AACR 2026 featured thousands of preclinical abstracts, and as usual, the signal-to-noise ratio was challenging. I have selected ten abstracts that I believe represent meaningful advances:
A note on KRAS: the meeting featured over two dozen abstracts spanning inhibitors, degraders, molecular glues, and peptides targeting various KRAS mutants. Rather than selecting the most polished dataset from this crowded space, I prioritized entries with differentiated mechanisms. Two KRAS abstracts made the list; the rest, while often impressive, are difficult to differentiate from summaries alone given that many claim sub-nanomolar potency and in vivo regressions.
Small-Molecule c-Myc Degradation via Non-Native Multimer Stabilization
RDP Pharma (Poster 4611)
c-Myc has been considered undruggable for three decades owing to its intrinsically disordered structure, lacking a stable conformation or binding pocket amenable to conventional small-molecule approaches. RDP Pharma's strategy circumvents this problem by stabilizing a non-native multimeric form of c-Myc that the cell recognizes as misfolded and routes for proteasomal degradation. The group reports oral in vivo activity, which elevates this from a biochemical curiosity to a potentially translatable therapeutic hypothesis. If this platform generalizes to other intrinsically disordered proteins, it would substantially expand the druggable proteome. However, the durability of target suppression in vivo and the selectivity window over wild-type Myc function will be critical questions for clinical advancement.
Coevolution-Guided Molecular Glue Degraders for Pan-KRAS
Poster 5800
The KRAS degrader space is becoming increasingly competitive, with multiple groups reporting PROTAC-, AUTOTAC-, and molecular-glue-mediated degradation of oncogenic KRAS variants at this meeting alone. This abstract distinguishes itself through its discovery methodology: a protein-first, coevolution-guided approach to rationally design E3 ligase-recruiting molecular glues that drive proteasomal degradation across multiple oncogenic KRAS mutants. The platform logic , leveraging evolutionary conservation patterns to identify glue interfaces, is fundamentally different from screening-based approaches and could, in principle, extend well beyond KRAS. That said, molecular glue programs have historically faced steep attrition in translating biochemical degradation into pharmacologically useful in vivo activity, and the absence of disclosed pharmacokinetic or in vivo efficacy data tempers enthusiasm at this stage.
ENA101: A Bispecific T-Cell Engager Targeting a DARKFOX altORF Antigen
Enara Bio (Presentation 4052)
The solid tumor immunotherapy field continues to be constrained by the scarcity of truly tumor-specific antigens. ENA101 attempts to address this by targeting DARKFOX, an antigen derived from an alternative open reading frame, a region of the genome not interrogated by conventional proteomics or standard antigen discovery pipelines. Enara pairs this novel target with a TCR-mimic bispecific T-cell engager, making this a first-in-class combination of target and modality. The broader implication is potentially more significant than the molecule itself: if altORF-derived antigens prove to be broadly expressed in solid tumors while remaining absent from normal tissue, this would open an entirely new target class for T-cell-redirecting therapies. However, the depth of clinical validation for altORF antigens remains limited, and the tumor specificity and expression heterogeneity of DARKFOX will need to be established across a meaningful number of tumor types before the platform thesis can be fully assessed.
Molecular Glue Degradation of HuR in BRAF-Mutant Colorectal Cancer
Presentation 6779
RNA-binding proteins (RBPs) have been pharmacologically inaccessible, despite their well-established roles in post-transcriptional regulation of oncogenic programs. This group presents structural evidence that HuR (ELAVL1), an RBP that stabilizes mRNAs encoding oncogenic and resistance-associated proteins, can be degraded by molecular glues. Critically, they connect HuR degradation mechanistically to altered BRAF splicing, the same splicing events that drive resistance in BRAF-mutant colorectal cancer patients treated with BRAF/MEK combinations. The therapeutic hypothesis is therefore well-defined: degrade HuR, shift BRAF splicing, and resensitize a specific patient population. The group reports near-IND status, suggesting tractable chemistry. While the first-in-class nature of RBP-targeted molecular glues makes clinical translation inherently uncertain, the clarity of the biological rationale is notable.
Coacervate-Based Delivery for Immune Cell Engineering
Poster 6705
Delivery remains the most significant bottleneck for non-viral cell and gene therapy manufacturing. Current approaches , electroporation, lipid nanoparticles, either damage cells or achieve insufficient editing in primary immune populations. This abstract describes a non-viral, endogenous protein-based coacervate delivery platform that reportedly achieves high transfection and CRISPR editing efficiency in primary and hard-to-transfect immune cells. If the efficiency data hold at manufacturing scale, this could materially reduce the cost and complexity of engineered cell therapy production. It is worth noting that many delivery platforms have demonstrated strong in vitro performance that did not translate to scalable manufacturing, so the critical data will come from process development rather than from the biology itself.
Systemic Nanoplasmid Gene Therapy With Cancer-Activated Promoters
Poster 4308
Cytokine-based immunotherapy has long been constrained by the toxicity associated with systemic administration and the limited applicability of intratumoral injection. This abstract presents IV-dosed nanoplasmids engineered with cancer-activated promoters that selectively drive interleukin expression within tumors. The concept , turning tumor cells into localized cytokine factories, is not new, but achieving tumor-selective expression after systemic dosing would represent a genuine step change if validated. The key unknowns are the magnitude of the therapeutic window between intratumoral and systemic cytokine exposure, and whether the promoter selectivity holds across tumor types with varying transcriptional landscapes. Without disclosed quantitative biodistribution or expression data, the strength of this abstract rests primarily on the elegance of the approach.
DNA-Templated, Spatially Controlled PROTACs for Cyclin D1-CDK4/6
Presentation 6735
Conventional degraders target individual proteins. This abstract introduces a DNA-templated approach that spatially coordinates the simultaneous degradation of a multi-protein complex - in this case, Cyclin D1 bound to CDK4/6. By controlling the arrangement of degrader warheads on a DNA scaffold, the system can engage protein-protein interfaces that are inaccessible to single-molecule PROTACs. This is an early-stage platform with clear translational hurdles, including the pharmacokinetic challenges inherent to DNA-based therapeutics. However, the intellectual contribution is significant: if the concept of complex-level degradation proves pharmacologically viable, it would extend the degrader modality into a space that current approaches cannot reach.
A Macrophage PD-1 Endocytic Checkpoint Governing Durable vs. Hyperprogressive Response
Poster 2797
Hyperprogression on anti-PD-1 therapy remains one of the most troubling and poorly understood phenomena in immuno-oncology, affecting an estimated 10-15% of treated patients. This group identifies a macrophage-intrinsic endocytic checkpoint involving CCDC88A/GIV that mechanistically separates durable responders from patients who hyperprogress. The implications are twofold: first, CCDC88A/GIV could serve as a patient stratification biomarker, which would have immediate clinical utility in treatment selection; second, it identifies a new druggable node in the PD-1 axis that resides in macrophages rather than T cells, potentially reframing anti-PD-1 biology beyond its current T-cell-centric paradigm. This is among the most clinically consequential basic science abstract on this list, though the validation of CCDC88A/GIV as a predictive biomarker will require prospective clinical data.
Engineered Probiotic Living Drug for Metastatic Pancreatic Cancer
Cold Spring Harbor Laboratory / Columbia University (Poster 5839)
Metastatic pancreatic ductal adenocarcinoma (PDAC) remains among the most treatment-resistant cancers, with a five-year survival rate below 5%. The KPC mouse model faithfully recapitulates this resistance and is notoriously difficult to treat in preclinical studies. This abstract reports that an engineered E. coli Nissle probiotic, delivering a tumor-selective pore-forming toxin, produced single-dose complete regressions and a survival benefit in KPC metastatic models. These are striking results for any modality in this setting. The living-drug approach is orthogonal to existing PDAC therapeutic strategies, which is precisely what makes it interesting - however, the translation of engineered bacterial therapies to the clinic has historically been challenging, with colonization efficiency, immunogenicity, and manufacturing consistency representing significant hurdles. The data warrant close attention, though clinical validation will be the critical test.
Oral Macrocyclic Peptide KRAS Inhibitor Entering Phase 1
Chugai Pharmaceutical (Poster 5138)
Macrocyclic peptides that are simultaneously orally bioavailable, cell-permeable, and capable of picomolar intracellular target engagement have been a long-standing aspiration of medicinal chemistry, with most prior attempts failing on ADME properties. Chugai reports a macrocyclic peptide that achieves KRAS isoform selectivity through recognition of a single amino acid difference (His95) and is entering Phase I. The selectivity mechanism is notable for its simplicity, and the progression to clinical testing suggests the pharmacokinetic challenges that typically eliminate peptide candidates have been adequately addressed. This is the only abstract on this list with a near-term clinical readout, which will provide a direct test of the intracellular peptide modality in an oncology setting.
Conclusion
Several themes emerge from these selections. The definition of "undruggable" continues to contract , c-Myc, HuR, altORF antigens, and multi-protein complexes all had first-in-class preclinical data at this meeting. Delivery and modality innovation, coacervates, nanoplasmids, engineered probiotics, DNA-templated degraders - are advancing alongside target biology, though many of these platforms will face their most difficult tests in clinical translation rather than in preclinical models. The KRAS space, meanwhile, has matured to the point where the relevant question is no longer whether KRAS can be targeted, but which of the many competing approaches will differentiate in the clinic. Based on what was shown at AACR 2026, the answer will likely depend as much on modality selection and trial design as on preclinical potency.
By Sakis Paliouras, PhD
