Lateral cephalometric radiography is a simple way to provide craniofacial soft/hard tissue profiles specific for patients with obstructive sleep apnea (OSA) and may thus offer diagnostic information on the disease. We hypothesized that a machine learning technology, a deep convolutional neural network (DCNN), could make it possible to detect OSA based solely on lateral cephalometric radiographs without the need for either large amounts of subjective/laboratory data or skilled analyses.

Methods

In this diagnostic study, a DCNN was developed (n=1,258) and tested (n=131) using data from 1,389 lateral cephalometric radiographs obtained from individuals diagnosed with severe OSA (n=867; apnea hypopnea index >30/hour) or non-OSA (n=522; apnea hypopnea index < 5) at a single center for sleep disorders from March, 2006 to February, 2017. Three kinds of data sets were prepared by changing the area of interest using a single image; original image without any modification (Full Image), image containing a facial profile, upper airway, craniofacial soft/hard tissues, and image containing part of the occipital region (upper left corner of the image; Head Only). A radiologist and an orthodontist also performed a manual cephalometric analysis of the Full Image for comparison. Observers were blinded to the patient groupings. Data analysis was performed from April, 2018 to August, 2019. When the predictive score obtained from the DCNN analysis exceeded the threshold (0.50), the patient was judged to have OSA. The primary outcome was diagnostic accuracy in terms of area under the receiver-operating characteristic curve.

Results

The sensitivity/specificity was 0.87/0.82 for Full Image, 0.88/0.75 for Main Region, 0.71/0.63 for Head Only, and 0.54/0.80 for the manual analysis. The area under the curve was the highest for Main Region (0.92): 0.89 for Full Image, 0.70 for Head Only, and 0.75 for the manual analysis.

Conclusion

A DCNN identified individuals with OSA with high accuracy. This is a useful approach that does not require any laborious analyses in a primary care setting or in remote areas where an initial specialized OSA diagnosis is not feasible.

Support

This study was supported in part by the Japan Society for the Promotion of Science (grant numbers 17K11793, 19K10236).

Prediction of oral appliance treatment outcome in obstructive sleep apnoea syndrome: a preliminary study

OBJECTIVE

Predictors of treatment outcome of oral appliances (OAs) in patients with obstructive sleep apnoea syndrome (OSAS) are not known. There is a pressing need for simple, clinically useful tools to predict treatment outcome. This study aimed to identify predictors of successful OA therapy for OSAS, including evaluation of pharyngeal morphology, which can be measured during routine examination by an otorhinolaryngologist.

METHODOLOGY

This was a prospective study of 26 OSAS patients treated with OAs. A favourable outcome was obtained in 14 patients (responders) but not in 12 patients (nonresponders). The baseline patient characteristics and polysomnography and rhinopharyngeal findings were analysed.

RESULTS

Body mass index (BMI) was significantly lower in responders versus nonresponders (23.6 ± 2.8 vs. 27.9 ± 4.7 kg/m2; p < 0.05). Pharyngeal morphology, age, sex and nasal resistance did not differ between the groups. Multiple regression analysis showed that BMI was a significant predictor of improvement in the apnoea/hypopnoea index after OA treatment (p < 0.05).

CONCLUSION

Here we demonstrated that BMI is a favourable predictor of OA treatment outcome in OSAS patients. Among the OSAS patients, responders had wider retroglossal spaces than nonresponders.

Catalytic lectin (leczyme) from bullfrog (Rana catesbeiana) eggs

Catalytic lectins (leczymes) of frog eggs are sialic acid-binding lectins that have intrinsic RNase activity. They inhibit tumor cell proliferation in vitro and in vivo, although their cytotoxic mechanism remains unclear. RNase A has no tumoricidal activity. It is hypothesized that leczymes bind to cell surface sialoglycoconjugate receptors, enter the cell, and subsequently degrade RNA. In order to investigate the cytotoxic mechanism of cSBL, a leczyme from Rana catesbeiana eggs, we established cSBL-resistant clone RC-150 from mouse leukemia P388 cells. cSBL-treated P388 cells showed extensive RNA degradation over the course of 1 h, whereas cSBL-treated RC-150 cells showed no RNA degradation even over the course of 24 h. Treatment of P388 cells with cSBL led to decreased concentration of intracellular Ca2+, decreased protein kinase A activity, and increased protein kinase G activity. Incubation with cSBL decreased glutathione levels and enhanced glutathione-S-transferase (GST) activity in P388 cells, but had no effect on RC-150 cells. We conclude that cSBL-specific degradation of RNA occurs in cSBL-sensitive tumor cells, that cSBL leads to alteration of signal transduction and an intracellular protein kinase cascade reaction, and that internalized cSBL is detoxified by GST or thioltransferase. Our findings support a bifunctional model in which a leczyme is both an adhesive protein (binding to sialoglycoconjugates) and an enzyme (displaying RNnase activity).

Optimal positive airway pressure predicts oral appliance response to sleep apnoea

Patients with less severe obstructive sleep apnoea (OSA) are usually prescribed oral appliances and/or smaller optimal nasal continuous positive airway pressure (P(nCPAP)) in nCPAP therapy. We hypothesised that OSA patients with greater P(nCPAP) would not respond favourably to oral appliances. Oral appliances were inserted in nCPAP users after washing-out the nCPAP effect. Follow-up polysomnography was undertaken with the adjusted oral appliance in place. The predictability of P(nCPAP) was evaluated with receiver-operating characteristic (ROC) curves. The median baseline apnoea/hypopnoea index (AHI) was reduced with the oral appliance from 36 to 12 events.h(-1) in 35 patients. When responders were defined as patients showing a follow-up AHI of <5 events.h(-1) with >50% reduction in baseline AHI, the area under the ROC curve for P(nCPAP) was 0.76. The best cut-off value of P(nCPAP) turned out to be 10.5 cmH(2)O with a high negative predictive value (0.93) and a low negative likelihood ratio (0.18). OSA patients with a P(nCPAP) of >10.5 cmH( 2)O are unlikely to respond to oral appliance therapy. This prediction is clinically helpful to both OSA patients and medical personnel in discussing oral appliances as a temporary substitute and/or alternative for nCPAP.

A simple method for evaluation of tongue position

Neck extension because of contraction of cervical extensor muscles often brings about a lower-positioned tongue secondary to jaw opening in patients with congenital myopathy (CM). We hypothesized that neck extension in control subjects would reproduce the lower position of the tongue similar to that found in a CM patient. A simple method was formulated to evaluate the tongue position in terms of tongue pressure on the maxillary molar. A pair of pressure sensors was attached to the buccal and lingual surfaces of the upper molar for both the CM patient and four control subjects. Changes in the buccal and tongue pressures were recorded at the neck extension position for the CM patient and during both the natural head position and neck extension for the control subjects. There was a remarkable difference between buccal and tongue pressures in the neck extension position in the CM patient: tongue pressure was not detected at all, indicating there was no contact between tongue and upper molar. The buccal and tongue pressures were approximately equal in the natural head position in the control subjects. However, both buccal and tongue pressures were reduced during neck extension in the control subjects, with a greater decrease in the tongue pressure than the buccal pressure. These findings suggest that neck extension in a control subject reproduces the lower position of the tongue observed in CM patients. We propose that the pressure sensor enables evaluation of the tongue position, but further investigation is required.

Effects of maxillary distraction osteogenesis on the upper-airway size and nasal resistance in subjects with cleft lip and palate

OBJECTIVES

To investigate the short- and long-term effects of maxillary distraction osteogenesis (DOG) on the upper-airway size and nasal resistance in nine patients with cleft lip and palate (CLP).

STUDY DESIGN

Changes in the upper-airway size were measured by using lateral cephalometric radiographs taken immediately before and after DOG, and 1 year later. Nasal resistance was measured with a rhinomanometer. An analysis of variance was used to establish statistical significance. Spearman correlation coefficient was used to evaluate the relationship between changes in the cross-sectional area of the upper airway and nasal resistance in association with DOG.

RESULTS

Immediately after DOG, the anteroposterior dimension of the superior part of the upper airway was significantly increased (p < 0.01) and nasal resistance was significantly decreased (p < 0.05). Moreover, the cross-sectional area of the total upper airway was significantly increased (p < 0.01). There was a significant correlation between the increase in the upper-airway cross-sectional area and the reduction in nasal resistance (p < 0.05). The upper-airway size was significantly augmented (p < 0.05) and nasal resistance was significantly reduced (p < 0.05) at 1 year after DOG compared with immediately before DOG.

CONCLUSION

An increase in the upper-airway size and a reduction in nasal resistance occurred after maxillary DOG in patients with CLP, and these changes were stable after 1 year.

Effects of mandibular advancement on airway curvature and obstructive sleep apnoea severity

In a curved tube, the amount of airflow appears to be influenced by the amount of curvature. The purpose of this study was to investigate changes in obstructive sleep apnoea (OSA) severity and awake velopharyngeal curvature in response to an anteriorly titrated mandibular position in 20 male OSA patients. Baseline supine cephalometry was obtained before the initial insertion of a titratable oral appliance and follow-up supine cephalometry was undertaken after titration of the mandibular position with the appliance in place. The mean apnoea/hypopnea index (AHI) before treatment (31.6 +/- 13.0 events x h(-1)) was significantly reduced (9.8 +/- 7.4 events x h(-1)) after titration of the mandibular position in all 20 patients. There was a significant increase in the anteroposterior calibre and the radius of the curvature of the anterior wall of the velopharynx in 14 good responders who exhibited an AHI reduction to < or = 15. Similar observations were not found in six poor responders. To conclude, an anteriorly titrated mandibular position reduced obstructive sleep apnoea severity, enlarged the velopharynx and diminished the curvature of the anterior velopharyngeal wall in good responders. It is proposed that this change in the upper airway curvature associated with mandibular advancement may effect obstructive sleep apnoea severity through its effect on airflow dynamics.

Effects of a titratable oral appliance on supine airway size in awake non-apneic individuals

STUDY OBJECTIVES

To define morphological changes in the upper airway and its surrounding structures after the insertion of a titratable mandibular repositioner.

DESIGN

Ten non-apneic adult males participated in this study. A set of supine lateral cephalograms was taken for each subject at the end of expiration with a titratable oral appliance in place in four mandibular positions: most retruded (RP), maximum protrusion (MAX), 33% of MAX (MAX33), and 67% of MAX (MAX67). Changes in the anteroposterior width of the upper airway, positions of the hyoid bone and the third cervical vertebra were compared between the four mandibular positions. An ANOVA was used to test for statistical significance.

SETTING

N/A.

PATIENTS OR PARTICIPANTS

N/A.

INTERVENTIONS

N/A.

MEASUREMENTS AND RESULTS

The anteroposterior width of the velopharynx significantly increased when the mandible was advanced from RP to MAX67 and MAX. However, there were no significant changes in the anteroposterior width of the oropharynx. Significant forward displacement of the hyoid bone and third cervical vertebra together with the mandible was found in MAX67 and MAX compared to RP.

CONCLUSION

Especially in MAX67 and MAX, the titratable oral appliance significantly enlarges upper airway size in the velopharynx and results in a forward displacement of the hyoid bone and the third cervical vertebra.

Functional divergence of human genioglossus motor units with respiratory-related activity

The genioglossus muscle has at least two types of motor unit with respiratory-related activity. Inspiratory motor units show phasic activity during inspiration, whereas inspiratory/expiratory motor units show phasic inspiratory activity superimposed on tonic activity. The purpose of this study was to investigate the physiological roles of these different genioglossus motor units. The unitary activities of 12 inspiratory and 12 inspiratory/expiratory motor units were recorded using fine-wire electrodes during quiet nasal breathing in eight normal adult males. The mean interspike interval and the SD of successive spikes were calculated for inspiratory and inspiratory/expiratory motor units, respectively. Scattergrams of the mean interspike interval versus SD were constructed for the two groups of motor units. The effects of changes in head position on the firing activity and the patterns of distribution of the mean interspike interval versus its SD were significantly different for inspiratory and inspiratory/expiratory motor units. These results suggest that the inspiratory and inspiratory/expiratory motor units have different functional roles in respiration; inspiratory motor units may be phasically active to counteract intraluminal negative pressure during inspiration, whereas inspiratory/expiratory motor units may be tonically active to maintain tongue posture.

Inhibition of cell proliferation by Rana catesbeiana and Rana japonica lectins belonging to the ribonuclease superfamily

Two frog egg lectins [Rana catesbeiana lectin (SBL-C) and Rana japonica lectin] preferentially agglutinate a large variety of human and animal tumor cells but not blood cells, lymphocytes, or fibroblasts. These lectins belong to the superfamily of pyrimidine base-specific RNases. The two lectins bound to a heparin-Sepharose column and were eluted from the column by an increase of NaCl molarity. Both their tumor cell-agglutinating activity and RNase activity were inhibited by heparin, and also by polyamines, such as spermine. Both lectins inhibited P388 leukemia cell proliferation. The inhibitory activity of SBL-C was blocked by addition of heparin. SBL-C inhibited protein synthesis by P388 cells, but RNase A did not. No lectin-induced antiproliferative effect was observed after sialidase treatment of cells. The antiproliferative activity of SBL-C was also inhibited by ammonium chloride treatment. These results suggest that internalization of the lectins by lectin receptor (sialoglycoconjugate)-mediated endocytosis is followed by cell death due to inhibition of protein synthesis. Administration of SBL-C i.p. delayed time to death in mice receiving i.p. transplants of Sarcoma 180 and Mep II cells.

Identification of a rat liver protein-tyrosine phosphatase similar to human placental PTPase-1B using quantitatively phosphorylated protein substrates

Erythrocyte Band 3 protein (Band 3), brain microtubule associated protein 2 (MAP2), and tubulin were phosphorylated to high stoichiometries (1-6 mol Pi/mol protein) on tyrosine residues using a rat spleen protein-tyrosine kinase in the presence of polylysine. After total removal of polylysine, the quantitatively phosphorylated proteins as well as tyrosine-glutamate copolymer [Poly(Glu4, Tyr1)], which was also phosphorylated (1.5 mol/mol) by the kinase, were employed to assay rat liver protein-tyrosine phosphatases (PTPases). Of the four partially purified PTPases termed L1, L2, L3, and L4, PTPase L1 was previously purified to homogeneity and demonstrated to be a novel enzyme with sequence similarity to src-homology region 2 [Hiraga, A. et al. (1992) Eur. J. Biochem. 209, 195-206]. In the present work PTPase L2 was purified to near homogeneity by a procedure involving chromatography on DEAE-cellulose, Blue Sepharose CL-6B, hydroxylapatite, Phenyl Sepharose CL-4B, and TSKgel Heparin-5PW. PTPase L2 was purified 20,000-fold with a recovery of 0.9% from the extract and 0.005 mg was isolated from 300 g of liver. The highly purified PTPase L2 showed a major protein band of 36 kDa on SDS/polyacrylamide gel electrophoresis. PTPase L2 had a specific activity of about 6,000 nmol of P1 released min-1.mg-1 toward either Band 3 or poly(Glu4, Tyr1), the values being within the range of those obtained for PTPases purified thus far. PTPase L2 dephosphorylated Band-3 9-fold and 5-fold faster than tubulin and MAP2, respectively, under the assay conditions employed.(ABSTRACT TRUNCATED AT 250 WORDS)

Multiple forms of mammalian sialidase: altered expression in carcinogenesis

We have demonstrated that rat liver contains at least four types of sialidase differing in subcellular location, in catalytic property and immunologically. They are intralysosomal, cytosolic and membrane-associated sialidases I and II. Membrane sialidase I locates mainly in plasma membrane and sialidase II in lysosomal membrane. Immunological study reveals that the same types of sialidase exist in various tissues of rat and of other mammalian species. Based on these results, we examined the sialidases in rat hepatomas and in transformed cells of JB6 mouse epidermal cell. Hepatomas were found to possess four types of sialidase and the three of them altered quantitatively. Intralysosomal sialidase activity was higher but cytosolic and lysosomal membrane sialidase activities were lower in hepatomas than in control liver. When the sialidases of transformants of JB6 cells were compared with those of control cells, the activities of two lysosomal sialidases were decreased and contrarily plasma membrane sialidase was increased. We discussed a possible significance of the sialidase alterations in carcinogenesis.

Enhanced expression of type 2C protein phosphatase gene during myogenic differentiation of C3H10T1/2 cells

Type 2C phosphatase is one of the four major protein serine.threonine phosphatases (types 1, 2A, 2B and 2C). The mRNA level of the enzyme protein was 11.5 times higher in mouse skeletal muscle than in undifferentiated mesenchymal cells (C3H10T1/2). The mRNA level was enhanced 4.2-fold in accordance with the differentiation of C3H10T1/2 cells into myoblasts induced by transfection with a MyoD1 expression vector, and this high level was maintained in terminally differentiated myotubes. Furthermore, type 2C phosphatase activity in both myoblast and myotube fractions was slightly higher than in control C3H10T1/2 cells. These results indicate that the expression of type 2C phosphatase gene is enhanced during the course of myogenic differentiation.

Purification and characterization of a rat liver protein-tyrosine phosphatase with sequence similarity to src-homology region 2

Utilizing three proteins plus tyrosine-glutamate copolymer as substrates, all of which are subjected to (near) stoichiometrical phosphorylation exclusively on tyrosine residues, we partially purified four different protein-tyrosine phosphatases (PTPases) from rat liver cytosol which differed in substrate preference. Of the four PTPases, tentatively termed L1, L2, L3, and L4, PTPase L1 was purified to apparent homogeneity by a procedure involving chromatography on DEAE-cellulose at pH 7.0, Blue Sepharose, DEAE-cellulose at pH 7.6, hydroxyapatite, Phenyl Sepharose, Mono Q, and TSKgel Heparin. PTPase L1 was purified about 7000-fold from the extract and 0.27 mg was isolated from 1000 g liver corresponding to a yield of 13% from the Blue Sepharose step where it had become freed from any other PTPases detectable by our assay procedure. The purified PTPase L1 showed a major protein band of 67 kDa on SDS/PAGE. Catalytically, PTPase L1 had a specific activity of about 6500 nmol Pi released min-1mg-1 toward tyrosine-glutamate copolymer phosphorylated on tyrosine residues. PTPase L1 exhibited very low sensitivities to PTPase inhibitors such as zinc acetate, sodium vanadate, and acidic compounds as compared with those of most of the PTPases purified thus far. Amino acid sequence analysis of the purified PTPase L1 revealed a partial peptide sequence showing similarity to the catalytic domain core sequences conserved in the PTPase family. PTPase L1 was most similar to a PTPase termed PTP1C encoded by a human breast carcinoma cDNA but the identity was 55% over 117 residues spanning nearly half of the catalytic domain of PTP1C. The analysis also revealed another partial peptide sequence (113 residues) 70% identical with the sequence corresponding to 68% of two adjacent copies of the src homology region 2(SH-2 domain) identified in PTP1C. Besides those peptide sequences, PTPase L1 had regional sequences which were 70-90% identical with the residues lying between the two SH-2 domains or between the more C-terminal SH-2 domain and the catalytic domain of the carcinoma PTPase.

Alteration of the level of protein phosphatase 2C (IA) mRNA during the course of differentiation of skeletal muscle cells

On Northern hybridization using the cDNA of type 2C protein phosphatase as a probe, substantial amount of the mRNA of type 2C protein phosphatase was detected in various organs of rats, suggesting that the type 2C protein phosphatase gene is a housekeeping gene. A relatively high level of the mRNA, however, was found in skeletal muscle compared with other organs. Similar results were obtained with the organs of mice. In addition, the mRNA level in C3H10T1/2 cells (embryonal mesenchymal cells of mice) was much lower than in mature skeletal muscle of mice. The mRNA level of type 2C protein phosphatase was enhanced in accordance with the differentiation of the C3H10T1/2 cells into myoblasts induced by the transfection of MyoD cDNA. These results suggest that type 2C protein phosphatase is related to the mechanism of differentiation of skeletal muscle cells.

mRNA levels of catalytic subunits of protein phosphatases 1, 2A, and 2C in hepatocarcinogenesis

The mRNA levels of three phosphoseryl/phosphothreonyl protein phosphatases, PP1, PP2A and PP2C, in rat liver have been determined by Northern blot analysis in various stages of rat chemical hepatocarcinogenesis using a Solt-Farber model. Five weeks after administration of diethylnitrosamine, the mRNA levels of PP1 alpha, PP2A and PP2C were elevated 8, 29 and 11 times, respectively, as compared to those of the control livers. However, in primary hepatoma induced according to the Solt-Farber model, the mRNA levels of all three protein phosphatases were dramatically decreased to normal levels or even to much lower levels, whereas the mRNA level of glutathione S-transferase placental form, a tumor marker protein, was greatly elevated as compared with that of the control livers. In a poorly differentiated hepatoma AH13, a line of rat ascites hepatoma, the mRNA level of PP1 alpha was 5.6 times higher than that of the control livers, whereas the mRNA lever of PP2C was almost the same as that of the control livers and the level of PP2A mRNA was distinctly lower than that of the control livers. These data appear to suggest some involvement of protein phosphatases in hepatocarcinogenesis.

Phosphorylation of keratin intermediate filaments by protein kinase C, by calmodulin-dependent protein kinase and by cAMP-dependent protein kinase

Keratins, constituent proteins of intermediate filaments of epithelial cells, are phosphoproteins containing phosphoserine and phosphothreonine. We examined the in vitro phosphorylation of keratin filaments by cAMP-dependent protein kinase, protein kinase C and Ca2+/calmodulin-dependent protein kinase II. When rat liver keratin filaments reconstituted by type I keratin 18 (molecular mass 47 kDa; acidic type) and type II keratin 8 (molecular mass 55 kDa; basic type) in a 1:1 ratio were used as substrates, all the protein kinases phosphorylated both of the constituent proteins to a significant rate and extent, and disassembly of the keratin filament structure occurred. Kinetic analysis suggested that all these protein kinases preferentially phosphorylate keratin 8, compared to keratin 18. The amino acid residues of keratins 8 and 18 phosphorylated by cAMP-dependent protein kinase or protein kinase C were almost exclusively serine, while those phosphorylated by Ca2+/calmodulin-dependent protein kinase II were serine and threonine. Peptide mapping analysis indicated that these protein kinases phosphorylate keratins 8 and 18 in a different manner. These observations gave the way for in vivo studies of the role of phosphorylation in the reorganization of keratin filaments.

Molecular cloning and sequence analysis of cDNA for the catalytic subunit 1 alpha of rat kidney type 1 protein phosphatase, and detection of the gene expression at high levels in hepatoma cells and regenerating livers as compared to rat livers

A cDNA clone containing the full coding sequence of a type 1 protein phosphatase catalytic subunit 1 alpha has been isolated from a rat kidney lambda gt 10 library. The protein sequence deduced from the cDNA contains 330 amino acid residues with a molecular mass of 38 kDa. The cDNA clone from rat kidney was 89% identical at the nucleotide level in the coding region to type 1 protein phosphatase 1 alpha from rabbit skeletal muscle. However, the two protein sequences were completely identical. The type 1 alpha protein phosphatase from rat kidney shows 49% homology of amino acid sequence to the rat type 2A alpha protein phosphatase. Thus, the protein sequence of type 1 alpha protein phosphatase was completely conserved between rat and rabbit. The mRNA levels of type 1 protein phosphatase were determined in rat liver, AH13, a strain of rat hepatoma, and regenerating rat liver by Northern blot analysis using the cDNA fragment as a probe, under which conditions a single mRNA of 1.5 kb was detected. The mRNA levels of AH13 were remarkably increased when compared to those of normal ivers, whereas the mRNA levels of regenerating livers were slightly but significantly increased. These results demonstrate a marked increase in gene expression of type 1 protein phosphatase in hepatoma cells, suggesting an important role of the type 1 protein phosphatase in hepatocarcinogenesis.

Tumor-promoting phorbol ester induces alterations of sialidase and sialyltransferase activities of JB6 cells

Sialidase and sialyltransferase activities were studied in JB6 mouse epidermal cells before and after exposure to phorbol ester, 12-O-tetradecanoyl phorbol-13-acetate (TPA), which irreversibly induces anchorage-independent growth and tumorigenicity. JB6 cells exhibited sialidase activities toward 4-methylumbelliferyl-alpha-D-N-acetylneuraminic acid (4MU-NeuAc) and gangliosides at pH 4.5 in the particulate fraction but apparently not in the cytosol at pH 4.5 or 6.0. In JB6 cells exposed to TPA and in the anchorage-independent transformants, the sialidase activity toward 4MU-NeuAc was decreased and the activity toward gangliosides was increased compared with those in untreated JB6 cells. Immunological analysis with antisera against membrane-associated sialidases I and II revealed that plasma membrane-associated sialidase I was increased and lysosomal membrane-associated sialidase II was decreased under these conditions. TPA treatment also affected the sialyltransferase activities of JB6 cells: and elevation of the transfer activities toward asialo-orosomucoid and asialo-porcine submaxillary mucin but a reduction of GM3 and GD3 synthase activities were observed on exposure to TPA and in cells transformed by TPA to retain anchorage-independency. These results suggest that an increase in sialic acid bound to glycoproteins and a decrease in that bound to glycolipids may occur in JB6 cells exposed to TPA and in the anchorage-independent transformants.

Neoplastic alteration of a membrane-associated sialidase of rat liver

Rat liver particulate fraction contains two types of membrane-associated and gangliosides-hydrolyzing sialidase, which have been shown to be identical to two membrane-associated sialidases of rat brain (I and II) chromatographically, immunologically and in substrate specificity. Chromatography on AH-Sepharose 4B of the membrane sialidases of rat primary hepatoma induced by 3'-methyl-4-dimethylaminoazobenzene (MeDAB) further revealed that hepatocarcinogenesis induces a marked decrease in sialidase II but no decrease in sialidase I. Using antisera against sialidases I and II of rat brain, immunoprecipitation studies of the solubilized particulate fractions of rat liver and MeDAB-hepatoma gave results similar to those obtained chromatographically. Using the same immunological technique, sialidase II but not sialidase I was found to be decreased in AH109 A hepatoma and in regenerating and fetal liver.

[Three patients with typical sandblaster's silicosis proven by mineralogical analysis]

Three family members who had worked as sandblasters in their own sandblasting factory showed innumerable small nodularities in both lung fields of their chest radiograms. One of those showed conglomerate shadows in the upper lung fields. Those shadows seemed to be consistent with those of silicosis. One of the patients was examined by transbronchial lung biopsy (TBLB) and showed typical silicotic nodules. Mineralogical studies were done on the abrasive particles and the deposited particles in the lung tissue specimen obtained via TBLB and bronchoalveolar lavage fluid sample (BALF) using polarized microscopy, X-ray diffraction and analytical electron microscopy. The particles which had accumulated on the floor of the factory mineralogically consisted of mostly (over 90%) silica quartz containing small amount of chlorite, and the deposited particles in the lung tissue and those in the BALF showed similar composition.

Immunological discrimination of intralysosomal, cytosolic, and two membrane sialidases present in rat tissues

Cytosolic sialidase was purified from rat skeletal muscle, and the purified enzyme migrated as a single band of Mr 43,000 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. A polyclonal antibody raised against the enzyme inhibited and immunoprecipitated rat liver cytosolic sialidase as well as the muscle enzyme but failed to cross-react with the intralysosomal sialidase of rat liver and membrane sialidases I (synaptosomal) and II (lysosomal) of rat brain. The antibody against brain membrane sialidase I (anti-I) and that against sialidase II (anti-II), which could be useful to discriminate the two enzymes, did not cross-react with the intralysosomal and cytosolic sialidases of liver. Although more than 90% of liver plasma membrane sialidase was immunoprecipitated with anti-I, only 60% of liver lysosomal membrane sialidase was immunoprecipitated with anti-II, the remainder being immunoprecipitated with anti-I. In confirmation of these data, liver lysosomal membrane exhibited two peaks of ganglioside sialidase corresponding to the membrane sialidases I and II on Aminohexyl-Sepharose chromatography while only one peak of ganglioside sialidase corresponding to sialidase I was observed for liver plasma membrane. These results indicate that the four types of rat sialidase are proteins distinct from one another and that the three kinds of antisera described above are useful for discriminating these sialidases qualitatively and probably quantitatively.

Biochemical and immunological studies on two distinct ganglioside-hydrolyzing sialidases from the particulate fraction of rat brain

Ganglioside-hydrolyzing sialidase activity was solubilized from rat brain particulate fraction by using Triton X-100 plus sodium deoxycholate. When chromatographed on AH-Sepharose 4B, the solubilized activity was resolved into two peaks, which were designated sialidases I and II in order of elution. The two sialidases were purified by using sequential chromatographies on Octyl-Sepharose CL-4B, Phenyl-Sepharose CL-4B, and Sephadex G-200. Sialidase II was purified further by Mono Q-FPLC. Overall purification was 450- and 2,150-fold, for sialidases I and II, respectively. Purified sialidases I and II were maximally active at near pH 5.0 and exhibited M = 70,000 by gel filtration. Sialidase I hydrolyzed gangliosides but scarcely other substrates including 4-methylumbelliferyl-NeuAc (4MU-NeuAc). Sialidase II hydrolyzed oligosaccharides, glycoproteins, and 4MU-NeuAc although gangliosides appeared to be preferential substrates. Sialidase II cleaved GM2 much faster than sialidase I. An antibody raised in rabbits against sialidase I reacted with only sialidase I and an antibody against sialidase II reacted with only sialidase II. A subcellular distribution study suggested sialidase I in the synaptosomal membrane and sialidase II in the synaptosomal and lysosomal membranes, and this was verified by using the above antibodies.

Characterization of protein phosphatases associated with the particulate fraction from rat liver

Protein phosphatases associated with the particulate fraction from rat liver were studied by chromatographing the fraction on a DEAE-cellulose column and assaying the eluate with phosphorylase alpha and glycogen synthase D as substrates. Phosphorylase phosphatase activity emerged as two peaks, termed P-1 and P-2 in order of elution, both of which were inhibited by Mn2+ and Mg2+. P-1 and P-2 were Mr = 50,000 and 32,000 proteins, respectively, and when treated with trypsin, P-1 converted to a form indistinguishable from P-2, to which protein phosphatase inhibitor-2 was a potent inhibitor. Thus P-2 appears to be the catalytic subunit of type-1 protein phosphatase even though it has been degrated proteolytically as evidenced by its relatively low Mr. The elution profile of glycogen synthase phosphatase activity was entirely different. The activity obtained with 5 mM Mn2+ resolved into three peaks, the second-migrating M-2 being the largest. M-2 is an Mr = 70,000 protein; but an attempt to purify it has been unsuccessful giving a product of Mr = 40,000 and closely similar to the type-1 catalytic subunit in properties including inhibition by inhibitor-2. These results suggest that phosphatases P-1 and M-2 have a common catalytic subunit (type-1), which is bound to different "regulatory" subunits. M-2 distributes in glycogen particles and microsomes evenly while P-1 is almost exclusively in microsomes.

Phosphorylation sites linked to glial filament disassembly in vitro locate in a non-alpha-helical head domain

Glial fibrillary acidic protein (GFAP), the intermediate filament component of astroglial cells, can serve as an excellent substrate for both cAMP-dependent protein kinase and protein kinase C, in vitro. GFAP phosphorylated by each protein kinase does not polymerize, and the filaments that do polymerize tend to depolymerize after phosphorylation. Dephosphorylation of phospho-GFAP by phosphatase led to a recovery of the polymerization competence of GFAP. Most of the phosphorylation sites for cAMP-dependent protein kinase and protein kinase C on GFAP are the same, Ser-8, Ser-13, and Ser-34. cAMP-dependent protein kinase has one additional phosphorylation site, Thr-7. All the sites are located within the amino-terminal non-alpha-helical head domain of GFAP. These observations pave the way for in vivo studies on organization of glial filaments.

Characterization of the major sialidases of various types of rat blood cells: their comparison with rat liver sialidases

The substrate specificity and subcellular location of the major sialidases of three types of rat blood cells were characterized and compared with those of the known three types of rat liver sialidase, which have been designated intralysosomal, cytosolic, and plasma membrane-associated sialidases. Platelets and leucocytes contain mainly an acid sialidase, which is highly active towards oligosaccharides and 4MU-NeuAc, and erythrocytes possess a high level of a sialidase acting on gangliosides. A Percoll gradient centrifugation study showed that the former is located in lysosomes and the latter in plasma membrane. When the sialidase was solubilized and partially purified from erythrocyte ghosts, the enzyme was found to hydrolyze actively gangliosides but only poorly other substrates such as 4MU-NeuAc, oligosaccharides, and glycoproteins. The sialidase partially purified from rat liver membrane fraction exhibited the same substrate specificity. It is concluded that the major sialidase of platelets and leucocytes corresponds to hepatic intralysosomal sialidase while erythrocytes contain almost exclusively a ganglioside sialidase which corresponds to hepatic plasma membrane sialidase.

Effects of OK-432 activation on the sialidase activities of rat peritoneal macrophages

Intraperitoneal treatment of rat peritoneal macrophages with OK-432 results in more than 9-fold increase in the activity of ganglioside sialidase, which seems to coincide with the appearance of a cell surface antigen, asialo-GM1. The results of subsequent studies suggest that the ganglioside sialidase is located in the plasma membrane, where the enzyme may be responsible for the formation of asialo-GM1 from GM1. In the macrophages activated with OK-432, sialidase activity toward 4-methylumbelliferyl-N-acetylneuraminic acid (4MU-NeuAc) is also increased. It appears that the 4MU-NeuAc sialidase is intralysosomal and is increased together with other acid hydrolases present in the lysosomes.

Particulate-associated protein phosphatases of rat hepatomas as compared with the enzymes of rat liver

In the course of investigating the neoplastic alterations of protein phosphatases, the particulate fractions of rat liver and AH-13, a strain of rat ascites hepatoma, were chromatographed on DEAE-cellulose and assayed for protein phosphatase using glycogen synthase D and phosphorylase a as substrates. The synthase phosphatase activity of rapidly growing AH-13 was due almost entirely to a divalent cation-inhibited protein phosphatase, tentatively designated phosphatase N, the level of which was elevated remarkably in the hepatoma as compared with liver. Other hepatomas including primary hepatoma induced with 3'-methyl-4-dimethylaminoazobenzene also exhibited high levels of this phosphatase. Phosphatase N exhibited Mr = 49,000 (gel filtration) and has been partially purified with little alteration in properties. Partially purified phosphatase N was inhibited by divalent cations, rabbit skeletal muscle polypeptide inhibitor-2 and heparin, and released the catalytic subunit of type-1 protein phosphatase upon tryptic digestion. It is therefore apparent that phosphatase N is a type-1 protein phosphatase. There is some evidence to suggest that the high levels of phosphatase N in neoplastic cells are due primarily to enhanced synthesis of its non-catalytic (regulatory) subunit.

Expression of rat protein phosphatase 2C (IA) in Escherichia coli

A cDNA containing the entire coding sequence of rat type 2C (IA) protein phosphatase was expressed in Escherichia coli. An extract of bacterial cells harboring the recombinant plasmid contained a major (Mr = 41,000 - 43,000) and a minor (Mr = 30,000) protein band; both of these reacted with an anti-type 2C protein phosphatase serum. The size of the major protein band agrees well with that of the 2C phosphatase conceptualized from the cognate cDNA. A Mg2+-dependent protein phosphatase activity was detected in extracts containing the recombinant protein, but not in host cell extracts. Based on these results, it is concluded that the isolated cDNA clone encodes a functional type 2C protein phosphatase.

Molecular cloning of rat type 2C (IA) protein phosphatase mRNA

A full-length cDNA encoding rat type 2C (IA) protein phosphatase was isolated from a kidney cDNA library. The cDNA was identified by screening the library with oligonucleotides based on a partial amino acid sequence determined from purified rat liver phosphatase. This clone is 2.35 kilobase pairs long and has a single extended translation reading frame that predicts a 382-amino acid protein of 42,416 daltons. The deduced amino acid sequence contains segments corresponding to three peptides from rat liver type 2C protein phosphatase and two peptides from rabbit skeletal muscle type 2C phosphatase. Rat kidney type 2C protein phosphatase is distantly related to yeast adenylate cyclase but is not related to the catalytic subunits of two other protein phosphatases (types 1 and 2A).

Molecular cloning of cDNA for the catalytic subunit of rat liver type 2A protein phosphatase, and detection of high levels of expression of the gene in normal and cancer cells

A cloned cDNA encoding a catalytic subunit of type 2A protein phosphatase from a rat liver cDNA library was obtained by use of a synthetic oligonucleotide corresponding to the tryptic peptide sequence of the purified enzyme. There was only a single amino acid difference between the deduced amino acid sequence of the clone obtained and those of the catalytic subunits, 2A alpha, of the rabbit skeletal muscle, porcine kidney and human liver enzymes, suggesting that this clone was a rat 2A alpha cDNA. On Northern blot analysis using a cDNA fragment as a probe, three mRNA species were detected in rat liver: a major mRNA of 2.0 kb and a minor one of 2.7 kb under high stringency conditions, and also a 1.1 kb mRNA under low stringency conditions. The 2A alpha gene was found to be highly expressed in various tissues of rat, especially the brain. High levels of expression of the gene were also detected in mouse NIH3T3 cells and their transformants, and in human cancer cell lines as well as a human immortalized cell line.

Tyrosine protein kinase in preneoplastic and neoplastic rat liver

When rats are subjected to chemical hepatocarcinogenesis according to the protocol of D. Solt and E. Farber ((1976) Nature (London) 263, 701-703), the liver exhibits elevated levels of tyrosine protein kinase activity as early as 3 weeks after the injection of diethylnitrosoamine. A more striking elevation in tyrosine protein kinase activity is noted in rat hepatomas induced by administration of chemical carcinogens, in particular that of 3'-methyl-4-dimethylaminoazobenzene (3'-Me-DAB). Tyrosine protein kinase solubilized from the particulate fraction of 3'-Me-DAB-induced hepatoma has a molecular weight identical to that of p60v-src, cross-reacts with p60v-src immunologically, phosphorylates the heavy chain of anti-p60v-src IgG, and probably belongs to a family of p60c-src. The tyrosine protein kinase from the particulate fraction of normal rat liver is indistinguishable from the hepatoma kinase in these properties; thus it apparently differs only in the level of activity. Whether the liver and hepatoma kinases differ merely quantitatively or whether they differ even qualitatively, however, remains to be elucidated.

Analysis of glucose-6-phosphate translocase and hexose-6-phosphate phosphohydrolase, the two obligatory components of microsomal glucose-6-phosphatase system, in rat liver

A membrane filter procedure developed by Igarashi et al. (1984) for the measurement of glucose 6-phosphate uptake by the microsomes has been demonstrated to be a good method for assaying glucose-6-phosphate translocase, an obligatory component of the microsomal glucose-6-phosphatase system. When glucose-6-phosphate translocase was assayed in developing and diabetic rat livers independently of hexose-6-phosphate phosphohydrolase, another obligatory component of the glucose-6-phosphatase system, the two activities were found to undergo alterations, whose profiles, however, were quite distinct from each other. The profile of the microsomal glucose-6-phosphatase activity resembles the profile of the phosphohydrolase activity rather than that of the translocase activity, suggesting that the phosphohydrolase may be rate-limiting at least under these conditions. AH-109A, a strain of transplantable rat ascites hepatoma, was found to lack both glucose-6-phosphate translocase and hexose-6-phosphate phosphohydrolase activities.

Comparative study of the levels of sialyltransferases responsible for the formation of sugar chains in glycoproteins and gangliosides in rat liver and hepatomas

Sialyltransferases responsible for the formation of sugar chains in glycoproteins were studied in rat hepatoma in comparison with rat liver. Hepatoma induced by feeding Wistar rats with 3'-methyl-4-dimethylaminoazobenzene (MeDAB) was more active than Wistar liver in sialylating asialo-orosomucoid, and this was due to an increased activity of Gal(beta 1----4)GlcNAc (alpha 2----6) sialyltransferase, the major sialyltransferase in these tissues. Gal(beta 1----3,4)GlcNAc (alpha 2----3) sialyltransferase and the sialyltransferase acting on asialo-bovine submaxillary mucin were, however, decreased in the hepatoma. A similar pattern of sialyltransferase alterations was observed in regenerating liver and other tumors such as AH-109A hepatoma and Sato lung cancer, both of which had been inoculated into Donryu rats. In contrast to these sialyltransferases, the activities of the sialyltransferases responsible for the formation of gangliosides were markedly different even between Wistar and Donryu livers. When compared with Wistar liver, MeDAB-induced hepatoma was higher in lactosylceramide- and lower in GM3-sialyltransferase activity, but these two activities were both lower in AH-109A compared with Donryu liver.

Membrane-associated sialidase of rat liver and its decrease in hepatomas

Using the particulate fraction of tissue homogenate, plasma membrane-associated sialidase was assayed at pH 4.5 with bovine brain mixed gangliosides as the substrate. The activity was lower in rat hepatoma induced by 3'-methyl-4-dimethylaminoazobenzene (MeDAB) and transplantable AH-109A rat hepatoma than in normal rat liver. The enzyme was almost quantitatively solubilized from liver particulate fraction by using 0.5% (w/v) sodium deoxycholate plus 0.2% (w/v) Triton X-100. When chromatographed on DEAE-cellulose, the solubilized activity emerged as a single peak. The enzyme thus obtained was maximally active at pH 4.5, and readily hydrolyzed mixed gangliosides but was less active toward 4-methylumbelliferyl-alpha-N-acetylneuraminic acid, 3'-sialyllactose and fetuin. The corresponding enzyme from MeDAB-induced hepatoma was indistinguishable from the liver enzyme in terms of ease of solubilization, pH-activity relationship, chromatographic behavior and substrate preference. It therefore appears that the plasma membrane-associated sialidase of hepatomas differs from that of liver only in the tissue level of activity.

[Liver phosphotyrosine protein phosphatase and tyrosine protein kinase in chemical hepatocarcinogenesis]

We found and characterized three forms of phosphotyrosine protein phosphatase, subsequently designated PTPP-1, -2 and -3, respectively, in rat liver. Chemical hepatocarcinogenesis according to Solt and Farber was accompanied by a slight increase in liver phosphotyrosine protein phosphatase activity and a remarkable increase in liver tyrosine protein kinase activity. A maximum 8-fold increase in tyrosine kinase activity was observed in hepatomas induced with 3'-methyl-4-dimethylaminoazobenzene (MeDAB). Tyrosine protein kinase that increased with the progress of chemical hepatocarcinogenesis was solubilized from the particulate fraction of MeDAB-induced hepatoma. The enzyme was shown to require Mg2+ for its activity, to immunoprecipitate with anti-pp-60src-IgG and to phosphorylate the IgG. Rat liver also contains another tyrosine protein kinase which requires Mn2+ and does not immunoprecipitate with anti-pp60src; the level of this enzyme appears to diminish with the progress of chemical hepatocarcinogenesis.

Activation of rat liver and rabbit skeletal muscle glycogen synthases by rat liver cytosolic protein phosphatases

To gain more insight into the nature of the substrate specificity of protein phosphatases, four forms of glycogen synthase D were used as substrates for previously characterized protein phosphatases, IA, IB, and II, from rat liver cytosol. The phosphatase activity was measured as the conversion of glycogen synthase D to synthase I. While glycogen synthase isolated from rat liver as the D-form was activated mainly by phosphatase IA, rabbit skeletal muscle glycogen synthase previously phosphorylated in vitro by cyclic AMP-dependent protein kinase or phosphorylase kinase was activated efficiently by phosphatases IA, IB, and II. Glycogen synthase isolated from rabbit skeletal muscle as the D-form, however, was a poor substrate for all three phosphatases. These results suggest that the phosphorylation state as well as the primary structure of synthase D markedly affects the rate of its activation by individual protein phosphatases. A protein phosphatase released from rat liver particulate glycogen, on the other hand, activated all forms of synthase D used here readily and at about the same rate.

ATP-Mn2+ stimulates the generation of a putative mediator of insulin action

Substantial evidence suggests that insulin receptor-associated protein kinase may play a pivotal role in the expression of the intracellular effects of insulin. This study was undertaken to determine whether insulin receptor kinase contributes to the generation of putative insulin mediators. The effect of ATP and divalent cation addition on the production of insulin mediators from liver plasma membranes was investigated. ATP (1 mM) added to liver plasma membranes in the absence of divalent cations enhanced insulin-stimulated release/generation of mediator slightly (approximately 3-fold). ATP in the presence of Mn2+ further increased release/generation of mediator markedly (approximately 100-fold). In contrast, ATP in the presence of Mg2+ had no stimulatory effect. Mn2+ and Mg2+ alone were ineffective. Addition of EDTA completely diminished the stimulatory effects of insulin, ATP, and Mn2+. The stimulation was ATP-specific since other nucleotides and nonhydrolyzable analogues of ATP had no or very weak activity. ATP-Mn2+ stimulated insulin-dependent mediator release/generation in a dose-dependent manner. These results suggest that insulin mediator release/generation is markedly stimulated by an ATP-Mn2+-dependent phosphorylation reaction, similar to insulin-stimulated receptor tyrosine kinase phosphorylation.

Cytosolic sialidases of rat tissues with special reference to skeletal muscle enzyme

Using 4-methylumbelliferyl-N-acetylneuraminic acid as substrate, the cytosolic fractions of various rat tissues were assayed for sialidase activity. The activity was about 16 times greater in skeletal muscle than in liver, and the enzymes from the two sources were identical in chromatographic behavior, pH optimum, and substrate specificity. Apparently the same enzyme was found to be distributed in brain, heart, stomach, intestine, and testis, but not in lung and spleen. Sialidase was also present in kidney cytosol at a high level, but the enzyme resembled liver lysosomal sialidase rather than liver cytosolic sialidase in substrate specificity.

Identification and characterization of Mg2+-dependent phosphotyrosyl protein phosphatase from rat liver cytosol

Although highly purified preparations of Mg2+-dependent phosphoseryl protein phosphatase (also designated phosphatase IA or phosphatase 2C) dephosphorylated phosphotyrosyl histone, the activity has been resolved from phosphatase IA by polyacrylamide gel electrophoresis at pH 9.5. This novel phosphotyrosyl-specific protein phosphatase absolutely requires Mg2+ or Mn2+ for activity, is inhibited by Zn2+, vanadate and fluoride, and has an optimal pH of 9.0 and Mr = 50,000. Certain properties of this phosphatase so closely resemble those of phosphatase IA that the two enzymes tend to be copurified through various separation procedures.

Evidence for sialidase hydrolyzing gangliosides GM2 and GM1 in rat liver plasma membrane

Rat liver plasma membrane removed sialic acid from mixed bovine brain gangliosides more efficiently than from sialyllactose and orosomucoid with an optimal pH of 4.5. When individual gangliosides, each labeled with [14C]sialic acid or [3H]sphingosine, were tested, not only GD1a and GM3 but also GM2 and GM1, both of which had been considered to resist mammalian sialidases, were desialylated. The products of GM2 and GM1 hydrolysis were identified as asialo-GM2 and asialo-GM1, respectively, by thin-layer chromatography.