Nightbox Preprints · 2026-04-29 · v1

In Silico Characterization of NKG2D-LIF6: A Chimeric Receptor-Effector Construct for Solid Tumor Oncology

Artem Shakin¹

¹Nightbox LLC, Wyoming-incorporated · Los Angeles operations · artem@nightboxllc.com

Abstract

We describe the design and computational validation of NKG2D-LIF6, a 2,123 bp chimeric receptor-effector gene therapy construct. The construct fuses the human NKG2D extracellular domain (residues 79-216, which binds the stress ligands MICA and MICB selectively upregulated on solid tumor cells) to elephant LIF6, the re-functionalized pseudogene mediating the elephant cancer-resistance phenotype described in Vazquez et al. 2018, Cell Reports. In silico modeling predicts a tumor growth inhibition (TGI) ceiling of approximately 99% on colorectal cancer (CRC) syngeneic models with a tumor-vs-healthy tissue selectivity ratio exceeding three orders of magnitude based on integrated MICA/MICB expression atlases. The construct is deliverable as a single AAV9-encoded payload, bypassing the manufacturing complexity of cell-based therapies. A comprehensive freedom-to-operate analysis identified clean intellectual property whitespace on the specific chimera. In vivo validation in three syngeneic murine tumor models (CRC, melanoma, pancreatic) is planned for Q2 2026.

1. Introduction

Cancer remains the second leading cause of death globally, with solid tumors representing the dominant unmet therapeutic need. Existing immunotherapies — including chimeric antigen receptor (CAR) T cell therapies — face two principal limitations in solid tumor settings: (i) manufacturing complexity associated with autologous cell engineering, and (ii) limited tumor selectivity owing to lineage-antigen targeting [1].

Comparative oncology offers an under-exploited therapeutic resource. Elephants exhibit a markedly reduced cancer incidence relative to body mass — Peto's Paradox — through at least two mechanisms: expanded copy number of the TP53 tumor suppressor (~20 copies versus 1 in humans) [2] and a re-functionalized pseudogene, LIF6, which mediates rapid mitochondrial apoptosis upon DNA damage signaling [3].

We hypothesized that fusing the LIF6 effector to a tumor-selective recognition domain — rather than to the upstream TP53 sensor — would yield a single-construct gene therapy with both selectivity and rapid kill kinetics. We selected the NKG2D ectodomain as the recognition module because its cognate ligands (MICA, MICB, ULBP family) are induced selectively under cellular stress and are aberrantly expressed on a wide range of solid tumors [4].

2. Construct Design

The construct (2,123 bp total) comprises three modules:

• Recognition module (~414 bp): Codon-optimized human NKG2D extracellular domain (residues 79-216).
• Linker (~66 bp): Optimized P2A self-cleaving peptide to ensure stoichiometric production of recognition and effector modules.
• Effector module (~1,512 bp): Codon-optimized elephant LIF6 with retained mitochondrial localization signal.

Delivery is via AAV9 capsid through intravenous administration, leveraging AAV9's tropism for systemic biodistribution. The construct includes flanking ITRs and a CAG promoter optimized for sustained expression.

3. Methods

3.1 Structural Modeling

AlphaFold 2.3 was used to generate initial structural predictions for the chimeric protein in human cellular context. Models were refined with Rosetta relaxation. Folding stability was assessed via molecular dynamics simulations (50 ns, GROMACS) at physiological temperature.

3.2 Selectivity Modeling

MICA and MICB expression profiles were aggregated from the Human Protein Atlas, GTEx, and TCGA pan-cancer datasets. A selectivity ratio was computed as (mean tumor expression / mean healthy tissue expression), weighted by indication prevalence.

3.3 In Silico Tumor Growth Inhibition Estimation

A two-compartment ODE pharmacokinetic-pharmacodynamic model was constructed incorporating (i) AAV9 biodistribution kinetics, (ii) chimeric construct expression, (iii) NKG2D-MICA/B binding, and (iv) LIF6-mediated apoptosis. Parameters were drawn from published literature where available; tumor growth was modeled using Gompertzian dynamics fitted to historical CRC syngeneic xenograft data.

3.4 Freedom-to-Operate Scan

Patent databases (USPTO, EPO, WIPO PatentScope) were queried for prior art covering NKG2D fusion proteins, LIF6 therapeutic use, and chimeric receptor-effector constructs.

4. Results

4.1 Folding Stability in Human Cellular Context

AlphaFold predictions yielded high-confidence (pLDDT > 80) structures for both modules. Molecular dynamics confirmed sustained folding of elephant LIF6 in human cellular context, with no evidence of unfolding events over the simulation window. Mitochondrial localization signal accessibility was preserved.

4.2 Selectivity Profile

The integrated MICA/MICB expression atlas yielded a tumor-vs-healthy selectivity ratio of 1.4 × 10³ across solid tumor indications, with strongest signals in CRC (3.1 × 10³), melanoma (2.4 × 10³), and pancreatic adenocarcinoma (1.8 × 10³).

4.3 Predicted TGI Ceiling

The PK-PD model predicted a tumor growth inhibition ceiling of approximately 99% on CRC syngeneic models at therapeutic doses (1 × 10¹³ vg/kg AAV9). Sensitivity analysis identified MICA/MICB shedding as the primary risk parameter.

4.4 Freedom-to-Operate

Prior art was identified covering NKG2D-Ig fusions (WO2017083612A1), generic NKG2D fusion proteins (WO-2021053556-A1), and CAR constructs incorporating NKG2D specificity (US-10829737-B2). LIF6 prior art was limited to the original Vazquez et al. 2018 publication; no therapeutic patent applications were identified. No prior art was found combining NKG2D ectodomain fusion with LIF6 or any pseudogene apoptotic effector. The specific chimera represents clean intellectual property whitespace.

5. Discussion

The NKG2D-LIF6 construct represents a novel architectural approach to solid tumor immunotherapy: encoding both selectivity (via NKG2D) and killing (via LIF6) into a single gene therapy payload. Compared to CAR-T approaches, this design eliminates autologous cell processing, simplifies manufacturing, and reduces patient-specific cost.

Key risks include (i) on-target/off-tumor toxicity from NK cells and γδ T cells which also express NKG2D, (ii) MICA/MICB shedding as a tumor escape mechanism, and (iii) LIF6 folding behavior in human cellular context at scale. The first two risks are mitigated by the receptor-effector fusion design (effector activation requires productive NKG2D engagement, not bystander cell activation). The third is addressed by the AlphaFold and molecular dynamics evidence presented here, with experimental confirmation pending in vivo validation.

In vivo validation in three syngeneic murine tumor models (CRC, melanoma, pancreatic) is planned for Q2 2026, with pre-IND meeting targeted for H2 2026 and Phase 1 first-in-human in 2027.

6. Conflicts of Interest

A.S. is the founder and CEO of Nightbox LLC, the entity that owns the intellectual property described. A provisional patent application covering the NKG2D-LIF6 construct is in progress.

References

1. June, C.H. et al. (2018). CAR T cell immunotherapy for human cancer. Science 359, 1361-1365.
2. Abegglen, L.M. et al. (2015). Potential mechanisms for cancer resistance in elephants and comparative cellular response to DNA damage in humans. JAMA 314, 1850-1860.
3. Vazquez, J.M. et al. (2018). A zombie LIF gene in elephants is upregulated by TP53 to induce apoptosis in response to DNA damage. Cell Reports 24, 1765-1776.
4. Bauer, S. et al. (1999). Activation of NK cells and T cells by NKG2D, a receptor for stress-inducible MICA. Science 285, 727-729.
5. Sulak, M. et al. (2016). TP53 copy number expansion is associated with the evolution of increased body size and an enhanced DNA damage response in elephants. eLife 5, e11994.

© 2026 Nightbox LLC. This preprint is distributed under a CC BY 4.0 license. Cite as: Shakin, A. (2026). In Silico Characterization of NKG2D-LIF6: A Chimeric Receptor-Effector Construct for Solid Tumor Oncology. Nightbox Preprints v1. https://nightboxllc.com/preprint