Development and external validation of a logistic and a penalized logistic model using machine-learning techniques to predict suicide attempts: A multicenter prospective cohort study in Korea.

Despite previous efforts to build statistical models for predicting the risk of suicidal behavior using machine-learning analysis, a high-accuracy model can lead to overfitting. Furthermore, internal validation cannot completely address this problem. In this study, we created models for predicting the occurrence of suicide attempts among Koreans at high risk of suicide, and we verified these models in an independent cohort. We performed logistic and penalized regression for suicide attempts within 6 months among suicidal ideators and attempters in The Korean Cohort for the Model Predicting a Suicide and Suicide-related Behavior (K-COMPASS). We then validated the models in a test cohort. Our findings indicated that several factors significantly predicted suicide attempts in the models, including young age, suicidal ideation, previous suicidal attempts, anxiety, alcohol abuse, stress, and impulsivity. The area under the curve and positive predictive values were 0.941 and 0.484 after variable selection and 0.751 and 0.084 in the test cohort. The corresponding values for the penalized regression model were 0.943 and 0.524 in the original training cohort and 0.794 and 0.115 in the test cohort. The prediction model constructed through a prospective cohort study of the suicide high-risk group showed satisfactory accuracy even in the test cohort. The accuracy with penalized regression was greater than that with the "classical" logistic model.

Evaluation of a Nondepleted Plasma Multiprotein-Based Model for Discriminating Psychiatric Disorders Using Multiple Reaction Monitoring-Mass Spectrometry: Proof-of-Concept Study.

Psychiatric evaluation relies on subjective symptoms and behavioral observation, which sometimes leads to misdiagnosis. Despite previous efforts to utilize plasma proteins as objective markers, the depletion method is time-consuming. Therefore, this study aimed to enhance previous quantification methods and construct objective discriminative models for major psychiatric disorders using nondepleted plasma. Multiple reaction monitoring-mass spectrometry (MRM-MS) assays for quantifying 453 peptides in nondepleted plasma from 132 individuals [35 major depressive disorder (MDD), 47 bipolar disorder (BD), 23 schizophrenia (SCZ) patients, and 27 healthy controls (HC)] were developed. Pairwise discriminative models for MDD, BD, and SCZ, and a discriminative model between patients and HC were constructed by machine learning approaches. In addition, the proteins from nondepleted plasma-based discriminative models were compared with previously developed depleted plasma-based discriminative models. Discriminative models for MDD versus BD, BD versus SCZ, MDD versus SCZ, and patients versus HC were constructed with 11 to 13 proteins and showed reasonable performances (AUROC = 0.890-0.955). Most of the shared proteins between nondepleted and depleted plasma models had consistent directions of expression levels and were associated with neural signaling, inflammatory, and lipid metabolism pathways. These results suggest that multiprotein markers from nondepleted plasma have a potential role in psychiatric evaluation.

Deep Learning of Speech Data for Early Detection of Alzheimer's Disease in the Elderly.

BACKGROUND: Alzheimer's disease (AD) is the most common form of dementia, which makes the lives of patients and their families difficult for various reasons. Therefore, early detection of AD is crucial to alleviating the symptoms through medication and treatment. OBJECTIVE: Given that AD strongly induces language disorders, this study aims to detect AD rapidly by analyzing the language characteristics. MATERIALS AND METHODS: The mini-mental state examination for dementia screening (MMSE-DS), which is most commonly used in South Korean public health centers, is used to obtain negative answers based on the questionnaire. Among the acquired voices, significant questionnaires and answers are selected and converted into mel-frequency cepstral coefficient (MFCC)-based spectrogram images. After accumulating the significant answers, validated data augmentation was achieved using the Densenet121 model. Five deep learning models, Inception v3, VGG19, Xception, Resnet50, and Densenet121, were used to train and confirm the results. RESULTS: Considering the amount of data, the results of the five-fold cross-validation are more significant than those of the hold-out method. Densenet121 exhibits a sensitivity of 0.9550, a specificity of 0.8333, and an accuracy of 0.9000 in a five-fold cross-validation to separate AD patients from the control group. CONCLUSIONS: The potential for remote health care can be increased by simplifying the AD screening process. Furthermore, by facilitating remote health care, the proposed method can enhance the accessibility of AD screening and increase the rate of early AD detection.

Network analysis of plasma proteomes in affective disorders.

The conventional differentiation of affective disorders into major depressive disorder (MDD) and bipolar disorder (BD) has insufficient biological evidence. Utilizing multiple proteins quantified in plasma may provide critical insight into these limitations. In this study, the plasma proteomes of 299 patients with MDD or BD (aged 19-65 years old) were quantified using multiple reaction monitoring. Based on 420 protein expression levels, a weighted correlation network analysis was performed. Significant clinical traits with protein modules were determined using correlation analysis. Top hub proteins were determined using intermodular connectivity, and significant functional pathways were identified. Weighted correlation network analysis revealed six protein modules. The eigenprotein of a protein module with 68 proteins, including complement components as hub proteins, was associated with the total Childhood Trauma Questionnaire score (r = -0.15, p = 0.009). Another eigenprotein of a protein module of 100 proteins, including apolipoproteins as hub proteins, was associated with the overeating item of the Symptom Checklist-90-Revised (r = 0.16, p = 0.006). Functional analysis revealed immune responses and lipid metabolism as significant pathways for each module, respectively. No significant protein module was associated with the differentiation between MDD and BD. In conclusion, childhood trauma and overeating symptoms were significantly associated with plasma protein networks and should be considered important endophenotypes in affective disorders.

Acoustic Analysis of Speech for Screening for Suicide Risk: Machine Learning Classifiers for Between- and Within-Person Evaluation of Suicidality.

BACKGROUND: Assessing a patient's suicide risk is challenging for health professionals because it depends on voluntary disclosure by the patient and often has limited resources. The application of novel machine learning approaches to determine suicide risk has clinical utility. OBJECTIVE: This study aimed to investigate cross-sectional and longitudinal approaches to assess suicidality based on acoustic voice features of psychiatric patients using artificial intelligence. METHODS: We collected 348 voice recordings during clinical interviews of 104 patients diagnosed with mood disorders at baseline and 2, 4, 8, and 12 months after recruitment. Suicidality was assessed using the Beck Scale for Suicidal Ideation and suicidal behavior using the Columbia Suicide Severity Rating Scale. The acoustic features of the voice, including temporal, formal, and spectral features, were extracted from the recordings. A between-person classification model that examines the vocal characteristics of individuals cross sectionally to detect individuals at high risk for suicide and a within-person classification model that detects considerable worsening of suicidality based on changes in acoustic features within an individual were developed and compared. Internal validation was performed using 10-fold cross validation of audio data from baseline to 2-month and external validation was performed using data from 2 to 4 months. RESULTS: A combined set of 12 acoustic features and 3 demographic variables (age, sex, and past suicide attempts) were included in the single-layer artificial neural network for the between-person classification model. Furthermore, 13 acoustic features were included in the extreme gradient boosting machine learning algorithm for the within-person model. The between-person classifier was able to detect high suicidality with 69% accuracy (sensitivity 74%, specificity 62%, area under the receiver operating characteristic curve 0.62), whereas the within-person model was able to predict worsening suicidality over 2 months with 79% accuracy (sensitivity 68%, specificity 84%, area under receiver operating characteristic curve 0.67). The second model showed 62% accuracy in predicting increased suicidality in external sets. CONCLUSIONS: Within-person analysis using changes in acoustic features within an individual is a promising approach to detect increased suicidality. Automated analysis of voice can be used to support the real-time assessment of suicide risk in primary care or telemedicine.

Latent class analysis of psychotic-affective disorders with data-driven plasma proteomics.

Data-driven approaches to subtype transdiagnostic samples are important for understanding heterogeneity within disorders and overlap between disorders. Thus, this study was conducted to determine whether plasma proteomics-based clustering could subtype patients with transdiagnostic psychotic-affective disorder diagnoses. The study population included 504 patients with schizophrenia, bipolar disorder, and major depressive disorder and 160 healthy controls, aged 19 to 65 years. Multiple reaction monitoring was performed using plasma samples from each individual. Pathologic peptides were determined by linear regression between patients and healthy controls. Latent class analysis was conducted in patients after peptide values were stratified by sex and divided into tertile values. Significant demographic and clinical characteristics were determined for the latent clusters. The latent class analysis was repeated when healthy controls were included. Twelve peptides were significantly different between the patients and healthy controls after controlling for significant covariates. Latent class analysis based on these peptides after stratification by sex revealed two distinct classes of patients. The negative symptom factor of the Brief Psychiatric Rating Scale was significantly different between the classes (t = -2.070, p = 0.039). When healthy controls were included, two latent classes were identified, and the negative symptom factor of the Brief Psychiatric Rating Scale was still significant (t = -2.372, p = 0.018). In conclusion, negative symptoms should be considered a significant biological aspect for understanding the heterogeneity and overlap of psychotic-affective disorders.

Comparison between self-reported smoking habits and daily ad-libitum smoking topography in a group of Korean smokers.

Tobacco smoking is associated with a high global mortality rate since it is known to cause cancers and lung and heart diseases. To control and reduce annual mortality attributed to smoking, it is essential to design applicable smoke cessation programs based on realistic tobacco exposure risk assessment. In this regard, understanding the smoking habits of the smoker is crucial. Using self-report smoking habit surveys is a common approach in measuring basic variables of smoking habits. However, smoking topography measurement devices have recently become available for investigating smoking habit variables accurately. In this study, we conducted a self-report survey to investigate a group of Korean smokers' smoking habit variables such as the number of smoked cigarettes per day, puff counts, and total smoking time. The survey also included items from the Fagerström Test for Nicotine Dependence (FTND). The results were compared with the corresponding variables from machine-determined data to investigate their correlation and reliability. Results indicate that Korean smokers have a reliable understanding of the average number of cigarettes they smoke daily (ρ = 0.517, Cronbach's α = 0.754) and the time to first cigarette (TTFC) after waking up (ρ = -0.587, Cronbach's α = 0.623), as fundamental items of the FTND score. Nevertheless, these smokers significantly under-reported the puff number and total smoking time, which can cause significant underestimation in the tobacco exposure risk assessment. Consequently, we suggest the application of self-report surveys that are based on the amount of daily smoked cigarettes (e.g. FTND) for clinical or risk assessment purposes. Using smoking topography measurement devices is recommended overusing self-report surveys in measuring smoking habit variables such as puff count and smoking time more accurately.

Rational Design and Synthesis of HSF1-PROTACs for Anticancer Drug Development.

PROTACs employ the proteosome-mediated proteolysis via E3 ligase and recruit the natural protein degradation machinery to selectively degrade the cancerous proteins. Herein, we have designed and synthesized heterobifunctional small molecules that consist of different linkers tethering KRIBB11, a HSF1 inhibitor, with pomalidomide, a commonly used E3 ligase ligand for anticancer drug development.

Exposure and Risk Assessment of Second- and Third-Hand Tobacco Smoke Using Urinary Cotinine Levels in South Korea.

Exposure to environmental tobacco smoke (ETS) is the reason for approximately 1% of global mortality. ETS exposure can happen either as inhalation of direct cigarette smoke (second-hand smoke) or its associated residue particles (third-hand smoke), especially when living with a smoker in the same family. This study investigated the association between the urinary cotinine levels, as biomarkers of exposure to tobacco smoke, of smokers and those exposed to second-hand and third-hand smoke while living in the same family, through a Korean nationwide survey. Direct assessment of ETS exposure and its lifetime effect on human health is practically difficult. Therefore, this study evaluated the internal estimated daily intake (I-EDI) of nicotine and equivalent smoked cigarette per day (CPD). The carcinogenic and non-carcinogenic inhalation risks of ETS exposure were assessed by considering the calculated equivalent CPD and composition of cigarette smoke of high-selling cigarette brands in South Korea. The results show that there is a statistically significant positive correlation between the cotinine levels of smokers and those of the non-smokers living in the same family. The risk assessment results yielded that hazard index (HI) and total excess lifetime cancer risk (ECR) for both second-hand and third-hand smoke exposure can exceed 1 and 1 × 10-6, respectively, especially in women and children. In the composition of the cigarette smoke, 1,3-butadiene and acrolein substances had the highest contribution to HI and ECR. Consequently, the provision of appropriate plans for smoking cessation as a strategy for the prevention of ETS exposure to women and children is deemed necessary.

Study on the daily Ad Libitum smoking habits of active Korean smokers and their effect on urinary smoking exposure and impact biomarkers.

INTRODUCTION: Understanding interactions of smoking topography with biomarkers of exposure to tobacco is essential for accurate smoking risk assessments. METHODS: In this study, the smoking topography and the levels of tobacco smoke exposure urinary biomarkers of a sample of active Korean smokers were quantified and measured. The results were used to investigate the effect of daily activities and smoking time on the smoking topography. Moreover, correlations between the smoking topography parameters and biomarkers were assessed. RESULTS: No significant effect of either the daily activities or time on the smoking topography of the subjects were observed. Synchronic correlations of the cigarette consumption per day (CPD) and the average flow per puff with both urinary cotinine and trans-3'-hydroxycotinine were significant. For the urinary nicotine metabolites, the peak levels appeared when the CPD was over 19 cigarettes per day and the average puff velocity was between 35 and 45 ml/s. Nevertheless, when the average flow was over 60 ml/s, the levels of cotinine and trans-3'-hydroxycotinine significantly dropped. CONCLUSIONS: The findings of this study may be beneficial for further smoking risk assessments with contributions of both the smoking topography and biomarkers to provide current smokers with applicable cession programs.Clinical significanceSmoking habits and levels of urinary biomarkers of Korean smokers are investigated.People with a higher dependency on nicotine smoke cigarettes with slower puffs.Effects of daily activities or time on smoking topography were not significant.Correlations between smoking topography and urinary biomarkers were significant.Peak biomarker levels were observed under certain smoking topography conditions.

Associated Factors on Quality of Life in Patients with Parkinson's Disease.

The objective of this study is to investigate the clinical and demographic factors that influence the quality of life in patients with Parkinson's disease (PD). This is a cross-sectional observational study of 47 patients in 2 hospitals with PD. All participants were asked to complete a disease-specific quality of life (QoL) questionnaire (PDQ-39). We gave a structured questionnaire interview and did a complete neurological examination on the same day. Additionally, we measured depression and dependency with the Geriatric Depression Scale-Short Form (GDS-SF) and the Korean version of the Modified Barthel Index (K-MBI). The PDQ-39 had a significant relationship with each motor part of the Unified Parkinson's Disease Rating Scale, the Korean Mini-Mental State Examination (K-MMSE), the GDS-SF, and the K-MBI (p < 0.05). The factors that independently contributed to the PDQ-39 scores were K-MMSE, GDS-SF, and K-MBI (p < 0.05). Factors having the greatest influence on the PDQ-39 were K-MBI, K-MMSE, and GDS-SF in that order. In addition, the mobility item in the K-MBI was independently a significant relating factor in the PDQ-39 (p < 0.05). These results demonstrated that dependency, especially with the mobility issue, was the greatest influence on the QoL in patients with PD.

Comparison of Suicide Risk by Mental Illness: a Retrospective Review of 14-Year Electronic Medical Records.

BACKGROUND: Korea is one of the countries with the highest rate of suicide, while suicidality is known to be closely related to mental illnesses. The study aimed to evaluate the suicide rates in psychiatric patients, to compare it to that of the general population, and to investigate the differences among psychiatric diagnoses and comorbidities. METHODS: Medical records and mortality statistics of psychiatric patients at Seoul National University Hospital from 2003 to 2017 were reviewed. The standardized mortality ratio (SMR) for suicide was calculated to compare the psychiatric patients with the general population. The diagnosis-specific standardized mortality rate and hazard ratio (HR) were adjusted by age, sex, and psychiatric comorbidity (i.e., personality disorder and/or pain disorder). RESULTS: A total of 40,692 survivors or non-suicidal deaths and 597 suicidal death were included. The suicide rate among psychiatric patients was 5.13-fold higher than that of the general population. Psychotic disorder had the highest SMR (13.03; 95% confidence interval [CI], 11.23-15.03), followed by bipolar disorder (10.26; 95% CI, 7.97-13.00) and substance-related disorder (6.78; 95% CI, 4.14-10.47). In survival analysis, psychotic disorder had the highest HR (4.16; 95% CI, 2.86-6.05), which was further increased with younger age, male sex, and comorbidity of personality disorder. CONCLUSION: All psychiatric patients are at a higher risk of suicide compared to the general population, and the risk is highest for those diagnosed with psychotic disorder.

Comparison of Serum Uric Acid in Major Depressive Disorder and Bipolar Disorder: a Retrospective Chart Review Study.

BACKGROUND: Uric acid (UA) has been suggested as a possible biomarker of bipolar disorder (BD) in recent studies. We aimed to provide a clearer comparison of UA levels between BD and major depressive disorder (MDD). METHODS: We retrospectively reviewed the medical chart records of psychiatric inpatients aged 19-60 years, whose main discharge diagnoses were either MDD or BD, with an admission between January 1, 2015 and December 31, 2018 at Seoul National University Hospital. Data such as sex, age, body mass index (BMI), medication usage, and serum UA levels were extracted. Patients with medical conditions or on medications that could influence UA levels were excluded. Age, sex, BMI, and psychiatric drug usage were considered in the comparison of serum UA between MDD and BD patients. RESULTS: Our sample consisted of 142 MDD patients and 234 BD patients. The BD patients had significantly higher serum UA levels compared to the MDD patients, without accounting for other confounding variables (5.75 ± 1.56 mg/dL vs. 5.29 ± 1.59 mg/dL, P = 0.006). T-test comparisons between psychiatric medication users and non-users revealed that mood stabilizers and antipsychotics may be relevant confounding factors in our sample analysis. The likelihood of BD diagnosis was significantly correlated with higher UA levels (odds ratio, 1.410; 95% confidence interval, 1.150-1.728; P = 0.001) when accounting for sex, age, and BMI in the logistic regression analysis. Also, accounting for mood stabilizers or antipsychotics, the likelihood of BD diagnosis was still significantly correlated with higher UA levels. CONCLUSION: Our study confirms that BD patients are significantly more likely to show higher serum UA levels than MDD patients. The high UA levels in BD point to purinergic dysfunction as an underlying mechanism that distinguishes BD from MDD. Further research is recommended to determine whether UA is a trait or a state marker and whether UA correlates with the symptoms and severity of BD.

Hybrid inhibitors of DNA and HDACs remarkably enhance cytotoxicity in leukaemia cells.

Chlorambucil is a nitrogen mustard-based DNA alkylating drug, which is widely used as a front-line treatment of chronic lymphocytic leukaemia (CLL). Despite its widespread application and success for the initial treatment of leukaemia, a majority of patients eventually develop acquired resistance to chlorambucil. In this regard, we have designed and synthesised a novel hybrid molecule, chloram-HDi that simultaneously impairs DNA and HDAC enzymes. Chloram-HDi efficiently inhibits the proliferation of HL-60 and U937 leukaemia cells with GI50 values of 1.24 µM and 1.75 µM, whereas chlorambucil exhibits GI50 values of 21.1 µM and 37.7 µM against HL-60 and U937 leukaemia cells, respectively. The mechanism behind its remarkably enhanced cytotoxicity is that chloram-HDi not only causes a significant DNA damage of leukaemia cells but also downregulates DNA repair protein, Rad52, resulting in the escalation of its DNA-damaging effect. Furthermore, chloram-HDi inhibits HDAC enzymes to induce the acetylation of α-tubulin and histone H3.

Design, synthesis and biological evaluation of a series of CNS penetrant HDAC inhibitors structurally derived from amyloid-ß probes.

To develop novel CNS penetrant HDAC inhibitors, a new series of HDAC inhibitors having benzoheterocycle were designed, synthesized, and biologically evaluated. Among the synthesized compounds, benzothiazole derivative 9b exhibited a remarkable anti-proliferative activity (GI50 = 2.01 µM) against SH-SY5Y cancer cell line in a dose and time-dependent manner, better than the reference drug SAHA (GI50 = 2.90 µM). Moreover, compound 9b effectively promoted the accumulation of acetylated Histone H3 and α-tubulin through inhibition of HDAC1 and HDAC6 enzymes, respectively. HDAC enzyme assay also confirmed that compound 9b efficiently inhibited HDAC1 and HDAC6 isoforms with IC50 values of 84.9 nM and 95.9 nM. Furthermore, compound 9b inhibited colony formation capacity of SH-SY5Y cells, which is considered a hallmark of cell carcinogenesis and metastatic potential. The theoretical prediction, in vitro PAMPA-BBB assay, and in vivo brain pharmacokinetic studies confirmed that compound 9b had much higher BBB permeability than SAHA. In silico docking study demonstrated that compound 9b fitted in the substrate binding pocket of HDAC1 and HDAC6. Taken together, compound 9b provided a novel scaffold for developing CNS penetrant HDAC inhibitors and therapeutic potential for CNS-related diseases.

A novel class of anthraquinone-based HDAC6 inhibitors.

Histone deacetylase 6 (HDAC6) is an important target for the treatment of diverse diseases including cancer, neurodegenerative diseases, autoimmune disorders, inflammation, drug addiction, and viral infection. Therefore, the discovery of HDAC6-isoform selective inhibitors is of high importance for clinical applications. Here, we present an approach to discover HDAC6-isoform selective inhibitors. To our best knowledge, we for the first time perform a virtual screening campaign in the surface and channel region of HDAC6 enzyme, followed by rational installation of zinc binding group for the development of HDAC6-isoform selective inhibitors. Consequently, this approach establishes the proof of principle for the discovery of HDAC6-isoform selective inhibitors and successfully provides our lead compound 3. In particular, compound 3 inhibits HDAC6 enzyme with an IC50 value of 56 nM and displays an excellent HDAC6 selectivity over other HDAC isoforms in HDAC enzyme assay. Furthermore, the exposure of SH-SY5Y cells with compound 3 significantly promotes the acetylation of α-tubulin at the low concentration of 0.5 µM, but not the acetylation of Histone H3 up to 20 µM. Thus, our lead compound 3 represents a novel HDAC6-isoform selective inhibitor and warrants further studies for therapeutic evaluation.

Aspirin-inspired acetyl-donating HDACs inhibitors.

Aspirin is one of the oldest drugs for the treatment of inflammation, fever, and pain. It is reported to covalently modify COX-2 enzyme by acetylating a serine amino acid residue. By virtue of aspirin's acetylating potential, we for the first time developed novel acetyl-donating HDAC inhibitors. In this study, we report the design, synthesis, in silico docking study, and biological evaluation of acetyl-donating HDAC inhibitors. The exposure of MDA-MB-231 cells with compound 4c significantly promotes the acetylation of α-tubulin and histone H3, which are substrates of HDAC6 and HDAC1, respectively. In silico docking simulation also indicates that compound 4c tightly binds to the deep substrate-binding pocket of HDAC6 by coordinating the active zinc ion in a bidentate manner and forming hydrogen bond interactions with Ser531 and His573 amino acid residues. In particular, compound 4c (GI50 = 147 µM) affords the significant enhancement of anti-proliferative effect on MDA-MB-231 cells, compared with its parent compound 2c (GI50 > 1000 µM) and acetyl-donating group deficient compound 6 (GI50 = 554 µM). Overall, compound 4c presents a novel strategy for developing acetyl-donating HDAC inhibitors.

Inhibitory Activities of Cudrania tricuspidata Leaves on Pancreatic Lipase In Vitro and Lipolysis In Vivo.

To identify effective herb to treat obesity, we screened 115 herbal extracts for inhibition of porcine pancreatic lipase (triacylg-ycerol acylhydrolase, EC 3.1.1.3) activity in vitro. Of the extracts tested, Cudrania tricuspidata leaves exhibited the most pronounced inhibitory effect on lipase activity with an IC(50) value of 9.91 µg/mL. Antilipid absorption effects of C. tricuspidata leaves were examined in rats after oral administration of lipid emulsions containing 50 or 250 mg C. tricuspidata/kg body weight. Plasma triacylglycerol levels 2 h after the oral administration of emulsions containing C. tricuspidata were significantly reduced compared to the untreated group (P < 0.05). These results suggest that C. tricuspidata leaves may be useful for the treatment of obesity.