Research Theme I: Cytokine Autoantibodies and Severe Infections in Adults

Recent studies have increasingly identified cytokine autoantibodies as a key cause of severe infections in adults without known primary immunodeficiency. Our laboratory focuses on elucidating the pathogenic mechanisms of anti–interferon‑gamma (anti‑IFN‑γ) autoantibodies and anti–granulocyte‑macrophage colony‑stimulating factor (anti‑GM‑CSF) autoantibodies, which are strongly associated with life‑threatening infectious diseases.

I. Anti–Interferon‑Gamma Autoantibodies (Anti‑IFN‑γ Autoantibodies, AIGA)
Anti‑IFN‑γ autoantibodies are frequently observed in patients from Taiwan and are commonly associated with nontuberculous mycobacterial (NTM) infections. We first identified a strong association between this disease and specific human leukocyte antigen (HLA) genotypes, DRB116:02 and DQB105:02, indicating a genetic predisposition to anti‑IFN‑γ autoantibody–associated disease (Chi et al., Blood, 2013).
By expanding patient recruitment to include 145 patients from multiple countries, particularly Southeast Asia, we further demonstrated a strong genetic association between anti‑IFN‑γ autoantibodies and the HLA genotypes DRB115:02/16:02 and DQB105:01/05:02. The strength of this association suggests that these HLA haplotypes may directly contribute to the induction of anti‑IFN‑γ autoantibodies (Ku et al., Journal of Allergy and Clinical Immunology, 2016).
The immunodeficiency caused by anti‑IFN‑γ autoantibodies also predisposes patients to Talaromyces marneffei infection. More than 95% of non‑HIV patients with talaromycosis were found to carry anti‑IFN‑γ autoantibodies. Clinically, this finding explains the geographic distribution of this fungal infection in southern China, while scientifically highlighting the critical role of interferon‑γ in controlling fungal infections in humans (Guo et al., Journal of Experimental Medicine, 2020).
Our research further indicates that anti‑IFN‑γ autoantibodies may be induced by Aspergillus infection through a mechanism of molecular mimicry. Based on this discovery, we developed a novel recombinant interferon‑γ variant lacking immunodominant epitopes, enabling it to evade binding by anti‑IFN‑γ autoantibodies and restore interferon‑γ signaling. This strategy represents a potential therapeutic approach for patients suffering from persistent mycobacterial infections (Lin et al., Nature Medicine, 2016).
To better understand the neutralizing mechanisms of anti‑IFN‑γ autoantibodies, we employed single memory B‑cell cloning technology to isolate monoclonal anti‑IFN‑γ antibodies from patients. Our analysis revealed that these autoantibodies inhibit interferon‑γ function through multiple mechanisms, including
1. Blocking ligand‑receptor binding,
2. Formation of immune complexes, and
3. Antibody‑dependent cell‑mediated cytotoxicity (ADCC)
    (Shih et al., Journal of Experimental Medicine, 2022).

II. Anti‑GM‑CSF Autoantibodies
Cryptococcosis, a fungal infection that frequently causes severe meningitis and pneumonia, was historically considered a disease primarily affecting individuals with HIV or overt immunodeficiency. In recent years, however, cryptococcal infections have also been identified in immunocompetent individuals. Our studies revealed that a subset of these patients carry anti‑GM‑CSF autoantibodies (Kuo et al., Journal of Clinical Immunology, 2017).To further delineate the clinical features associated with these autoantibodies, we expanded our cohort and found that anti‑GM‑CSF autoantibodies are strongly associated with Cryptococcus gattii infections, particularly involving the central nervous system. Clinically, anti‑GM‑CSF autoantibodies are also known to be present in pulmonary alveolar proteinosis (PAP), another autoimmune disease.
Using serological comparisons between patients in Taiwan, we observed that although the prevalence of anti‑GM‑CSF autoantibodies did not differ significantly between cryptococcosis and PAP cohorts, the coexistence of these two diseases was rare. These findings suggest that the pathogenic mechanisms mediated by anti‑GM‑CSF autoantibodies are more complex than previously understood (Wang et al., Journal of Clinical Immunology, 2022).
Nocardiosis, another infectious disease capable of causing severe meningitis and pneumonia, shares similarities with cryptococcal infection. We identified anti‑GM‑CSF autoantibodies in immunocompetent patients with nocardial central nervous system infections (Lo et al., Journal of Clinical Immunology, 2024).
Our laboratory continues to investigate the neutralizing mechanisms of anti‑GM‑CSF autoantibodies to improve understanding of their role in infectious and immune‑mediated diseases (Lo et al., manuscript in preparation).
.

Research Theme II: Development of Emerging Immunotherapies for Pathogenic Cytokine Autoantibodies

Cytokine autoantibodies can neutralize or block essential cytokine signaling pathways, leading to immune dysregulation and severe infections. Currently, effective treatments for these conditions remain limited. Conventional approaches, such as long‑term antimicrobial therapy, provide only temporary infection control and do not eliminate pathogenic autoantibodies. Although B‑cell depletion therapies (e.g., anti‑CD20 monoclonal antibodies) have shown partial efficacy in some patients, recurrent disease, non‑responsiveness in certain individuals, and nonspecific impairment of humoral immunity limit their long‑term utility.
To address these challenges, our research team is developing next‑generation precision immunotherapies that selectively target pathogenic cytokine autoantibodies or their antibody‑producing cells. By integrating innovative protein and cellular engineering strategies, our goal is to preserve normal humoral immune function while achieving durable, safe, and effective therapeutic outcomes.

I. Development of a Cytokine Chimeric Autoantibody Receptor (CAAR) Screening Platform
To overcome the limitations of current therapies, we designed T cell–based Chimeric Autoantibody Receptor (CAAR) T cells that can specifically recognize and eliminate autoreactive B cells expressing anti‑IFN‑γ autoantibody (AIGA) B‑cell receptors. By integrating patient‑derived autoantibodies, structural information of interferon‑gamma (IFN‑γ) and its receptor, and a Jurkat NFAT reporter system, we established a comprehensive CAAR screening platform targeting anti‑cytokine autoantibodies.
Using this platform, we successfully identified modified IFN‑γ variants that lack biological activity but retain autoantibody recognition capacity, leading to the development of IFN‑γ CAAR T cells. In vitro and in vivo studies demonstrated that IFN‑γ CAAR T cells exhibit high specificity and safety, with no detectable receptor cross‑reactivity or Fc‑mediated off‑target toxicity. Importantly, these cells maintained functionality even under conditions mimicking high circulating AIGA levels.
We further demonstrated that IFN‑γ CAAR T cells can be successfully generated from patient‑derived autologous T cells and effectively eliminate autoreactive B cells within peripheral blood mononuclear cells. Collectively, these results establish IFN‑γ CAAR T cells as a promising therapeutic strategy for AIGA‑associated diseases (Peng et al., Science Immunology, 2025).
Notably, this CAAR platform is highly adaptable and can be extended to other anti‑cytokine autoantibody–associated diseases, such as those involving anti‑GM‑CSF or anti–type I interferon autoantibodies, which play critical roles in multiple life‑threatening infections and autoimmune conditions. This work provides a strong proof of concept and opens new avenues for immunotherapy targeting cytokine autoantibody–mediated diseases.

II. Enhancing the Clinical Translational Potential of Cytokine CAAR Therapy
While Cytokine CAAR T cells can selectively eliminate autoantibody‑producing B cells, high circulating levels of anti‑cytokine autoantibodies in patients may bind to CAARs, potentially suppressing function or triggering immune‑mediated rejection. Therefore, a major focus of this research direction is to elucidate rejection mechanisms induced by autoantibody binding, including NK cell–mediated antibody‑dependent cellular cytotoxicity (ADCC), phagocyte‑mediated antibody‑dependent cellular phagocytosis (ADCP), and complement‑dependent cytotoxicity (CDC).
Based on mechanistic insights, we are developing strategies to mitigate these immune rejection pathways, such as mimicking viral immune‑evasion mechanisms, incorporating “don’t‑eat‑me” signals to prevent phagocytosis, or introducing complement‑inhibitory proteins. These approaches are being integrated into Cytokine CAAR T cells to enhance their in vivo persistence and durability.
In parallel, we are developing in vivo Cytokine CAAR T‑cell generation platforms, including lipid nanoparticle (LNP) systems encapsulating CAAR‑encoding mRNA with T‑cell–targeting antibodies on the particle surface, as well as T‑cell‑tropic lentiviral vectors for direct in vivo T‑cell engineering. These approaches aim to significantly reduce manufacturing costs and accelerate the clinical translation of Cytokine CAAR therapies.
Beyond T cells, we are extending the CAAR strategy to other immune effector cells. For example, macrophages, which possess strong phagocytic and immune complex–clearing capabilities, may be particularly suitable for treating anti‑GM‑CSF autoantibody–mediated pulmonary alveolar proteinosis (PAP). By matching optimal effector cells to specific disease characteristics, we aim to establish a flexible and versatile CAAR effector cell platform.

III. Development of Anti‑Cytokine Autoantibody Degraders
To selectively eliminate pathogenic cytokine autoantibodies, we are developing anti‑cytokine autoantibody degraders. These molecules use engineered cytokine fragments as decoys to capture disease‑causing autoantibodies and redirect them to lysosomal degradation pathways via specific receptors such as IGF2R or ASGPR.
Unlike FcRn inhibitors that broadly suppress IgG recycling, this strategy enables selective removal of target autoantibodies, minimizing global immunosuppression and infection risk. Anti‑cytokine autoantibody degraders therefore represent a precise and safer therapeutic modality for autoantibody‑mediated diseases.

IV. Identification of Novel Therapeutic Targets in Anti‑Cytokine Autoantibody Diseases
Beyond autoreactive B cells, other pathogenic immune populations—including autoreactive T follicular helper (Tfh) cells and long‑lived plasma cells—remain difficult to identify and target. Their roles in disease initiation, tissue localization, and interactions with other immune cells remain poorly understood.
Our goal is to develop strategies to label, track, and functionally characterize these cells, enabling deeper insights into disease networks. Through systematic analysis, we aim to uncover novel, druggable molecular targets and open new therapeutic avenues for diseases driven by anti‑cytokine autoantibodies.
.

Research Theme III: Cancer Immunity and Therapy

The core concept of cancer immunity lies in harnessing the immune system’s ability to recognize and eliminate tumor cells, thereby overcoming the lack of specificity and adverse side effects associated with conventional therapies such as surgery, chemotherapy, and radiotherapy. Cancer immunotherapy is based on targeting tumor‑associated antigens or tumor‑specific antigens to activate or reprogram immune cells, enabling them to effectively identify and destroy cancer cells while maintaining immune tolerance toward normal tissues.
Immune checkpoint blockade (ICB) and chimeric antigen receptor T‑cell (CAR‑T) therapy have profoundly transformed the treatment landscape of certain cancers, including melanoma, lung cancer, and B‑cell malignancies, offering therapeutic benefits beyond conventional approaches. ICB restores the function of exhausted T cells by blocking inhibitory pathways such as PD‑1/PD‑L1 and CTLA‑4, whereas CAR‑T cells are genetically engineered to express tumor‑specific receptors that directly mediate cytotoxicity against cancer cells.
Despite these advances, significant challenges remain. ICB shows limited efficacy in so‑called “immunologically cold” tumors and is frequently associated with immune‑related adverse events. Although CAR‑T therapy has achieved remarkable success in hematological malignancies, its application to solid tumors is hindered by the immunosuppressive tumor microenvironment, antigen heterogeneity, and insufficient persistence of infused cells.
These limitations have shifted research focus toward next‑generation cellular therapies, including non‑conventional lymphocyte populations such as γδ T cells and natural killer (NK) cells, as well as engineered platforms such as CAR‑NK and CAR‑γδ T cells. These strategies enable MHC‑independent tumor recognition, carry a lower risk of alloreactivity, and hold promise as off‑the‑shelf cancer immunotherapies.
To explore the potential of next‑generation cell therapies, our laboratory is dedicated to the development of γδ T cells, which bridge innate and adaptive immunity. Our approach encompasses three major directions:
(1) developing and optimizing ex vivo expansion protocols for γδ T cells;
(2) performing high‑throughput, genome‑wide CRISPR/Cas9 screening to identify editable genetic modules; and
(3) collaborating with Chang Gung Memorial Hospital to identify tumor‑specific γδ T‑cell receptors (γδ TCRs) from cancer patients, enabling the development of precision TCR‑T cell therapies.
.

Research Theme IV: Severe Viral Infections and Inborn Errors of Immunity

Accumulating evidence indicates that severe viral infections in a subset of children may result from dysregulated antiviral pathways caused by inborn errors of immunity (IEI). Based on this central hypothesis, we aim to establish a national research cohort of Taiwanese children with severe viral infections. By integrating detailed clinical manifestations with immunophenotypic analyses, we systematically investigate the molecular and genetic mechanisms underlying disease severity.
Identification of Inborn Errors of Immunity Using Sequencing and Next-Generation Sequencing Technologies.
Our previous research has demonstrated that severe enteroviral encephalitis can be caused by functional defects in TLR3 signaling (Kuo, Journal of Clinical Immunology, 2022). In the context of chronic mucocutaneous candidiasis (CMC), both prior literature and our recent work indicate that abnormalities in STAT1 or genes involved in the Th17 pathway are closely associated with disease susceptibility (Lei, Journal of Clinical Immunology, 2024). These findings support the concept that specific genetic defects confer susceptibility to specific infectious diseases.
Building on this foundation, we will employ next-generation sequencing (NGS) combined with functional validation to identify pathogenic variants in Taiwanese children with severe viral infections. Our goal is to link genetic variants to dysregulated immune pathways and to develop translational genetic diagnostic tools and precision therapeutic strategies. Ultimately, this research is expected to have a substantial clinical impact on disease prevention, prognostic stratification, and targeted treatment.
.

Research Theme V: Pathogenic Mechanisms of Renal Autoantibody‑Mediated Diseases

Our research team is dedicated to investigating the pathogenic mechanisms underlying immune‑mediated kidney diseases, with a particular focus on how autoimmune responses and pathogenic autoantibodies contribute to renal injury and disease progression.
Our research team is dedicated to elucidating the mechanisms underlying immune‑mediated kidney diseases, with a particular emphasis on how autoimmune responses lead to renal injury. Our current research focuses on membranous nephropathy (MN), especially the pathogenic processes driven by anti‑phospholipase A2 receptor (anti‑PLA2R) autoantibodies.
Membranous nephropathy is a common autoimmune kidney disease characterized by autoantibody‑mediated injury to glomerular podocytes, resulting in heavy proteinuria and progressive deterioration of renal function. Among the identified autoantibodies, anti‑PLA2R is the most prevalent and representative pathogenic factor worldwide.
Our laboratory has established a long‑term cohort of MN patients in Taiwan and integrates clinical sample analysis, single‑cell technologies, and antibody engineering to investigate the structural features of anti‑PLA2R antibodies, their IgG subclass distribution, and their roles in complement activation and kidney damage. Through these studies, we aim to delineate the causal relationship between autoantibodies and disease progression and to uncover the underlying molecular mechanisms of MN pathogenesis. Ultimately, our goal is to translate these findings into novel diagnostic tools and targeted therapeutic strategies, enabling more precise and effective clinical care for patients.
.
Go Back