FO-35

• Molecular Formula: C65H113N3O8
• Molecular Weight: 1064.6 g/mol
• InChI: InChI=1S/C65H113N3O8/c1-10-15-18-20-22-24-32-59(33-25-23-21-19-16-11-2)76-63(72)35-27-26-34-60(69)68(50-29-28-49-67(13-4)14-5)65(64(73)66-48-17-12-3,46-30-36-61(70)74-51-55-38-42-57(43-39-55)53(6)7)47-31-37-62(71)75-52-56-40-44-58(45-41-56)54(8)9/h38,40,57-59H,6,8,10-37,39,41-52H2,1-5,7,9H3,(H,66,73)
• InChIKey: IZGCSXONYIFNNC-UHFFFAOYSA-N

FO-35 for mRNA-LNP Formulation: AI-Discovered Ionizable Cationic Lipid for Extrahepatic Delivery Research

FO-35 is an ionizable cationic lipid (pKa = 7.67) utilized in the formulation of lipid nanoparticles (LNPs) for messenger RNA (mRNA) delivery in vivo . It was identified through an artificial intelligence (AI)-driven deep learning strategy called Lipid Optimization using Neural networks (LiON), which screened 1.6 million lipids in silico to discover structures with predicted nucleic acid delivery potential beyond the liver [1]. The compound has demonstrated local mRNA delivery to mouse muscle and nasal mucosa, and efficient delivery to the lungs of ferrets, a key preclinical model for pulmonary gene therapy [1].

Workflow mRNA LNP formulation for non-viral nucleic acid delivery

Selection Ionizable lipid with reported pKa profile for pH-dependent endosomal escape

Use Context Preclinical pulmonary and tissue-specific mRNA delivery research

• [1] Witten J, Raji I, Manan RS, et al. Artificial intelligence-guided design of lipid nanoparticles for pulmonary gene therapy. Nat Biotechnol. 2024 Dec 10. doi: 10.1038/s41587-024-02490-y. View Source

Why FO-35 Cannot Be Replaced by Off-the-Shelf Ionizable Lipids for Pulmonary and Mucosal mRNA Delivery

Ionizable lipids are the critical functional component of LNPs, dictating endosomal escape efficiency, tissue tropism, and overall transfection potency. While benchmark lipids such as SM-102 and ALC-0315 have been optimized for intramuscular or hepatic delivery [1], they do not inherently confer efficient delivery to the pulmonary epithelium or nasal mucosa. FO-35, in contrast, was specifically identified by a deep learning model to address the challenge of extrahepatic delivery [2]. Substitution with a generic ionizable lipid not validated in these specific anatomical contexts would result in unpredictable and likely reduced transfection in lung and nasal tissues, undermining the validity of experiments designed to probe pulmonary gene therapy or mucosal vaccination. The precise chemical structure of FO-35, a product of a vast AI screen, is not interchangeable with common catalog lipids without a high risk of experimental failure [2][3].

Ionization profile mismatch

Ionizable lipid pKa varies across analogs, directly influencing endosomal escape efficiency; substituting without pKa verification may alter cytoplasmic mRNA release.

Tissue tropism divergence

Structurally similar analog (FO-32) showed distinct in vivo tissue distribution; FO-35’s delivery profile may not be replicated by generic or closely related lipids.

Model-dependent performance

Delivery outcomes may differ across species and administration routes; FO-32 was highlighted in mouse studies, highlighting potential model-specific differences.

• [1] PreciGenome. Targeted Delivery of Lipid Nanoparticles Through Surface Modification. Application Note. View Source
• [2] Witten J, Raji I, Manan RS, et al. Artificial intelligence-guided design of lipid nanoparticles for pulmonary gene therapy. Nat Biotechnol. 2024 Dec 10. doi: 10.1038/s41587-024-02490-y. View Source
• [3] DC Chemicals. FO-35 COA. Product Page. View Source

Quantitative Evidence for FO-35 Selection: Comparative Performance Against Benchmark Lipids and Analogs

Differentiation by pKa: FO-35 vs. FO-32 for Fine-Tuning Endosomal Escape

FO-35 exhibits a higher pKa (7.67) compared to its closest analog FO-32, which has a reported pKa of 6.36 or 6.46 . This difference in acid dissociation constant is critical for LNP function. The pKa of an ionizable lipid determines the pH at which it becomes protonated, a key event for both encapsulating mRNA during LNP formulation at low pH and facilitating endosomal escape after cellular uptake at the acidic pH of the endosome. The higher pKa of FO-35 may contribute to different endosomal escape kinetics or efficiency compared to FO-32, providing a distinct tool for optimizing LNP performance for specific cell types or tissues .

pKa Comparison: FO-35 vs FO-32

Head-to-head

FO-35: pKa 7.67
FO-32: pKa ~6.36–6.46

Reported higher ionization may support endosomal escape interpretation.

Difference ~1.2–1.3 pH units; review assay conditions.

Evidence Dimension	pKa Value
Target Compound Data	7.67
Comparator Or Baseline	FO-32: 6.36 or 6.46
Quantified Difference	~1.21 to 1.31 pH units
Conditions	Measurement conditions not specified in vendor data; values are from product datasheets.

Why This Matters

This quantitative difference in pKa provides a scientific basis for selecting between these two closely related AI-discovered lipids to tune the pH-dependent behavior of LNPs for specific experimental requirements.

Comparative In Vivo Efficacy: Ferret Lung mRNA Delivery by FO-35 vs. State-of-the-Art Lipids

In a direct in vivo study using the ferret model for pulmonary delivery, LNPs formulated with FO-35 efficiently delivered mRNA to ferret lungs. The study's abstract states that 'both FO-32 and FO-35 efficiently delivered mRNA to ferret lungs' [1]. While a specific quantification is not provided in the abstract, this is presented as a key achievement, with a MIT source describing this as 'representing the first published example of mRNA delivery to ferret conducting airways' [2]. This performance is set against a backdrop where FO-32, the analog, matched the state-of-the-art for nebulized mRNA delivery to the mouse lung [1]. The efficiency of FO-35 in the more challenging ferret model differentiates it from many conventional lipids that fail to achieve robust pulmonary transfection.

Ferret Lung mRNA Delivery

Data to verify

Comparable delivery efficiency to FO-32; tdTomato expression confirmed.

Supports ferret model pulmonary delivery endpoint review.

Quantitative reporter data not delineated; model-specific context.

Evidence Dimension	In vivo mRNA delivery to ferret lungs (trachea to alveoli)
Target Compound Data	Efficient delivery reported; described as first published example for ferret conducting airways.
Comparator Or Baseline	Benchmark state-of-the-art LNPs (implied, e.g., SM-102, ALC-0315) not achieving broad ferret lung delivery.
Quantified Difference	Qualitative differentiation: 'efficient delivery' vs. lack of published examples for comparators in this model.
Conditions	Ferret model; LNP formulation with FO-35; tracheal administration.

Why This Matters

This evidence demonstrates that FO-35 possesses functional activity in a highly relevant preclinical model for human lung diseases, a property not shared by all ionizable lipids, making it a critical component for research in pulmonary gene therapy.

• [1] Witten J, Raji I, Manan RS, et al. Artificial intelligence-guided design of lipid nanoparticles for pulmonary gene therapy. Nat Biotechnol. 2024 Dec 10. doi: 10.1038/s41587-024-02490-y. View Source
• [2] MIT Industrial Liaison Program. Artificial Intelligence for State-of-the-Art Gene Therapy: Jacob Witten. Presentation Summary. View Source

Discovery Method: AI-Driven Screening of 1.6 Million Candidates for FO-35 vs. Conventional Rational Design

FO-35 is not a product of traditional, iterative rational design. It was identified through a deep learning strategy (LiON) that evaluated 1.6 million lipid structures in silico [1]. This is a fundamentally different approach compared to the discovery of first-generation ionizable lipids. The LiON model was trained on a dataset of >9,000 LNP activity measurements, enabling it to predict RNA delivery and extrapolate to novel structures divergent from the training set [1]. The identification of FO-35 from a vast chemical space, guided by AI, provides a level of confidence in its unique structural and functional properties that is distinct from a manually selected analog. This differentiates FO-35 as a tool emerging from a state-of-the-art computational discovery paradigm.

Evidence Dimension	Discovery methodology
Target Compound Data	AI-discovered via LiON deep learning model screening 1.6 million candidates.
Comparator Or Baseline	Conventional lipids discovered via experimental screening and/or rational design.
Quantified Difference	Not applicable (methodological difference).
Conditions	In silico screening and subsequent in vivo validation.

Why This Matters

The AI-driven origin of FO-35 signals a new class of ionizable lipids with potentially non-obvious structure-activity relationships, making it a valuable and distinct tool for researchers aiming to push the boundaries of LNP delivery beyond well-trodden chemical space.

• [1] Witten J, Raji I, Manan RS, et al. Artificial intelligence-guided design of lipid nanoparticles for pulmonary gene therapy. Nat Biotechnol. 2024 Dec 10. doi: 10.1038/s41587-024-02490-y. View Source

Comparative Tissue Tropism: Distinct Local Delivery Profile of FO-35

FO-35 enables local mRNA delivery to a specific set of tissues: mouse muscle and nasal mucosa, in addition to ferret lungs [1]. This profile is distinct from that of its closest analog, FO-32, which, while also achieving delivery to ferret lungs, matched the state-of-the-art for nebulized delivery to mouse lung [1]. FO-35's combination of muscle, nasal, and lung tropism suggests a distinct and potentially broader or differently tuned in vivo distribution profile compared to lipids optimized for a single route or organ. This is a key differentiator for researchers designing studies that require multi-tissue or mucosal delivery. The quantitative data on expression levels in these specific tissues are presented in the primary research article [1].

Evidence Dimension	Local mRNA delivery sites in mice
Target Compound Data	Muscle and nasal mucosa.
Comparator Or Baseline	FO-32: Matched state-of-the-art for nebulized delivery to mouse lung.
Quantified Difference	Qualitative difference in primary tropism profile; FO-35 shows nasal and muscle delivery, while FO-32 is highlighted for lung delivery.
Conditions	In vivo mouse models, local administration routes.

Why This Matters

The distinct tissue tropism profile of FO-35 allows researchers to select the most appropriate AI-discovered lipid for their specific application, whether it be pulmonary gene therapy, intramuscular vaccination, or intranasal mucosal immunization.

• [1] Witten J, Raji I, Manan RS, et al. Artificial intelligence-guided design of lipid nanoparticles for pulmonary gene therapy. Nat Biotechnol. 2024 Dec 10. doi: 10.1038/s41587-024-02490-y. View Source

Top 3 Experimental Applications for FO-35 Based on Validated Preclinical Performance

Pulmonary Gene Therapy and Inhalable mRNA Formulations

The primary evidence for FO-35's utility is its efficient delivery of mRNA to ferret lungs, a preclinical model highly relevant to human respiratory diseases [1]. Researchers developing inhalable gene therapies for conditions such as cystic fibrosis, alpha-1 antitrypsin deficiency, or other genetic lung disorders should prioritize FO-35. Its performance in this model differentiates it from lipids that are primarily effective in the liver or at intramuscular injection sites [1].

Intranasal Mucosal Vaccination and Immunization Studies

The validated local mRNA delivery to the nasal mucosa in mice positions FO-35 as a candidate for intranasal vaccine development . For studies aimed at generating a mucosal immune response as a first line of defense against respiratory pathogens, FO-35 offers a demonstrated, albeit qualitative, advantage over lipids not validated in this specific anatomical niche [2].

Intramuscular mRNA Delivery for Comparative Vaccine Research

FO-35 has been shown to effectively deliver mRNA to mouse muscle [2]. This application scenario is not about claiming superiority over clinically validated vaccine lipids like SM-102 or ALC-0315. Instead, FO-35 serves as a powerful research tool for direct, side-by-side comparisons. Researchers can use FO-35 to study how a novel, AI-discovered lipid's intramuscular performance compares to established benchmarks in terms of expression kinetics, immunogenicity, and biodistribution, thereby advancing the fundamental understanding of LNP design [1].

Application

Selection Property

Validation Focus

Pulmonary mRNA delivery research (ferret model)

Reported ionization profile for lung endosomal escape

Ferret lung mRNA expression and biodistribution endpoints

Comparative LNP tropism studies

Structural analog (FO-32) differentiation

Tissue-specific delivery profiles across respiratory sites

Endosomal escape mechanistic studies

Defined pKa for ionization-dependent trafficking assays

Endosomal release kinetics and mRNA cytosolic delivery efficiency

• [1] Witten J, Raji I, Manan RS, et al. Artificial intelligence-guided design of lipid nanoparticles for pulmonary gene therapy. Nat Biotechnol. 2024 Dec 10. doi: 10.1038/s41587-024-02490-y. View Source
• [2] GlpBio. FO-35. Product Page. View Source