3rd Place Global Finalist, SARC 2026

Abstract:

Glioblastoma (GBM) is still the most fatal and the most common primary brain tumor which occurs

among adult individuals. Its treatment has brought only a median of 15-month survival through

surgery, radiotherapy, and temozolomide. As a result, it is really common to face problems again

since the infiltrative residual cells remain heterogeneous and are protected from the blood-brain

barrier (BBB). This research project is about de novo trispecific miniprotein to be designed

to induce DR5-mediated apoptosis selectively only in EGFR+/B7-H3+ GBM cells while at the

same time preserving BBB integrity. The central hypothesis in this project is that a compact

trispecific construct can concentrate on EGFR+/B7-H3+ cells first, then cluster DR5 enough for

apoptosis. This target is utilized instead of a basic mono- or bispecific construct because in

selective binding only non-specific anchoring cannot trigger death signaling, but, passive DR5

agonism only cannot solve the problem of GBM heterogeneity. The overall approach combines

computational design, commercial synthesis and fluorescent labeling of the two top-ranked

candidates, and proof-of-concept testing in a human BBB/GBM-on-a-chip assay within a bounded

BSL-2 workflow.

​

Introduction:​​​

GBM remains difficult to treat because failure reflects more than inadequate cytotoxicity. Standard

treatment reduces tumor burden, but recurrence remains frequent because residual cells infiltrate

surrounding brain, persist across diverse signaling states, and are difficult to reach with large

therapeutics. A useful strategy must therefore address selectivity, transport, and mechanism

together. The present study uses a design-based approach rather than a repurposed drug. Its

therapeutic concept is a compact engineered protein that activates a death pathway only under a

defined surface condition. Dual anchoring to EGFR and B7-H3 is intended to restrict activity to

a GBM cell population with a high-risk molecular profile, after which a DR5-facing surface is

intended to promote apoptosis.

Research Question. Can a de novo trispecific miniprotein be computationally designed

to selectively bind EGFR+/B7-H3+ glioblastoma cells, preserve BBB integrity, and induce

DR5-mediated apoptosis in a human BBB/GBM-on-a-chip model?

​

Literature Review:​ 

There are three points that constitute the biological basis of the study. Dysregulation of the EGFR

is often found in GBM and is a signal of a more aggressive tumor behavior, so it is easy to set

it as the target anchor (Ezzati et al., 2024). B7-H3/CD276 is predominant in the glioblastoma

disease and is related to the poor prognosis, so the co-expression of EGFR/B7-H3 provides a more

selective surface than a single-marker approach (Babic et al., 2024). Clustering of DR5-linked

apoptotic signaling has always found its relevance in the investigations on GBM, thus making it

a sensible explanation as death-receptor clustering for the fast tumor-cell destruction (Qiao et al.,

2025; Thang et al., 2023). Large biologic agents, in most cases, are not advantageous in the central

nervous system contexts as BBB transport is the limiting factor, but smaller protein scaffolds are

only a testable starting point and not an automatic BBB solution, so the chip still needs to check

transport and barrier safety (Hajal et al., 2022; Luo et al., 2022; Straehla et al., 2022). The latest

generative techniques for protein design have turned new therapeutic discovery into a reality, as

they enable backbone generation, sequence refinement, complex evaluation, and stability testing

in a single pipeline (Abramson et al., 2024; Dauparas et al., 2022; Watson et al., 2023).

​

Methodology:

First, extracellular epitopes on EGFR and B7-H3 will be chosen from the structural databases and

GBM literature according to the criteria of extracellular site, recurrence, and possible co-expression

in aggressive cancer cells. Large libraries of target-conditioned binder backbones for both receptors

will be produced by RFdiffusion. ProteinMPNN will give out the sequences, and the candidates

will be selected after they have been checked for compactness, buried unsatisfied polar atoms,

aggregation liability, monomer fold plausibility, and predicted soluble protein expression. The

highest-affinity EGFR and B7-H3 binders will be integrated into a compact DR5-agonizing

module to create a trispecific construct, and the aim of this is to cluster DR5 only following dual

tumor-antigen engagement. AlphaFold 3 will consider the folded monomers and the assemblies

that are bound to the receptor, while the GROMACS simulations will still check the interface

persistence, the linker behavior, and the conformational stability. Only two leads will be

moved forward, and only if they meet the pre-registered conditions of high predicted assembly

confidence, low aggregation risk, acceptable post-synthesis solubility, and preservation of the

original DR5-presenting geometry after fluorescent labeling.

The computational tasks will be completed first before synthesis and chip work. This proposal

discusses a supervised partner-lab pilot rather than student-owned wet-lab work; University of the

Philippines or Ateneo de Manila may support research supervision, cell culture, flow cytometry,

and imaging, while constructs are synthesized commercially and the chip is used only through

a partner with platform or ready-made device access. A reasonable timeline is four weeks for

design and down-selection, two weeks for synthesis and labeling, and four weeks for pilot chip

experiments and analysis. The chip assay will compare the two leads with the vehicle, scrambled

miniprotein, and a full-length IgG-sized control. Specificity controls will entail EGFR+/B7-H3+,

EGFR+/B7-H3-, EGFR-/B7-H3+, EGFR-/B7-H3-, and a non-tumor human comparator that

is relevant to BBB safety. The readouts will consist of trans-endothelial transport, dextran

permeability, endothelial viability, tight-junction integrity, CellTiter-Glo 3D viability, live/dead

imaging, Annexin V/propidium iodide flow cytometry, caspase-3/7 activation, and invasion area.

The main targets will be transport with respect to IgG, barrier disruption, apoptotic fraction, and

viability of the spheroids. No less than three independent chip runs will be done, with technical

triplicates when possible. Group comparisons will apply ANOVA or mixed-effects models together

with corrected post hoc testing. The preliminary success thresholds will be set in advance as

transport at least twofold above IgG, barrier-permeability change of less than 15%, at least 60%

reduction in viability, and dual-positive spheroids showing at least a twofold increase in apoptotic

fraction relative to all control groups. In the case of weak transport but high selectivity, the construct

will be kept as a locoregional or redesign candidate rather than being discarded as a failed lead.

 

Conclusion:

The anticipated result is not a developed clinical therapy, but rather a translational screen that

identifies whether the de novo trispecific miniprotein can simultaneously comply with three

conditions that are rarely met by GBM drugs: tumor-selective surface recognition, measurable

BBB-transport potential, and DR5-linked apoptotic killing. One of the beneficial outcomes would

be finding one lead compound that would be able to surpass the previously established transport

and safety thresholds while contributing significantly to the increase in apoptosis in dual-positive

GBM spheroids. A negative transport result would still be informative since it would differentiate

the receptor-binding success from the delivery failure and help in redesigning the molecule.

 

References :

Abramson, J., Adler, J., Dunger, J., Evans, R., Green, T., Pritzel, A., Ronneberger, O., Willmore, L., Ballard, A. J.,

Bambrick, J., Bodenstein, S. W., Evans, D. A., Hung, C.-C., O’Neill, M., Reiman, D., Tunyasuvunakool, K., Wu,

Z., Zemgulyte, A., Arvaniti, E., ... Jumper, J. M. (2024). Accurate structure prediction of biomolecular interactions

with AlphaFold 3. Nature, 630(8016), 493–500. https://doi.org/10.1038/s41586-024-07487-w

Babic, D., Jovcevska, I., & Zottel, A. (2024). B7-H3 in glioblastoma and beyond: Significance and therapeutic

strategies. Frontiers in Immunology, 15, 1495283. https://doi.org/10.3389/fimmu.2024.1495283

Ezzati, S., Salib, S., Balasubramaniam, M., & Aboud, O. (2024). Epidermal growth factor receptor inhibitors in

glioblastoma: Current status and future possibilities. International Journal of Molecular Sciences, 25(4), 2316.

https://doi.org/10.3390/ijms25042316

Dauparas, J., Anishchenko, I., Bennett, N., Bai, H., Ragotte, R. J., Milles, L. F., Wicky, B. I. M., Courbet, A., de Haas,

R. J., Bethel, N., Leung, P. Y., Huddy, T. F., Pellock, S., Tischer, D., Chan, F., Koepnick, B., Nguyen, H., Kang,

A., Sankaran, B., ... Baker, D. (2022). Robust deep learning-based protein sequence design using ProteinMPNN.

Science, 378(6615), 49–56. https://doi.org/10.1126/science.add2187

Hajal, C., Offeddu, G. S., Shin, Y., Zhang, S., Morozova, O., Hickman, D., Knutson, C. G., & Kamm, R. D. (2022).

Engineered human blood-brain barrier microfluidic model for vascular permeability analyses. Nature Protocols,

17(1), 95–128. https://doi.org/10.1038/s41596-021-00635-w

Luo, R., Liu, H., & Cheng, Z. (2022). Protein scaffolds: antibody alternatives for cancer diagnosis and therapy. RSC

Chemical Biology, 3, 830–847. https://doi.org/10.1039/D2CB00094F

O’Halloran, L., Akinsete, O., Kogan, A. L., Wrona, M., & Mahdi, A. F. (2025). 3D in vitro blood-brain barrier models:

recent advances and their role in brain disease research and therapy. Frontiers in Pharmacology, 16, 1637602.

https://doi.org/10.3389/fphar.2025.1637602

Qiao, X., Guo, S., Meng, Z., Gan, H., Wu, Z., Sun, Y., Liu, S., Dou, G., & Gu, R. (2025). Advances in the study of death

receptor 5. Frontiers in Pharmacology, 16, 1549808. https://doi.org/10.3389/fphar.2025.1549808

Rogers, S., Gross, M., Ermis, E., et al. (2024). Re-irradiation for recurrent glioblastoma: a pattern of care analysis.

BMC Neurology, 24, 462. https://doi.org/10.1186/s12883-024-03954-z

Sadowski, K., Jazdzewska, A., Kozlowski, J., Zacny, A., Lorenc, T., & Olejarz, W. (2024). Revolutionizing

glioblastoma treatment: A comprehensive overview of modern therapeutic approaches. International Journal

of Molecular Sciences, 25(11), 5774. https://doi.org/10.3390/ijms25115774

Shi, Y., He, X., Wang, H., Dai, J., Fang, J., He, Y., Chen, X., Hong, Z., & Chai, Y. (2023). Construction of a novel

blood brain barrier-glioma microfluidic chip model: Applications in the evaluation of permeability and anti-glioma

activity of traditional Chinese medicine components. Talanta, 253, 123971. https://doi.org/10.1016/j.

talanta.2022.123971

Straehla, J. P., Hajal, C., Safford, H. C., Offeddu, G. S., Boehnke, N., Dacoba, T. G., Wyckoff, J., Kamm, R. D.,

& Hammond, P. T. (2022). A predictive microfluidic model of human glioblastoma to assess trafficking of

blood-brain barrier-penetrant nanoparticles. Proceedings of the National Academy of Sciences of the United States

of America, 119(23), e2118697119. https://doi.org/10.1073/pnas.2118697119

Thang, M., Mellows, C., Mercer-Smith, A., Nguyen, P., & Hingtgen, S. (2023). Current approaches in enhancing

TRAIL therapies in glioblastoma. Neuro-Oncology Advances, 5(1), vdad047. https://doi.org/10.1093/

noajnl/vdad047

Watson, J. L., Juergens, D., Bennett, N. R., Trippe, B. L., Yim, J., Eisenach, H. E., Ahern, W., Borst, A. J., Ragotte,

R. J., Milles, L. F., Wicky, B. I. M., Hanikel, N., Pellock, S. J., Courbet, A., Sheffler, W., Wang, J., Venkatesh, P.,

Sappington, I., Vazquez Torres, S., ... Baker, D. (2023). De novo design of protein structure and function with

RFdiffusion. Nature, 620(7976), 1089–1100. https://doi.org/10.1038/s41586-023-06415-8