IL-17 and IL-21: Their Immunobiology and Therapeutic Potentials.

Studies over the last 2 decades have identified IL-17 and IL-21 as key cytokines in the modulation of a wide range of immune responses. IL-17 serves as a critical defender against bacterial and fungal pathogens, while maintaining symbiotic relationships with commensal microbiota. However, alterations in its levels can lead to chronic inflammation and autoimmunity. IL-21, on the other hand, bridges the adaptive and innate immune responses, and its imbalance is implicated in autoimmune diseases and cancer, highlighting its important role in both health and disease. Delving into the intricacies of these cytokines not only opens new avenues for understanding the immune system, but also promises innovative advances in the development of therapeutic strategies for numerous diseases. In this review, we will discuss an updated view of the immunobiology and therapeutic potential of IL-17 and IL-21.

STAT3 Decoy Oligodeoxynucleotides Suppress Liver Inflammation and Fibrosis in Liver Cancer Cells and a DDC-Induced Liver Injury Mouse Model.

Liver damage caused by various factors results in fibrosis and inflammation, leading to cirrhosis and cancer. Fibrosis results in the accumulation of extracellular matrix components. The role of STAT proteins in mediating liver inflammation and fibrosis has been well documented; however, approved therapies targeting STAT3 inhibition against liver disease are lacking. This study investigated the anti-fibrotic and anti-inflammatory effects of STAT3 decoy oligodeoxynucleotides (ODN) in hepatocytes and liver fibrosis mouse models. STAT3 decoy ODN were delivered into cells using liposomes and hydrodynamic tail vein injection into 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC)-fed mice in which liver injury was induced. STAT3 target gene expression changes were verified using qPCR and Western blotting. Liver tissue fibrosis and bile duct proliferation were assessed in animal experiments using staining techniques, and macrophage and inflammatory cytokine distribution was verified using immunohistochemistry. STAT3 decoy ODN reduced fibrosis and inflammatory factors in liver cancer cell lines and DDC-induced liver injury mouse model. These results suggest that STAT3 decoy ODN may effectively treat liver fibrosis and must be clinically investigated.

Boosting with variant-matched adenovirus-based vaccines promotes neutralizing antibody responses against SARS-CoV-2 Omicron sublineages in mice.

Global spread of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) Omicron subvariants, such as BA.4, BA.5 and XBB.1.5, has been leading the recent wave of coronavirus disease 2019 (COVID-19). Unique mutations in the spike proteins of these emerging Omicron subvariants caused immune evasion from the pre-existing protective immunity induced by vaccination or natural infection. Previously, we developed AdCLD-CoV19-1, a non-replicating recombinant adenoviral vector that encodes the receptor binding domain of the spike protein of the ancestral SARS-CoV-2 strain. Based on the same recombinant adenoviral vector platform, updated vaccines that cover unique mutations found in each Omicron subvariant, including BA.1, BA.2, BA.4.1 and BA.5, were constructed. Preclinical studies revealed that each updated vaccine as a booster shot following primary vaccination targeting the ancestral strain improved neutralizing antibody responses against the pseudovirus of its respective strain most effectively. Of note, boosting with a vaccine targeting the BA.1 or BA.2 Omicron subvariant was most effective in neutralization against the pseudovirus of the BA.2.75 strain, whereas BA.4.1/5-adapted booster shots were most effective in neutralization against the BQ.1, BQ1.1 and BF.7 strains. Therefore, it is imperative to develop a vaccination strategy that can cover the unique spike mutations of currently circulating Omicron subvariants in order to prevent the next wave of COVID-19.

The Clinical Significance of Myosteatosis in Survival Outcomes in Patients with Hepatocellular Carcinoma Treated with Sorafenib.

The role of body composition parameters in sorafenib-treated hepatocellular carcinoma (HCC) patients is still not fully elucidated. Here, we aimed to evaluate the impact of computed tomography (CT)-based body composition parameters on the survival of such patients. In this multicenter study, we analyzed the data of 245 sorafenib-treated HCC patients from January 2008 to December 2019. Sarcopenia, visceral obesity, and myosteatosis were defined by using cross-sectional CT images at the third lumbar vertebra level. The effects of these parameters on overall survival (OS) and progression-free survival (PFS) were evaluated. The median age was 67.0 years (interquartile range: 61.0-78.0 year), and 211 patients (86.1%) were male. The median OS and PFS were 7.9 months and 4.8 months, respectively. Vascular invasion (hazard ratio (HR), 1.727; 95% confidence interval (CI), 1.258-2.371; p = 0.001), extrahepatic metastasis (HR, 1.401; 95% CI, 1.028-1.908; p = 0.033), alpha-fetoprotein level > 200 ng/mL (HR, 1.559; 95% CI, 1.105-2.201; p = 0.012), and myosteatosis (HR, 1.814; 95% CI, 1.112-2.960; p = 0.017) were associated with OS. Patient mortality was significantly higher in the group with two or more risk factors than in the group with fewer risk factors. In conclusion, myosteatosis may be a novel prognostic CT-based radiological biomarker in sorafenib-treated HCC patients.

CD8 T-cell subsets: heterogeneity, functions, and therapeutic potential.

CD8 T cells play crucial roles in immune surveillance and defense against infections and cancer. After encountering antigenic stimulation, naïve CD8 T cells differentiate and acquire effector functions, enabling them to eliminate infected or malignant cells. Traditionally, cytotoxic T cells, characterized by their ability to produce effector cytokines and release cytotoxic granules to directly kill target cells, have been recognized as the constituents of the predominant effector T-cell subset. However, emerging evidence suggests distinct subsets of effector CD8 T cells that each exhibit unique effector functions and therapeutic potential. This review highlights recent advancements in our understanding of CD8 T-cell subsets and the contributions of these cells to various disease pathologies. Understanding the diverse roles and functions of effector CD8 T-cell subsets is crucial to discern the complex dynamics of immune responses in different disease settings. Furthermore, the development of immunotherapeutic approaches that specifically target and regulate the function of distinct CD8 T-cell subsets holds great promise for precision medicine.

A Personalized Cancer Vaccine that Induces Synergistic Innate and Adaptive Immune Responses.

To demonstrate potent efficacy, a cancer vaccine needs to activate both innate and adaptive immune cells. Personalized cancer vaccine strategies often require the identification of patient-specific neoantigens; however, the clonal and mutational heterogeneity of cancer cells presents inherent challenges. Here, extracellular nanovesicles derived from alpha-galactosylceramide-conjugated autologous acute myeloid leukemia (AML) cells (ECNV-αGC) are presented as a personalized therapeutic vaccine that activates both innate and adaptive immune responses, bypassing the need to identify patient-specific neoantigens. ECNV-αGC vaccination directly engages with and activates both invariant natural killer T (iNKT) cells and leukemia-specific CD8+ T cells in mice with AML, thereby promoting long-term anti-leukemic immune memory. ECNV-αGC sufficiently serves as an antigen-presenting platform that can directly activate antigen-specific CD8+ T cells even in the absence of dendritic cells, thereby demonstrating a multifaceted cellular mechanism of immune activation. Moreover, ECNV-αGC vaccination results in a significantly lower AML burden and higher percentage of leukemia-free survivors among cytarabine-treated hosts with AML. Human AML-derived ECNV-αGCs activate iNKT cells in both healthy individuals and patients with AML regardless of responsiveness to conventional therapies. Together, autologous AML-derived ECNV-αGCs may be a promising personalized therapeutic vaccine that efficiently establishes AML-specific long-term immunity without requiring the identification of neoantigens.

Triterpenoid saponins from Camellia sinensis roots with cytotoxic and immunomodulatory effects.

Camellia sinensis L. (Theaceae) leaves have been used as a beverage in both Eastern and Western cultures for a long time, while its root has not been intensively studied. In this study, seven undescribed triterpenoid saponins (1-7) and twelve known saponins (8-19) with different combinations of substituents, such as oxygenated isoprenyl substituents and sugar moieties, and lengths of sugar chains, were isolated from the C. sinensis roots. Their structures were unequivocally determined using one- and two-dimensional nuclear magnetic resonance data and acid hydrolysis analysis. Investigation of the biological activities of isolated compounds revealed that only those without functional acetyl groups exhibited cytotoxic activities against mouse and human cancer cells (B16F10) and human cervical cancer cell line (HeLa) at 50 µM. Compounds with an aldehyde group at C-23 of aglycone showed immunomodulatory activity against Th1 and Th17 cells at 10 µM. Ten compounds with biological activities from C. sinensis roots extracts, including three previously undescribed ones (3, 6, and 7), were identified.

The Clinical Courses and Prognosis of Cirrhotic Patients after First Acute Decompensation: Prospective Cohort Study.

BACKGROUND: The European Foundation for the Study of Chronic Liver Failure (EF-CLIF) consortium suggested that the clinical courses after acute decompensation (AD) stratify the long-term prognosis: stable decompensated cirrhosis (SDC), unstable decompensated cirrhosis (UDC), pre acute-on-chronic liver failure (pre ACLF), and ACLF. However, previous studies included patients with a history of previous AD and had limitations associated with identifying the clinical factors related to prognosis after the first AD. METHOD: The prospective Korean Acute-on-Chronic Liver Failure (KACLiF) cohort included cirrhotic patients who were hospitalised with first AD between July 2015 and August 2018. We analysed the factors associated with readmission after the first AD and compared the characteristics and prognosis among each subgroup to evaluate the risk factors for the occurrence of pre ACLF after AD. RESULT: A total of 746 cirrhotic patients who were hospitalised with first AD were enrolled. The subgroups consisted of SDC (n = 565), UDC (n = 29), pre ACLF (n = 28), and ACLF (n = 124). Of note, pre ACLF showed a poorer prognosis than ACLF. The risk factors associated with readmission within 3 months of first AD were non-variceal gastrointestinal (GI) bleeding, hepatic encephalopathy (HE), and high MELD score. Viral aetiology was associated with the occurrence of pre ACLF compared with alcohol aetiology regardless of baseline liver function status. CONCLUSION: Cirrhotic patients with first AD who present as non-variceal GI bleeding and HE can easily relapse. Interestingly, the occurrence of AD with organ failure within 3 months of first AD (pre ACLF) has worse prognosis compared with the occurrence of organ failure at first AD (ACLF). In particular, cirrhotic patients with viral hepatitis with/without alcohol consumption showed poor prognosis compared to other aetiologies. Therefore, patients with ACLF after AD within 3 months should be treated more carefully and definitive treatment through LT should be considered.

The FoxO4/DKK3 axis represses IFN-γ expression by Th1 cells and limits antimicrobial immunity.

Forkhead box O transcriptional factors, especially FoxO1 and FoxO3a, play critical roles in physiologic and pathologic immune responses. However, the function of FoxO4, another main member of the FoxO family, in lymphoid cells is still poorly understood. Here, we showed that loss of FoxO4 in T cells augmented IFN-γ production of Th1 cells in vitro. Correspondingly, conditional deletion of FoxO4 in CD4+ T cells enhanced T cell-specific responses to Listeria monocytogenes infection in vivo. Genome-wide occupancy and transcriptomic analyses identified Dkk3 (encoding the Dickkopf-3 protein) as a direct transcriptional target of FoxO4. Consistent with the FoxO4-DKK3 relationship, recombinant DKK3 protein restored normal levels of IFN-γ production in FoxO4-deficient Th1 cells through the downregulation of lymphoid enhancer-binding factor 1 (Lef1) expression. Together, our data suggest a potential FoxO4/DKK3 axis in Th1 cell differentiation, providing what we believe to be an important insight and supplement for FoxO family proteins in T lymphocyte biology and revealing a promising target for the treatment of immune-related diseases.

Platelet-to-White Blood Cell Ratio Is Associated with Adverse Outcomes in Cirrhotic Patients with Acute Deterioration.

Background: The platelet-to-white blood cell ratio (PWR) is a hematologic marker of the systemic inflammatory response. Recently, the PWR was revealed to have a role as an independent prognostic factor for mortality in patients with hepatitis B virus (HBV)-related acute-on-chronic failure (ACLF) and HBV-related liver cirrhosis (LC) with acute decompensation (AD). However, the prognostic role of the PWR still needs to be investigated in LC patients with AD. In this study, we analyzed whether the PWR could stratify the risk of adverse outcomes (death or liver transplantation (LT)) in these patients. Methods: A prospective cohort of 1670 patients with AD of liver cirrhosis ((age: 55.2 ± 7.8, male = 1226 (73.4%)) was enrolled and evaluated for 28-day and overall adverse outcomes. Results: During a median follow-up of 8.0 months (range, 1.9−15.5 months), 424 (25.4%) patients had adverse outcomes (death = 377, LT = 47). The most common etiology of LC was alcohol use (69.7%). The adverse outcome rate was higher for patients with a PWR ≤ 12.1 than for those with a PWR > 12.1. A lower PWR level was a prognostic factor for 28-day adverse outcomes (PWR: hazard ratio 1.707, p = 0.034) when adjusted for the etiology of cirrhosis, infection, ACLF, and the MELD score. In the subgroup analysis, the PWR level stratified the risk of 28-day adverse outcomes regardless of the presence of ACLF or the main form of AD but not for those with bacterial infection. Conclusions: A lower PWR level was associated with 28-day adverse outcomes, indicating that the PWR level can be a useful and simple tool for stratifying the risk of 28-day adverse outcomes in LC patients with AD.

Deep Learning-Based Four-Region Lung Segmentation in Chest Radiography for COVID-19 Diagnosis.

Imaging plays an important role in assessing the severity of COVID-19 pneumonia. Recent COVID-19 research indicates that the disease progress propagates from the bottom of the lungs to the top. However, chest radiography (CXR) cannot directly provide a quantitative metric of radiographic opacities, and existing AI-assisted CXR analysis methods do not quantify the regional severity. In this paper, to assist the regional analysis, we developed a fully automated framework using deep learning-based four-region segmentation and detection models to assist the quantification of COVID-19 pneumonia. Specifically, a segmentation model is first applied to separate left and right lungs, and then a detection network of the carina and left hilum is used to separate upper and lower lungs. To improve the segmentation performance, an ensemble strategy with five models is exploited. We evaluated the clinical relevance of the proposed method compared with the radiographic assessment of the quality of lung edema (RALE) annotated by physicians. Mean intensities of segmented four regions indicate a positive correlation to the regional extent and density scores of pulmonary opacities based on the RALE. Therefore, the proposed method can accurately assist the quantification of regional pulmonary opacities of COVID-19 pneumonia patients.

Nanovesicle-Mediated Targeted Delivery of Immune Checkpoint Blockades to Potentiate Therapeutic Efficacy and Prevent Side Effects.

Despite the clinically proven efficacies of immune checkpoint blockades, including anti-cytotoxic T lymphocyte-associated protein 4 antibody (αCTLA-4), the low response rate and immune-related adverse events (irAEs) in cancer patients represent major drawbacks of the therapy. These drawbacks of αCTLA-4 therapy are mainly due to the suboptimal activation of tumor-specific cytotoxic T lymphocytes (CTLs) and the systemic nonspecific activation of T cells. To overcome such drawbacks, αCTLA-4 is delivered by dendritic cell-derived nanovesicles presenting tumor antigens (DCNV-TAs) that exclusively interact with tumor-specific T cells, leading to selective activation of tumor-specific CTLs. Compared to conventional αCTLA-4 therapy, treatment with αCTLA-4-conjugated DCNV-TAs significantly inhibits tumor growth and reduces irAEs in syngeneic tumor-bearing mice. This study demonstrates that the spatiotemporal presentation of both αCTLA-4 and tumor antigens enables selective activation of tumor-specific T cells and potentiates the antitumor efficacy of αCTLA-4 without inducing systemic irAEs.

Inhibition of topoisomerase I shapes antitumor immunity through the induction of monocyte-derived dendritic cells.

Understanding the rationale of combining immunotherapy and other anticancer treatment modalities is of great interest because of interpatient variability in single-agent immunotherapy. Here, we demonstrated that topoisomerase I inhibitors, a class of chemotherapeutic drugs, can alter the tumor immune landscape, corroborating their antitumor effects combined with immunotherapy. We observed that topotecan-conditioned TC-1 tumors were occupied by a vast number of monocytic cells that highly express CD11c, CD64, and costimulatory molecules responsible for the favorable changes in the tumor microenvironment. Ly6C+MHC-II+CD11chiCD64hi cells, referred to as topotecan-induced monocyte-derived dendritic cells (moDCs), proliferate and activate antigen-specific CD8+ T cells to levels equivalent to those of conventional DCs. Phenotypic changes in Ly6C+ cells towards moDCs were similarly induced by exposure to topotecan in vitro, which was more profoundly facilitated in the presence of tumor cells. Notably, anti-M-CSFR reversed the acquisition of DC-like properties of topotecan-induced moDCs, leading to the abolition of the antitumor effect of topotecan combined with a cancer vaccine. In short, topoisomerase I inhibitors generate monocyte-derived antigen-presenting cells in tumors, which could be mediated by M-CSF-M-CSFR signaling.

Changing Trends in Liver Cirrhosis Etiology and Severity in Korea: the Increasing Impact of Alcohol.

BACKGROUND: Chronic hepatitis B is the most common cause of liver cirrhosis in South Korea. However, alcoholic liver disease has shown an increasing trend. Although the clinical implications surrounding liver cirrhosis have been changing over the years, few studies have recently examined cirrhosis epidemiology. Therefore, we aimed to investigate changes in liver cirrhosis etiology and severity in Korea. METHODS: We retrospectively reviewed 16,888 records of cirrhotic patients from six tertiary hospitals in Korea from 2008 to 2017. Continuous and non-continuous variables were processed via linear and Poisson regression, expressed as beta (B) coefficients and as exponentiated values of coefficients (Exp[B]), respectively. RESULTS: Chronic hepatitis B showed a decreasing trend (Exp[B] = 0.975, P < 0.001), whereas alcohol showed an increasing trend (Exp[B] = 1.013, P = 0.003), occupying the most common etiology in 2017. The Child-Turcotte-Pugh (CTP) score and decompensated liver cirrhosis prevalence did not change over the 10-year period. The incidence of variceal bleeding, severe ascites, hepatic encephalopathy, and spontaneous bacterial peritonitis significantly decreased from 12.3% to 7.7%, 7.8% to 4.1%, 1.0% to 0.5%, and 1.9% to 1.1%, respectively (P < 0.05 for all). In the subgroup analysis, the chronic hepatitis B group showed improving CTP scores (B = -0.025, P < 0.001) and decreasing decompensated liver cirrhosis rates (Exp[B] = 0.977, P = 0.016), whereas the alcohol group demonstrated increasing CTP class C (Exp[B] = 1.031, P = 0.005) and model for end-stage liver disease scores (B = 0.081, P = 0.005) over 10 years. CONCLUSION: The chronic hepatitis B group exhibited improved results, whereas the alcohol group still presented poor liver functions and outcomes. Future national policies and systematic approaches addressing the incidence, prevention, and treatment of alcoholic liver cirrhosis are indispensable.

Diagnostic usefulness of the spot urine sodium/potassium ratio in cirrhotic patients with ascites.

BACKGROUND AND AIMS: The low-salt diet is considered important for control of ascites in cirrhotic patients. To validate whether the spot urine sodium (Na)/potassium (K) ratio could replace 24-h urine Na (uNa) excretion in assessing low-salt diet compliance. METHODS: We prospectively studied 175 patients. 24-h urine collection and spot urine collection were performed. Subsequently, 24-h uNa, urine creatinine (uCr), and spot urine Na and K were assessed. A complete urine collection was confirmed based on 24-h uCr excretion levels of 15mg/kg/day for men and 10mg/kg/day for women. The area under the receiver operating characteristic (AUROC) curve analysis was performed to evaluate the feasibility of spot urine Na/K ratio in predicting 24-h uNa greater than 78mmol/day. RESULTS: Out of 175 patients, 24-h urine samples were completely collected in 57 patients only. Moreover, urine samples were not completely collected in 118 patients because their 24-h uCr excretion level was less than the established criteria. In complete urine collection group, AUROC curve for spot urine Na/K ratio in predicting 24-h uNa greater than 78mmol/day was 0.874±0.051 (P<0.001). In the incomplete urine collection group, the AUROC was 0.832±0.039 (P<0.001). In complete urine collection group, the classical cutoff value greater than 1.0 of spot urine Na/K ratio showed 90.9% sensitivity and 56.0% specificity. CONCLUSIONS: The spot urine Na/K ratio reflects 24-h uNa, but the AUROC value obtained in this study is lower than that of a previous study. Considered the large number of patients with incomplete urine collection, validating 24-h complete urine collection criteria is necessary.

Type 17 immunity promotes the exhaustion of CD8+ T cells in cancer.

BACKGROUND: Multiple types of immune cells producing IL-17 are found in the tumor microenvironment. However, their roles in tumor progression and exhaustion of CD8+ tumor-infiltrating lymphocytes (TILs) remain unclear. METHODS: To determine the role of type 17 immunity in tumor, we investigated the growth of B16F10 melanoma and the exhaustion of CD8+ TILs in Il17a-/- mice, Il17aCreR26DTA mice, RORγt inhibitor-treated mice, or their respective control mice. Adoptive transfer of tumor-specific IL-17-producing T cells was performed in B16F10-bearing congenic mice. Anti-CD4 or anti-Ly6G antibodies were used to deplete CD4+ T cells or CD11b+Gr-1hi myeloid cells in vivo, respectively. Correlation between type 17 immunity and T cell exhaustion in human cancer was evaluated by interrogating TCGA dataset. RESULTS: Depletion of CD4+ T cells promotes the exhaustion of CD8+ T cells with a concomitant increase in IL-17-producing CD8+ T (Tc17) cells in the tumor. Unlike IFN-γ-producing CD8+ T (Tc1) cells, tumor-infiltrating Tc17 cells exhibit CD103+KLRG1-IL-7Rαhi tissue resident memory-like phenotypes and are poorly cytolytic. Adoptive transfer of IL-17-producing tumor-specific T cells increases, while depletion of IL-17-producing cells decreases, the frequency of PD-1hiTim3+TOX+ terminally exhausted CD8+ T cells in the tumor. Blockade of IL-17 or RORγt pathway inhibits exhaustion of CD8+ T cells and also delays tumor growth in vivo. Consistent with these results, human TCGA analyses reveal a strong positive correlation between type 17 and CD8+ T cell exhaustion signature gene sets in multiple cancers. CONCLUSION: IL-17-producing cells promote terminal exhaustion of CD8+ T cells and tumor progression in vivo, which can be reversed by blockade of IL-17 or RORγt pathway. These findings unveil a novel role for IL-17-producing cells as tumor-promoting cells facilitating CD8+ T cell exhaustion, and propose type 17 immunity as a promising target for cancer immunotherapy.

A critical role for Th17 cell-derived TGF-ß1 in regulating the stability and pathogenicity of autoimmune Th17 cells.

Pathogenic conversion of Th17 cells into multifunctional helper T cells or Th1 cells contributes to the pathogenesis of autoimmune diseases; however, the mechanism regulating the plasticity of Th17 cells remains unclear. Here, we found that Th17 cells expressed latent TGF-ß1 in a manner dependent on autocrine TGF-ß1. By employing IL-17-producing cell-specific Tgfb1 conditional knockout and fate-mapping systems, we demonstrated that TGF-ß1-deficient Th17 cells are relatively susceptible to becoming IFN-γ producers through IL-12Rß2 and IL-27Rα upregulation. TGF-ß1-deficient Th17 cells exacerbated tissue inflammation compared to TGF-ß1-sufficient Th17 cells in adoptive transfer models of experimental autoimmune encephalomyelitis and colitis. Thus, TGF-ß1 production by Th17 cells provides an essential autocrine signal for maintaining the stability and regulating the pathogenicity of Th17 cells in vivo.

Deep metric learning-based image retrieval system for chest radiograph and its clinical applications in COVID-19.

In recent years, deep learning-based image analysis methods have been widely applied in computer-aided detection, diagnosis and prognosis, and has shown its value during the public health crisis of the novel coronavirus disease 2019 (COVID-19) pandemic. Chest radiograph (CXR) has been playing a crucial role in COVID-19 patient triaging, diagnosing and monitoring, particularly in the United States. Considering the mixed and unspecific signals in CXR, an image retrieval model of CXR that provides both similar images and associated clinical information can be more clinically meaningful than a direct image diagnostic model. In this work we develop a novel CXR image retrieval model based on deep metric learning. Unlike traditional diagnostic models which aim at learning the direct mapping from images to labels, the proposed model aims at learning the optimized embedding space of images, where images with the same labels and similar contents are pulled together. The proposed model utilizes multi-similarity loss with hard-mining sampling strategy and attention mechanism to learn the optimized embedding space, and provides similar images, the visualizations of disease-related attention maps and useful clinical information to assist clinical decisions. The model is trained and validated on an international multi-site COVID-19 dataset collected from 3 different sources. Experimental results of COVID-19 image retrieval and diagnosis tasks show that the proposed model can serve as a robust solution for CXR analysis and patient management for COVID-19. The model is also tested on its transferability on a different clinical decision support task for COVID-19, where the pre-trained model is applied to extract image features from a new dataset without any further training. The extracted features are then combined with COVID-19 patient's vitals, lab tests and medical histories to predict the possibility of airway intubation in 72 hours, which is strongly associated with patient prognosis, and is crucial for patient care and hospital resource planning. These results demonstrate our deep metric learning based image retrieval model is highly efficient in the CXR retrieval, diagnosis and prognosis, and thus has great clinical value for the treatment and management of COVID-19 patients.