Targeted Delivery – A Holy Grail for CRISPR Gene Editing

Steve Wolk, 14JAN2026

Targeted Delivery.  In simple terms, targeted delivery refers to getting a drug to, and only to, the places in the body where it can provide therapeutic benefit.  In the gene editing world, the various constructs that have been developed for CRISPR gene editing, such as nuclease editing, base editing, and prime editing, work extremely well for editing cells in a dish.  However, we are still not very good at delivering these gene editing machines in a highly specific manner to an organ and a cell type of interest.  This is one of the holy grails of gene editing, since it has the potential to significantly increase potency and reduce toxicity of CRISPR-based therapeutics.

Vehicles for Targeted Delivery.  Due to the success of mRNA/LNP-based vaccines, considerable focus is being placed on using mRNA/LNP technology to deliver the CRISPR machinery, which requires both a guide RNA (gRNA) and CRISPR associated (Cas) protein.  Typically, the gRNA and an mRNA coding for a Cas protein are encapsulated within an LNP.   Currently, work with viral vectors has become more limited due to a number of factors including size limitations of the cargo, potential immune response/limited redosing, and manufacturing reproducibility challenges.  Alternate nonviral delivery (NVD) opportunities under early-stage development include extracellular vesicles, synthetic nanoparticles, and inorganic nanoparticles, as well physical delivery (e.g., inhalation for delivery to lungs, or direct injection into tumors, eyes, brain, or CSF).

Clinical Trials for in vivo CRISPR Therapeutics based on mRNA/LNP.  Because the body naturally funnels LNPs to the liver (Hosseini-Kharat et al., 2025), liver diseases are a logical opportunity to apply this technology.  Numerous CRISPR therapeutics for liver-based treatments are currently in clinical development, including hATTR (Intellia), HAE (Intellia), AATD-1 (Beam), GSD1 (Beam), and FH (Verve/Lilly), though some have been halted due to adverse events or financial limitations.

Extrahepatic Delivery and tLNPs.  In order to get LNPs to places other than the liver, two important advances have been made recently.  The first involves altering the lipid ratios to keep the LNPs in circulation, including compositions that disfavor the endogenous ApoE/LDL-R pathway and evade macrophage uptake (Hosseini-Kharat et al., 2025).  Methods to achieve this include the structure of the ionizable lipid (iLipid), higher PEG lipid levels, lower or no cholesterol, and phospholipid structures and ratios (Su et al., 2024).

The second is the addition of targeting elements to the surface of the LNPs (tLNPs (e.g., Abbrederis et al., 2025)).  These targeting elements are designed to bind to receptors on the surface of cell types of interest, which can facilitate uptake into the cells.  A number of factors must be considered to optimize cellular uptake, including:

• density and nature of the binding moiety (Ab variant, aptamer, peptide, small molecule, etc.) on the tLNP

• cell surface receptor properties such as cell surface density and recycle rates, as well as variations in the receptor properties with different cell-types and different species

• affinity of the epitope-paratope pair (both on- and off-rates)

• aggregation and nonspecific binding of the tLNP

In addition, developing the additional analytics needed to characterize these more complex moieties must also be on the drawing board. 

Capstan (recently acquired by AbbVie) currently has a tLNP in the clinic that targets CD8+ T cells, which can then produce Anti-CD19 CARs for treatment for B-cell mediated autoimmune diseases.

Extrahepatic Delivery via Bispecific tLNPs:  The “or gate” vs. the “and gate”.  One problem to consider is that it is rare to find a receptor that is highly specific for only one cell type.  An idea for the next generation of tLNPs is the addition of a second targeting element, since it is more likely to find a combination of two receptors that is specific to the cell type of interest.  The fundamental problem with this approach is that the presence of two different binding species allows the tLNPs to now bind to and potentially incorporate into cells that have either type of receptor that the targeting elements are designed for.  In digital logic circuitry, this is referred to as an “or gate”, e.g., either answer allows passing through the gate successfully.  What we really want is an “and gate”, meaning both receptors must be present to successfully bind to the cell.

There are a number of potential approaches to create the “and gate” therapeutically.  The first is taking advantage of the concept of avidity (e.g., Erbe et al., 2020).  In this model, the two targeting elements are chosen two cell receptors, and they are designed/selected to have weak affinities for their intended receptors.  Therefore, if only one surface moieties binds to its target receptor, the fast off-rate will reduce uptake efficiency.  However, for cells that contain both target receptors, simultaneous binding to both receptors will result in a stronger K_d, which can facilitate cell-specific uptake.  This approach, of course, would require sufficient receptor densities to create the potential to bind two adjacent receptors simultaneously.  This motif allows, and would almost certainly require, optimization of the individual Kds, linker types and lengths, as well as the targeting element densities and ratios.

A second approach, which may be more powerful, is utilizing two separate LNPs.  One LNP would encapsulate the gRNA only and have one of the targeting elements on the surface.  The second LNP would contain the mRNA for the Cas protein and have the targeting element for the second receptor.  In this model, only cells that have both receptor types and encapsulate both LNP types would get the full CRISPR machinery (gRNA + Cas protein).  In this paradigm, parameters for each binding moiety (e.g., K_d) can be optimized independently to facilitate uptake.  Another advantage of this approach is that the timing and the concentration of two components can be optimized.  For example, when the mRNA and gRNA are delivered simultaneously, the gRNA must “wait around” in the cytoplasm for the mRNA to be translated, and loss of gRNA can occur due to degradation and/or sequestration.  In the two-part model, the tLNP with mRNA can be delivered first, and the timing of delivering the second tLNP with gRNA can be optimized to facilitate formation of the CRISPR enzyme complex.

Potential risks of this approach include the complications of developing a two-component therapeutic from both the manufacturing and regulatory perspectives.

Extrahepatic Delivery: Targeting vs Syphoning.  As a final consideration, the word “targeted” typically implies an active process to achieve an objective.  This would be an accurate description for bone marrow stem cells that have been removed from the body and are then reintroduced into circulation, as in done in ex vivo sickle cell treatments.  Through active chemotaxis processes, based on receptor responses that initiate complex biological cascades within the cell (Liu et al., 2024), and transmigration steps to cross the bone marrow endothelial layers (Kalafati & Chavakis, 2020), these stem cells are able to return highly specifically to the bone marrow compartments.  This is in strong contrast to LNPs, which are “brainless” entities that are acted upon by the body’s endogenous transport and uptake systems.  For the LNPs, the newly designed lipid formulation changes helps keep them out of the liver, but this results in broad distribution, where the lungs and the spleen are next in line but by no means exclusive.  The targeting elements allow interactions with the cell types of interest when the LNPS arrive mainly through passive processes. Therefore, it is more correct to think of tLNPs reaching their target cells as a small percentage of them being “syphoned off” from the endogenous/passive processes.

Extrahepatic Delivery: Hitchhiking.  Based on this idea, a third variation for targeted delivery is the addition of a surface moiety element designed to facilitate transport.  This “hitchhiking” component is designed hijack existing biological pathway to mimic true targeted delivery.  For example, an ionizable lipid can be chosen that binds albumin to reduce uptake in the liver (e.g., Feng et al., 2025), or a third surface element could be added to achieve the same end.

And the Winner Is……?  It will be very exciting to see which of these technical approaches will lead to success in the clinic, or whether another novel approach and/or alternate delivery system is developed that renders these current approaches obsolete.  As is sometimes the case, a spectrum of solutions, where different approaches are best for particular diseases, may be the final outcome.

References:

Abbrederis et al (2025) Trends in Chemistry 7(12), 827-840. Smart lipid nanoparticles: the chemistry driving targeted therapeutics.

Erbe AK et al. (2020) J Immunol, May 1, 2020, 204 (1 Supplement) 91.2.  Specific Targeting of Tumors Through Bispecific SNIPER Antibodies.

Feng Y et al. (2025) Nature Materials 24, 1826–1839. Albumin-recruiting lipid nanoparticle potentiates the safety and efficacy of mRNA vaccines by avoiding liver accumulation.

Hosseini-Kharat et al. (2025) Molecular Therapy: Methods & Clinical Development 33, 1-14.  Why do lipid nanoparticles target the liver? Understanding of biodistribution and liver-specific tropism.

Kalafati L, Chavakis T (2020) Haematologica 105(12), 2700–2701.  Hematopoietic stem and progenitor cells take the route through the bone marrow endothelium.

Liu et al. (2024) Biochem Soc Trans 52(6):2427-2437.  Molecular regulators of chemotaxis in human hematopoietic stem cells.

Su et al. (2024) Nature Communications volume 15, Article number: 5659. Reformulating lipid nanoparticles for organ targeted mRNA accumulation and translation.