Treating Late-Stage Tumor Immune Suppression: Multi-Modal TGFBR1 Drug Design

Hi, i'm Mohamed and I hope you enjoy my video! :)

Cancer remains one of the leading causes of death globally, and in many late-stage cancers, treatment failure is not only due to tumor growth, but because the tumor actively suppresses the immune system. In these environments, known as the tumor microenvironment (TME), signaling pathways such as TGF-beta inhibit natural killer (NK) cell activity, which prevents the body from effectively destroying cancer cells.

This project was inspired by my personal experience and interest in cancer, as I have dealt with it first hand through family, and my deep interest in cancer biology, as well as the urgent need for more effective treatment strategies for advanced cancers. Rather than focusing solely on tumor growth, I aimed to explore how restoring immune function could improve outcomes, particularly in late-stage disease where current therapies fail.

My question of this project was: Can targeted inhibition of TGFBR1 restore immune activity in suppressed tumor environments, while maintaining drug-like properties and delivery potential?

To address this, I combined bioinformatics, computational drug design, and pharmacokinetic analysis to develop and evaluate novel TGFBR1-targeted compounds. Additionally, I explored a modular click-chemistry approach to enable future targeted delivery strategies.

This work has the potential to contribute to more effective, adaptable cancer treatments by shifting the focus from tumor suppression alone to restoring the body's natural immune response.

1. Bioinformatic Analysis (Tumor Microenvironment)

I began by conducting background research using scientific literature and gene expression databases such as GEPIA2, STRING, Reactome, and the NCBI Gene Expression Omnibus. Through this process, I identified key genes involved in immune suppression within the tumor microenvironment (TME), including TGFBR1, HIF1A, and LDHA. I compared gene expression between tumor and normal tissues and analyzed patterns across cancer types and stages to confirm the role of TGF-beta signaling in suppressing immune function.

2. Molecular Docking and Virtual Screening

To design my potential drug, I used Schrödinger, which is a computational chemistry tool. The 3D structure of TGFBR1 was obtained from the RCSB Protein Data Bank and the protein was prepared in Schrödinger by removing water molecules and adding hydrogens, after that I removed the co-crystallized ligand. I screened multiple candidate molecules using virtual screening methods and selected top performing compounds based on docking scores. These candidates were further analyzed using induced-fit docking (IFD) to better simulate real binding interactions.

2. ADME and Drug-Likeness Analysis

In order to evaluate whether the compounds could function as real drugs, I used SwissADME to analyze their pharmacokinetic properties. Each molecule was converted into SMILES format and assessed for absorption, distribution, metabolism, and excretion (ADME). I specifically evaluated properties such as blood-brain barrier permeability, P-glycoprotein interaction, and Lipinski's Rule of Five to identify compounds with strong drug like characteristics.

3. Lead Optimization and Interaction Analysis

I then analyzed ligand-protein interactions using visualization tools within Schrödinger. This allowed me to identify key amino acid residues involved in binding to determine which parts of the molecule were essential for interaction.

3. Click Chemistry Design (Modular Targeting Strategy)

A solvent-exposed region of the molecule was identified and modified with a click-compatible handle, allowing future attachment of targeting ligands for improved delivery.

5. Bioinformatic Analysis Results

-Gene Expression Results (GEPIA2):

Gene expression analysis using GEPIA2 showed that TGFBR1 is consistently expressed across multiple cancer types, supporting its role in tumor progression and immune suppression. Additional genes associated with the tumor microenvironment, including HIF1A and LDHA, demonstrated patterns consistent with hypoxia-driven metabolic activity. In contrast, cytotoxic markers such as GZMB and PRF1 were reduced in later stages, indicating suppressed NK cell function.

-Protein Interaction Network (STRING)

Protein-protein interaction analysis using GEPIA2 showed that TGFBR1 is highly connected within networks related to immune suppression, hypoxia signaling, and metabolic regulation. These interactions reinforce its role as a central regulator in the tumor microenvironment and support its selection as a therapeutic target.

-Pathway Mapping (Reactome)

Pathway analysis using Reactome identified that TGFBR1 is involved in key biological pathways, including TGF-beta signaling, hypoxia response, and metabolic reprogramming. These pathways are known to contribute to immune suppression and tumor progression, further supporting TGFBR1 as a relevant therapeutic target.

6. Molecular Docking and Virtual Screening Results

Virtual Screening Results

Virtual screening of a large compound library ( 4 million compounds) identified multiple candidate molecules with strong predicted binding potential to TGFBR1. These hits were ranked based on docking scores, and top performing compounds were selected for further analysis. The screening process allowed efficient identification of promising candidates beyond the initially designed molecule.

Molecular Docking Results

Docking results demonstrated that the designed lead compound binds strongly to the TGFBR1 active site. Binding affinity scores were comparable to, and in some cases exceeded, known inhibitors such as Galunisertib and Vactosertib. Induced fit docking further confirmed stable ligand protein interactions, with key residues contributing to hydrogen bonding and hydrophobic stabilization. My drug also predicted lower binding to the off-target protein p38a.

6. Comparative Drug Analysis

When compared to clinically studied inhibitors, the designed compound showed competitive binding performance while maintaining a unique structural profile. This finding suggests that alternative scaffolds can achieve similar or improved interaction with TGFBR1, supporting the potential for novel inhibitor development.

7. ADME and Drug-Likeness Results

ADME analysis revealed that structural optimization improved binding and certain drug-like properties; however, increased molecular complexity reduced predicted blood-brain barrier (BBB) permeability. This highlights a key tradeoff in drug design, where enhancing interaction specificity can negatively impact drug delivery.

This project demonstrated that TGFBR1 is in fact a biologically relevant and strategically valuable therapeutic target within late-stage "cold" tumors, where immune suppression limits treatment effectiveness. Bioinformatic analysis confirmed that immune dysfunction is driven by not only a lack of immune cell infiltration, but by inhibitory signaling and metabolic pathways within the tumor microenvironment.

Through computational drug design, a novel small-molecule inhibitor was developed and shown to exhibit strong and selective binding to TGFBR1, with docking performance stronger to clinically studied inhibitors such as Galunisertib and Vactosertib. Importantly, reduced binding towards the off-target protein p38a suggests improved specificity, which is crucial for minimizing side effects.

Through ADME analysis, I further demonstrated that the compound maintains favorable drug-like properties, including high absorption, low efflux risk, and minimal metabolic interference. While limitations in blood-brain barrier permeability were identified, this project addressed these challenges by proposing modular click chemistry strategy, which enables targeted delivery without compromising binding performance.

Overall, this work highlights the importance of integrating biological validation, molecular modeling, and pharmacokinetic analysis in drug development. It also shows that computational approaches can be used to design competitive therapeutic candidates while identifying and addressing real-world limitations, especially in the world of cancer.

The next step is to validate these findings through in vitro testing using cancer cell lines such as glioblastoma or pancreatic models. Experiments would assess whether the compound inhibits TGFBR1 signaling and improves immune related activity by measuring markers such as GZMB, PRF1, HIF1A, and LDHA. Co-culture systems with immune cells could be used to test effects on NK cell function. Further work would focus on improving BBB permeability, through testing the proposed click chemistry delivery strategy. Ultimately, in vivo studies would be required to evaluate therapeutic effectiveness, pharmacokinetics, and safety.

I'd like to first give a huge thanks to my mentor Samra Khan (PhD) from the University of Windsor, for assisting and giving me tremendous feedback for my project.

Next I'd like to thank Will Hosie (PhD) from the University of Windsor for helping me find my mentor.

Now I want to thank my chemistry teacher, Miss Mariam Angeli (PhD), for helping me with any chemistry questions I had, and amazing feedback!

Another huge thanks to Sara Waqas, a previous best project CWSF winner, for her advice when I had questions during the beginning of my project.

A tremendous thank you to the Trant Team at the University of Windsor for being so helpful and supportive throughout my journey, and allowing me to work in their lab.

I'd also like to thank a previous multiple time CWSF winner Tasnia Nabil (PhD), for being so supportive and reviewing my work.

National Center for Biotechnology Information. (2024). Research article. https://pmc.ncbi.nlm.nih.gov/articles/PMC11016125/

Reactome. (n.d.). TGF-beta receptor signaling activates SMADs. https://reactome.org/content/detail/R-HSA-2173789

Sigma-Aldrich. (n.d.). Click chemistry reagents overview. https://www.sigmaaldrich.com/CA/en/products/chemistry-and-biochemicals/chemical-biology/click-chemistry-reagents

National Center for Biotechnology Information. (2024). Article summary. https://pubmed.ncbi.nlm.nih.gov/40546335/

Cell Press. (2024). Research article. https://www.sciencedirect.com/science/article/pii/S2211124724015572

National Center for Biotechnology Information. (2024). Research article.

https://pmc.ncbi.nlm.nih.gov/articles/PMC11529905/

Cell Press. (2021). Research article. https://www.sciencedirect.com/science/article/pii/S1936523321000346

National Center for Biotechnology Information. (2024). Research article. https://pmc.ncbi.nlm.nih.gov/articles/PMC12816280/

National Center for Biotechnology Information. (2023). Research article. https://pmc.ncbi.nlm.nih.gov/articles/PMC10458686/

National Center for Biotechnology Information. (2019). Research article. https://pubmed.ncbi.nlm.nih.gov/30995507/

National Center for Biotechnology Information. (2012). Research article. https://pubmed.ncbi.nlm.nih.gov/22304911/

National Center for Biotechnology Information. (2009). Research article.

https://pubmed.ncbi.nlm.nih.gov/19460998/

National Center for Biotechnology Information. (2018). Research article. https://pubmed.ncbi.nlm.nih.gov/30476243/

National Center for Biotechnology Information. (2016). Research article. https://pubmed.ncbi.nlm.nih.gov/26884601/

National Center for Biotechnology Information. (2022). Research article. https://pmc.ncbi.nlm.nih.gov/articles/PMC9623302/

National Center for Biotechnology Information. (2022). Research article. https://pmc.ncbi.nlm.nih.gov/articles/PMC9545774/

National Center for Biotechnology Information. (2018). Research article. https://pubmed.ncbi.nlm.nih.gov/29686425/

Nucleic Acids Research. (2020). Database update. https://academic.oup.com/nar/article/48/D1/D498/5613674

CB-Dock. (n.d.). Protein–ligand docking tool. http://183.56.231.194:8001/cb-dock2/index.php

SwissADME. (n.d.). Pharmacokinetics prediction tool. https://www.swissadme.ch/

GEPIA2. (n.d.). Gene expression analysis platform. http://gepia2.cancer-pku.cn/#index

Royal Society of Chemistry. (n.d.). Fluorine element data. https://periodic-table.rsc.org/element/9/fluorine

Videos

YouTube. (n.d.). Short video. https://www.youtube.com/shorts/1J6TO7cWndY

YouTube. (n.d.). Educational video. https://www.youtube.com/watch?v=3PwVWX28dEE

YouTube. (n.d.). Educational video. https://www.youtube.com/watch?v=gIoJRJiZGSQ