Dr. Ho's laboratory studies biochemistry and cell biology of tumor-associated proteins in broad scientific fields of molecular and cellular biology including receptor/ligand interactions, signaling pathways, antibody/protein engineering, structure and computational biology, and functional genomics. We also develop new antibody engineering tools such as mammalian cell surface display and single domain antibody libraries to advance drug discovery against cancer and virus infection. Some of our research has direct clinical application.

Dissecting cell surface receptors in cancer: structure and function

Antibody-based therapy has shown promising efficacy in hematologic tumors. However, this approach has shown limitations in most solid cancers. One of the major challenges is the lack of cancer-specific targets in many cancers. Our long-term research interests lie primarily in the biochemistry and cell biology of cell surface receptors including glypicans (GPC3, GPC2, and GPC1) for characterizing them as new cancer targets.

Heparan and chondroitin sulfate proteoglycans (HSPGs and CSPGs, respectively) are important modulators of signal transduction pathways during development and disease. HSPGs mainly consist of glycosylphosphatidylinositol (GPI)-anchored glypicans and transmembrane syndecans. Six members (GPC1 to GPC6) of the human glypican fall into two broad subfamilies, GPC1/2/4/6 and GPC3/5. Much of our work has focused on fundamental aspects of glypicans in connection to Wnt, Yap and other oncogenic signaling pathways. Our goal is to investigate glypicans such as GPC3, GPC2 and GPC1 as new therapeutic targets in cancer.

GPC3 is highly expressed in over 70% of hepatocellular carcinoma, a major type of primary liver cancer, but not in normal adult tissues. We identified two Wnt functional binding sites on GPC3, one in the protein core and the other in the heparan sulfate chains. We built a structural model of the GPC3/Wnt complex in collaboration with Byungkook Lee (NCI) and experimentally determined the hydrophobic groove at the N-lobe of the protein core of GPC3 for the Wnt functional binding site, providing evidence for GPC3 as a co-receptor for regulating Wnt. Our research also revealed the role of GPC3 in regulating Yap for cancer cell proliferation and discovered the interaction between GPC3 and HGF in liver cancer. Recently, we identified FAT1, a potential cell surface receptor of Yap signaling in mammalian cells, as a novel GPC3 interacting protein in liver cancer cells. Furthermore, we found that a single domain antibody (HN3) was capable of inhibiting Wnt/β-catenin signaling by reaching the Wnt functional binding groove on GPC3. Additionally, we used our HS20 human antibody specific for the heparan sulfate chains of GPC3 to determine the Wnt binding motif which may function as a Wnt storage/transporter site on heparan sulfate chains. In collaboration with Jian Liu (University of North Carolina), we found that Wnt recognized a heparan sulfate motif containing IdoA2S and GlcNS6S, and that the 3-O-sulfation in GlcNS6S3S could enhance the binding of Wnt.

In recent years, we started to apply what we learned from our work on GPC3 to explore other glypicans in cancer. We found that GPC2 protein was expressed in nearly half of neuroblastoma cases and that GPC2 knockout inactivated Wnt/β-catenin signaling and reduced the expression of the target gene, N-Myc, an oncogenic driver of neuroblastoma tumorigenesis. Furthermore, we conducted a proof-of-concept study that showed CAR T cells and immunotoxins targeting GPC2 could inhibit neuroblastoma growth in mouse models. As a result, we reported GPC2 to be a new therapeutic target in neuroblastoma.

GPC1 is overexpressed in multiple types of cancers, including pancreatic cancer. GPC1 modulates various signaling pathways including FGF, VEGF-A, TGF-β, Wnt, Hh, and BMP through its interactions with pathway ligands and receptors. For treating cancer, we have constructed the GPC1-targeted immunotoxin. The immunotoxin inhibits pancreatic tumor growth via degradation of internalized GPC1, downregulation of Wnt signaling, and inhibition of protein synthesis. We find that GPC1 has a uniquely much shorter intracellular half-life (15 minutes) than those of GPC2 (1 hour) and GPC3 (2 hours), other members of the glypican family. Immunotoxin treatment could reduce the expression levels of short-lived GPC1 through lysosomal degradation following internalization. The generation of new GPC1 molecules is blocked by immunotoxin-induced protein synthesis inhibition. The downregulated GPC1 can inhibit the downstream Wnt signaling cascade by reducing the levels of active–β-catenin.

Mesothelin is highly expressed in many solid tumors including mesothelioma and ovarian cancer. To understand its biological function, we collaborated with Byungkook Lee (NCI) to make a structure model and experimentally identified the functional binding domain for MUC16/CA125 (residues 296-359; named IAB) at the N-terminus of cell surface mesothelin. Along with Ira Pastan (NCI) and Manish Patankar (University of Wisconsin-Madison), we found that the MUC16-associated N-glycans were required for its binding to mesothelin. Our work reveals the molecular mechanism by which mesothelin regulates MUC16/CA125 in cancer pathogenesis.

Engineering antibody therapeutics for treating cancer: immunotoxins and CAR T cells

To overcome the limitations of antibody-based immunotherapy for solid tumors, we used multiple antibody technologies to generate large panels of human and xenogeneic antibodies and utilized multiple effector mechanisms to improve the anti-tumor activity of our tumor antigen-specific monoclonal antibodies. We have developed multiple strategies with different mechanisms of action, such as immunotoxins, ADCs and CAR T cells. Our research program focuses on the development and implementation of antibody-based immunotherapeutic strategies for the treatment of solid tumors, including hepatocellular carcinoma (HCC), pancreatic cancer, neuroblastoma, and mesothelioma. The strategies developed in our program may be applicable to other solid tumors.

To validate GPC3 as a cancer therapeutic target, we generated antibodies (HN3 and YP7) that bind GPC3 on liver cancer cells. HN3 is a human single domain antibody isolated using phage display technology, which recognizes a cryptic Wnt binding site in the N-lobe of GPC3. YP7 is a monoclonal antibody isolated using mouse hybridoma technology, which binds an epitope at the C-lobe of GPC3 close to cell surface. Using these antibodies, we generated immunotoxins and ADCs to target GPC3-expressing cancers. We constructed a novel immunotoxin that can produce tumor regression via a dual mechanism of action. These immunotoxins inactivate cancer signaling (Wnt and Yap) via the antibody domain and inhibit protein synthesis via the toxin portion. This dual mechanism of action may be applicable to other antibody-toxin/drug conjugates for better anti-tumor efficacy. Immunogenicity and a short serum half-life may limit the ability of immunotoxins to transition to the clinic. To address these concerns, we engineered HN3-based immunotoxins in collaboration with Ira Pastan (NCI) to use various deimmunized Pseudomonas exotoxin (PE) domains and added an albumin binding single domain. We also designed CAR T cells using our GPC3 antibodies and demonstrated that humanized YP7 (hYP7) antibody-derived CAR T cells had potent and persistent anti-tumor activity in orthotopic liver cancer mouse models. Our collaborator Tim Greten (NCI) is leading the clinical trial at the NIH using our hYP7 CAR T cells for treating liver cancer patients.

Using RNA-seq analysis, we showed that exon 3 and exons 7-10 of GPC2 were expressed in cancer but were minimally expressed in normal tissues. Accordingly, we discovered a monoclonal antibody (CT3) that recognized exons 3 and 10 and visualized the complex structure of CT3 and GPC2 by electron microscopy. The potential of this approach was exemplified by designing CT3-derived CAR T cells that regressed neuroblastoma in mice. Genomic sequencing of T cells recovered from mice revealed the CAR integration sites that may contribute to CAR T cell proliferation and persistence. These studies demonstrate how RNA-seq data can be exploited to help identify tumor-associated exons that can be targeted by CAR T cell therapies. The CT3-derived CAR T cells are currently being developed for a clinical trial at the NIH for treating neuroblastoma patients in our collaboration with the Pediatric Oncology Branch at the NCI.

Mesothelin is highly expressed in mesothelioma and other major solid tumors. We previously isolated HN1, a human monoclonal antibody that disrupts the mesothelin-MUC16 interaction. Our recent research has been focused on the targeting of the membrane-proximal region of mesothelin for enhanced anti-tumor activities. As a result, we isolated YP218, a unique rabbit monoclonal antibody targeting the C-terminal portion of mesothelin close to the tumor cell surface. We developed CAR T cells based on the humanized YP218 and showed they were very effective against mesothelin-positive tumors in mouse models in collaboration with Raffit Hassan (NCI).

Developing antibody engineering technology: mammalian cell surface display and single domain antibodies

We have developed a variety of antibody engineering technologies to produce therapeutic antibodies. These include development of mammalian cell surface display, construction of shark and camel single domain antibody phage libraries, and humanization of rabbit antibodies. Our antibody engineering technologies, such as single domain antibody libraries, can be used to identify therapeutic drugs and diagnostic agents for human diseases including cancer, infectious disease, and neurological disease.

Single domain antibodies (also commonly called nanobodies) are known to bind restricted epitopes that may be inaccessible to conventional antibodies. To make a large and diverse single domain phage library, we established a method based on PCR-Extension Assembly and Self-Ligation (named ‘EASeL’) to construct a V_NAR single domain antibody library from nurse sharks (Ginglymostoma cirratum) in collaboration with Martin Flajnik (University of Maryland). We conducted next-generation sequencing analysis of 1.2 million shark V_NAR single domains and showed that our shark single domain library is highly diverse. In addition, we applied the EASeL method to construct V_HH single domain phage libraries from dromedary camels (Camelus dromedaries). Single domain antibodies isolated from our phage libraries can have a high affinity (KD < 1 nM) for their tumor or viral antigens. We have used single domain antibodies to engineer CAR-T cells, immunotoxins, and other therapeutic scaffolds for treating cancer.

Antibody engineering is typically carried out by displaying antibody fragments on the surface of microorganisms (e.g. phage, bacteria and yeast). We established a new antibody engineering method known as 'mammalian cell surface display' that is adapted from yeast cell surface display. Using this approach, antibody fragments are expressed on human HEK-293 cells, and high affinity antigen binders are isolated from a combinatory library via flow cytometry.

To develop rabbit antibodies for clinical application, we analyzed the complex structures of rabbit antibodies with their antigens from the Protein Data Bank and identified antigen-contacting residues on the rabbit Fv within the 6 Angstrom distance to its antigen. Based on the structural analysis, we designed a humanization strategy by grafting the combined Kabat/IMGT/Paratome CDRs into a human germline framework sequence. The immunotoxins composed of the humanized rabbit Fvs (e.g. hYP218) fused to a clinically used toxin showed stronger cytotoxicity against tumor cells than the immunotoxins derived from their original rabbit Fvs. The CAR T cells based on the hYP218 antibody show effective inhibition of tumor growth in mice. Our method can be used as a general approach to humanize rabbit antibodies.

Drug resistance is an important component of tumor biology that requires a complex cellular environment for study. We have established ex vivo tumor spheroid models using cell lines and primary patient cells in collaboration with Shuichi Takayama (University of Michigan) and V. Courtney Broaddus (UCSF). This allows us to study the molecular mechanisms of drug resistance in a physiologically relevant cellular model. We have also used microarrays to profile gene expression in both spheroids and monolayers to identify new targets specific to the 3D biological structure of cancer.

Discovering antibody drugs for treating COVID-19: nanobodies broadly neutralizing SARS-CoV-2 variants

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the etiologic agent of COVID-19 that enters human cells by binding its envelope anchored type I fusion protein (spike) to angiotensin-converting enzyme 2 (ACE2). Numerous neutralizing antibodies targeting the spike, particularly its ACE2 receptor binding domain (RBD), have been developed to treat COVID-19 using common strategies such as single B cell cloning, animal immunization, and phage display. However, SARS-CoV-2 variants such as D614G, B.1.1.7, B.1.351, P.1, and B.1.617.2 have exhibited increased resistance to neutralization by monoclonal antibodies. Highly effective and broadly neutralizing antibody therapy is demanded for COVID-19 patients worldwide.

Due to their small size, nanobodies can recognize protein cavities that are not accessible to conventional antibodies. Dromedary camels have been found as potential natural reservoirs of MERS-CoV. We speculated that dromedary camels would be an ideal source of neutralizing nanobodies against coronaviruses. Using our dromedary camel (Camelus dromedarius) V_HH phage libraries, we successfully discovered the potent neutralizing nanobodies against SARS-CoV-2 variants. In collaboration with Mario Borgnia (NIEHS) and Raul Cachau (NCI), we solved the cryo-EM complex structures that reveal one V_HH nanobody (8A2) binds the S1 subunit of the viral spike protein, and the other (7A3) targets a deeply buried region that uniquely extends to the S2 subunit beyond the S1 subunit. In collaboration with Hang Xie (FDA), we found that these nanobodies could protect mice from the lethal challenge of variants B.1.351 or B.1.617.2, suggesting the therapeutic potential of these nanobodies against COVID-19. The dromedary camel V_HH libraries could be helpful to isolate neutralizing nanobodies against future emerging viruses quickly.

Teaching Interests

BIOC301/302 - Biochemistry I/II