Synergistic integration of mRNA-LNP with CAR-engineered immune cells: Pioneering progress in immunotherapy

Abstract

Chimeric antigen receptor T cell (CAR-T) therapy has emerged as a revolutionary approach in the treatment of malignancies. Despite its remarkable successes, this field continues to grapple with challenges such as scalability, safety concerns, limited therapeutic effect, in vivo persistence, and the need for precise control over CAR expression. In the post-pandemic era of COVID-19 vaccine immunization, the application of messenger RNA (mRNA) encapsulated within lipid nanoparticles (LNPs) has recently garnered significant attention as a potential solution to address these challenges. This review delves into the dynamic landscape of mRNA-LNP technology and its potential implications for CAR-engineered immune cell-based immunotherapy.

Graphical abstract

Mei and colleagues demonstrate how mRNA-LNP technology enhances CAR-engineered immune cells, offering a novel approach to overcome challenges in immunotherapy. This review highlights the integration of these advancements, potentially revolutionizing cancer treatment by improving manufacturing limits and therapeutic efficacy.

Introduction

CAR-T therapy, a pioneering immunotherapeutic approach, involves the genetic modification of patient-derived T lymphocytes to express synthetic receptors, enabling targeted recognition and elimination of malignant cells.¹ To date, the US Food and Drug Administration (FDA) and China Food and Drug Administration have granted approvals for six and five CAR-T products respectively, to be marketed for the treatment of refractory and relapsed (r/r) CD19+ B cell lineage leukemia and lymphoma, as well as B cell maturation antigen (BCMA)+ multiple myeloma.^{2,3,4,5,6,7,8,9,10,11,12} This innovative approach has transformed the landscape of cancer treatment by equipping the immune system with the precision and potency to combat hematologic malignancies and solid tumors. Over the years, the success of CAR-T therapy has prompted a surge in the exploration of other CAR-based immunotherapy approaches, extending beyond CAR-T cells. This expansion includes alternate T cell-regulatory T cells (Tregs), γδT, natural killer T (NKT) cells, and mucosal-associated invariant T cells (MAIT), other lymphocyte subsets—NK cells and B cells, and myeloid cells including dendritic cells (DC), monocytes, macrophages (MΦs), and neutrophils, as well as cytokine-induced killer (CIK) cells, etc.^13,14,15 These diverse CAR-based strategies offer promise in addressing various cancer types and immunological challenges.

Notwithstanding the remarkable clinical successes, CAR-T therapy faces challenges such as large-scale production, safety concerns including cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS),¹⁶ therapeutic outcomes that fall short of the desired efficacy,¹⁷ maintaining in vivo persistence,¹⁸ and achieving precise control over CAR expression, driving ongoing research to refine and optimize this groundbreaking therapeutic modality to enhance both its efficacy and safety, especially in solid tumors. In addition, as of December 2023, the FDA has received reports of 22 cases of secondary malignancies of T cell origin following CAR-T therapy, involving five out of six commercially available CAR-T therapies in the United States, all of which are based on lentiviral vectors.¹⁹ Although there is not enough evidence to confirm the exact relationship between CAR-T therapy and secondary T cell malignancies, CAR transgenes were detected within malignant T cell clones in three cases where genetic sequencing was completed,²⁰ suggesting a potential association with the random integration of lentiviral vectors into the T cell genome.

mRNA technology exhibits extensive applicability, enabling the expression of virtually any protein. Through the design of tailored mRNA sequences, efficient synthesis of specific proteins can be achieved.^21,22,23 The inherent instability of mRNA mandates the utilization of a delivery system to shield it from enzymatic degradation, fostering optimal cellular uptake, intracellular release, and subsequent translation into proteins. Various packaging systems, encompassing lipid nanoparticles (LNPs), extracellular vesicles, cells, and biomimetics, have been employed to deliver mRNA.²¹

mRNA encapsulated within LNPs represents an innovative biotechnological approach with profound implications in various medical fields, particularly in the development of vaccines and gene therapies.^24,25,26,27 This cutting-edge technology leverages the efficient delivery of mRNA, providing temporal control over protein expression. The versatility and adaptability of mRNA-LNP platforms hold significant promise in revolutionizing the landscape of personalized medicine, infectious disease control, and cancer immunotherapy.²⁸ During the COVID-19 pandemic, the world witnessed the groundbreaking success of mRNA-LNPs in the development of highly effective vaccines.^29,30 This achievement has captured significant attention and prompted researchers to explore the potential of mRNA-LNP technology as a solution to the ongoing challenges in CAR-based immunotherapy.

This review aims to provide a comprehensive overview of the dynamic landscape of mRNA-LNP technology and its profound implications for CAR-engineered immune cell-based immunotherapy. By examining the intersection of these two cutting-edge advancements, we can explore innovative strategies to address the existing challenges and further enhance the future of immunotherapy.

mRNA design and construction, and mechanisms of lysosomal escape

mRNA and LNPs are two key components of mRNA-LNPs, in which LNPs, serving as carriers, play a pivotal role in the delivery of mRNA-based therapeutics, which typically comprise four essential components: ionizable cationic lipids, phospholipids, cholesterol, and PEGylated lipids.^26,31 The development of LNPs, along with their design, construction, optimization, and modification, has been extensively covered in previous reviews.^26,32 Therefore, this review intends to sidestep this topic and instead focus on the design and modification of CAR mRNA, as well as the mechanisms underlying lysosomal escape.

The successful application of mRNA-LNP technology in CAR therapy hinges on the meticulous design and modification of mRNA to ensure optimal translation efficiency, stability, and minimal immunogenicity. The basic structural units of conventional linear mRNA include the 5′ cap, 5′ untranslated region (UTR), coding sequence (CDS), 3′ UTR, and the poly(A) tail^33,34 (Figure 1). Each of these regions can be precisely engineered to enhance mRNA performance in therapeutic contexts. The 5′ cap is crucial for mRNA stability and translation initiation.³⁵ Modifications, such as the incorporation of an anti-reverse cap analog, can improve mRNA recognition by the translational machinery and protect against exonuclease degradation.³⁶ The 5′ and 3′ UTRs also play critical roles in mRNA stability and translation efficiency. By selecting UTR sequences that interact favorably with the host cell’s translational machinery, or by engineering UTRs to contain specific regulatory elements, the stability and translation of mRNA can be significantly enhanced.³⁷ These modifications are particularly important in the context of CAR-T therapy, where sustained expression of the CAR protein is required for effective treatment.

Figure 1.

Schematic representation of the basic structure of conventional linear mRNA as well as the generational evolution of CAR

The basic structural units of conventional linear mRNA consist of: a 5′ cap structure, a 5′ UTR, the encoding CAR sequence, a 3′ UTR, and a 3′ poly(A) tail. Based on the differences in intracellular domain design and the introduction of cytokines and ligands, CAR molecules can currently be divided into five different generations. The first generation features an antigen-binding domain and a CD3ζ signaling domain, providing T cell activation but lacking co-stimulatory signals, leading to limited persistence. The second generation adds a co-stimulatory domain (e.g., CD28 or 4-1BB), enhancing T cell proliferation and antitumor activity. The third generation incorporates multiple co-stimulatory domains, further boosting T cell function. The fourth generation, known as TRUCKs (T cells redirected for universal cytokine killing), includes genes for cytokine expression (e.g., IL-12), modifying the tumor microenvironment to enhance immune response. The fifth generation integrates a cytokine receptor signaling domain (e.g., IL-2Rβ) that engages the JAK-STAT pathway, combining CAR and cytokine receptor signaling to improve T cell activity and persistence.

The CDS, which encodes the CAR proteins that have evolved through five generations, each improving functionality, efficacy, and safety (Figure 1). The poly(A) tail, typically consisting of 100–250 adenine residues, enhances the stability of mRNA by protecting it from degradation and facilitating the export of mRNA from the nucleus to the cytoplasm.³⁵ Polyadenylation can be fine-tuned to achieve the desired balance between mRNA stability and translation efficiency, further contributing to the sustained expression of the CAR protein.³⁸

During in vitro transcription (IVT), mRNA synthesis requires nucleotides such as adenosine (A), guanosine (G), cytidine (C), and uridine (U). Chemical modifications of these nucleotides, such as the incorporation of pseudouridine (Ψ) or 5-methylcytosine (m5C), can significantly reduce the immunogenicity of the mRNA by preventing activation of pattern recognition receptors such as Toll-like receptors.^39,40,41 This is crucial in systemic administration, where minimizing the immune response to exogenous mRNA is necessary to avoid adverse reactions and ensure the safety and efficacy of the treatment. To further enhance mRNA performance, additional functional elements can be incorporated into the mRNA structure. These may include sequences that enhance mRNA localization, translation, or stability, or that allow for targeted delivery within the host cell. Such modifications are particularly useful in the context of CAR-T cell therapies, where precise control over mRNA behavior can significantly impact therapeutic outcomes.

In summary, the design and construction of mRNA for CAR-immune cell applications involve a series of intricate modifications and optimizations that collectively enhance mRNA stability, reduce immunogenicity, and ensure efficient translation. These advancements are essential for the successful implementation of CAR therapies, where the precise expression of CAR proteins is critical for therapeutic efficacy.

Also, the successful delivery of mRNA to target cells and its subsequent expression depend critically on the ability of the mRNA to escape from the lysosomal system.^42,43 Once internalized by cells through endocytosis, the LNP complex enters the endosomal pathway, where it faces the challenge of traversing the highly acidic environment of the endosomes and lysosomes. Efficient mRNA delivery and expression require the mRNA to avoid degradation within these compartments and to reach the cytoplasm, where translation occurs.

The lipid composition of LNPs plays a critical role in lysosomal escape. Ionizable lipids promote the formation of non-bilayer structures under acidic conditions, which may destabilize the lysosomal membrane and facilitate mRNA release.⁴⁴ Designing LNPs to inhibit the fusion of endosomes with lysosomes, thereby preventing the mRNA from entering the highly degradative lysosomal compartment.^45,46 Certain molecules, such as polyethylenimine, could act as proton sponges, sequestering protons and neutralizing the acidic environment within endosomes, leading to osmotic swelling and rupture of the endosomal membrane, allowing the mRNA to escape into the cytoplasm.^47,48,49

Understanding the mechanisms of mRNA solubilization and lysosomal escape is essential for the design of effective mRNA delivery systems. Ongoing research aims to optimize these processes to improve the therapeutic potential of mRNA-based drugs, such as those used in CAR-T cell therapies.

mRNA-LNP-mediated manufacturing of CAR-immune cells

Traditional lentiviral vector-based CAR-T manufacturing poses safety risks due to irreversible genetic insertion mutations⁵⁰ and has limitations in delivering large gene sequences.^51,52 In addition, its complex, time-consuming, and costly production process complicates large-scale preparation.^53,54 However, these deficiencies may not exist in mRNA-LNPs, a viral-free delivery system (Table 1). To begin with, once the target mRNA is delivered into the cytoplasm, it can be translated to express the desired protein upon endosomal escape, without the need to enter the cell nucleus or integrate into the genome, eliminating the risk of insertional mutations.⁵⁵ Moreover, the expression of mRNA is transient,⁵⁶ implying that, with the passage of time and the gradual expansion of CAR-immune cells, the surface CAR expression will gradually diminish. Therefore, the transient expression of CAR effectively circumvents the toxicity associated with the prolonged and uncontrolled expansion of CAR-immune cells in vivo, such as CRS and ICANS.^16,57 Theoretically, any target protein molecule can be efficiently delivered and expressed via an mRNA-based platform. In addition, the simultaneous expression of more complex CAR elements or other regulatory factors can be achieved through the co-delivery of multiple distinct mRNAs encapsulated in respective LNPs. Last but not least, the production of IVT mRNA is uncomplicated and economically efficient,^58,59 and the preparation of mRNA-LNP is easily scalable and standardizable.⁶⁰ Preclinical and clinical studies of mRNA-LNP-based CAR-immune cell production are summarized below (Table 2).

Table 1.

Comparison between lentiviral vectors and mRNA-LNP

Characteristics	Lentiviral vectors	mRNA-LNP
Genome integration	yes50	no182
Insertion mutation	high risks183	none182
Transfection efficiency	low to moderate14	various/high14
Expression duration	permanent50	transient184
Payload capacity	∼9 kb185	∼6 kb
Regulatory factors	common to co-deliver186,187	easy, by co-encapsulation111

Table 2.

Overview of integration of LNPs with CAR-engineered immune cells in preclinical and clinical studies

Targeting receptor	Delivery payloads	Cell type	In vivo/vitro	Target	Indication	Clinical phase	Ref/NCT
None	ahCD19-4-1BBζ-CAR mRNA	CAR-T	in vivo	hCD19	none	preclinical study	Billingsley et al.64
None	hCD19-4-1BBζ-CAR mRNA	CAR-T	in vivo	hCD19	none	preclinical study	Billingsley et al.63
hCD3 + hCD28	hCD19-4-1BBζ-CAR mRNA	CAR-T	in vitro	hCD19	B cell leukemia	preclinical study	Metzloff et al.70
None	hCD19 Flag tag-CD28ζ-CAR mRNA; hBCMA-4-1BBζ-CAR mRNA	CAR-NK	in vitro	hCD19; hBCMA	B cell leukemia	preclinical study	Golubovskaya et al.62
None	bmGPC3-CD28ζ-CAR mRNA	CAR-NK	in vitro	mGPC3	hepatocellular carcinoma	preclinical study	Shin et al.71
None	hCD19-4-1BBζ-CAR mRNA	CAR-T/MΦ	in vitro	hCD19	none	preclinical study	Ye et al.72
mCD5	mFAP-CD28ζ-CAR mRNA	CAR-T	in vivo	mFAP	cardiac injury	preclinical study	Rurik et al.75
hCD3	iPB7-hCD19-4-1BBζ-CAR + IL6 shRNA pDNA	CAR-T	in vivo	hCD19	B cell lymphoma	preclinical study	Zhou et al.78
mCD3/5/7	mCD19-CD28ζ-CAR mRNA	CAR-T	in vivo	mCD19	none	preclinical study	Billingsley et al.79
SORT molecules	mCD19-4-1BB/CD28ζ-CAR mRNA	CAR-T	in vivo	mCD19	B cell lymphoma	preclinical study	Álvarez-Benedicto et al.92
SORT molecules	mGPC3-CD3ζ-CAR mRNA; Siglec-GΔITIMs mRNA	CAR-MΦ	in vivo	mGPC3	hepatocellular carcinoma	preclinical study	Yang et al.74
CRV peptide	SasA-CAR mRNA; CASP11 siRNA	CAR-MΦ	in vivo	SasA	MRSA in sepsis	preclinical study	Tang et al.97
Oxidized lipids	hCD19-CAR mRNA	CAR-monocytes	in vivo	hCD19	none	preclinical study	Mukalel et al.99
None	hCD19-CD28ζ-CAR mRNA; PD-1 siRNA	CAR-T	in vitro	CD19	none	preclinical study	Hamilton et al.111
None	hBCMA-CAR mRNA	CAR-T	in vitro	BCMA	multiple myeloma	phase 1 (recruiting)	NCT06359509
Unknown	hTROP2-CAR mRNA	CAR-myeloid cells	in vivo	chTROP2	epithelial tumors	phase 1 (recruiting)	NCT05969041
Unknown	hGPC3-CAR mRNA	CAR-myeloid cells	in vivo	hGPC3	hepatocellular carcinoma	phase 1 (recruiting)	NCT06478693
None	CLDN6 mRNA	CAR-T	in vivo	CLDN6	solid tumors	preclinical study	Reinhard et al.137
None	CLDN6 mRNA	CAR-T	in vivo	CLDN6	solid tumors	phase 1 (completed)	Mackensen et al.139

^ahCD19, the “h” in this case refers to human, and the same as follows in Table 2.

^bmGPC3, the “m” in this case refers to mouse, and the same as follows in Table 2.

^chTROP2, human trophoblast cell surface antigen 2.

Ex vivo generation of CAR-immune cells through mRNA-LNPs

Standard LNP components used in vaccines are generally not suitable for efficient transfection of T cells. For example, the formulation of Moderna’s mRNA-LNP-based COVID-19 vaccine (mRNA-1273) includes several key components: heptadecan-9-yl 8-((2-hydroxyethyl) (6-oxo-6-(undecyloxy)hexyl)amino)octanoate (SM-102), 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC), cholesterol, and polyethylene glycol 2000-dimyristoylglycerol (PEG2000-DMG).²⁵ These components are designed to deliver mRNA to antigen-presenting cells (APCs) to elicit an immune response.⁶¹ However, these cell types differ significantly from T cells in terms of biological characteristics and uptake mechanisms, resulting in low transfection efficiency of standard LNPs in T cells. In addition, LNPs are typically designed to target specific endocytic pathways, whereas T cells may rely on different pathways, further affecting transfection efficiency. To enhance T cell transfection efficiency, chemical modifications and targeted delivery optimizations of LNPs are necessary. By modifying the lipid composition of LNPs, such as increasing the proportion of ionizable lipids, interactions with the T cell membrane can be enhanced. Furthermore, modifying the LNP surface with ligands, some sort of targeting molecules, or antibodies specific to T cell receptors can increase LNP binding to T cells, thereby improving CAR expression. In recent years, several studies have emerged focusing on the ex vivo generation of CAR-engineered immune cells by engineering mRNA-LNPs as the delivery platform^62,63,64 (Figure 2B). These investigations represent notable advancements in the application of mRNA-LNP technology for the controlled and efficient synthesis of CAR constructs in immune cells outside the biological milieu.

Figure 2.

Schematic illustration of ex vivo generation of CAR-immune cells through mRNA-LNP

Primary T cells, NK cells, and MΦs are isolated and multiple in vitro, and then transduced by mRNA-LNP encoding CAR to generate CAR-engineered immune cells. Subsequently, they would be transfused back to the patients. CAR-immune cells would specifically recognize and kill tumor cells.

Billingsley et al. screened these LNPs generated from a library of 24 ionizable lipids, and delivered luciferase mRNA to Jurkat cells.⁶⁴ The most effective LNP formulation, C14-4, was selected for delivering CAR mRNA to primary human T cells. Notably, CAR T cells prepared by C14-4 mRNA-LNPs showed obvious killing activity against CD19+ Nalm-6 cells.⁶⁴ This pioneering research underscores the potential of the C14-4 LNP platform for achieving efficient and safe CAR mRNA delivery. Nevertheless, this screening was conducted using a standard excipient molar ratio for mRNA delivery.^65,66 Subsequently, the research team further employed an orthogonal design of experiment methodology to unveil novel formulations of C14-4 LNPs tailored for T cell delivery.⁶³ This involved sequential library screens of C14-4 LNPs with varying excipient molar ratios of C14-4, cholesterol, phospholipid, and lipid-anchored PEG. Through this systematic approach, they identified B10 LNPs as the most potent mRNA delivery platform for T cell engineering, based on their minimal cytotoxicity to T cells and robust killing activity against tumor cells.⁶³

In traditional practices, for engineering T cells under ex vivo conditions, researchers frequently employ magnetic beads covalently coated with human CD3 and CD28 antibodies (CD3/CD28 Dynabeads) to simulate T cell activation.⁶⁷ This entails adding the T cell activation beads to the T cell culture medium. Subsequently, co-incubation proceeds for 24 h to allow T cell activation. Following activation, the activated beads are removed using a magnet, followed by the introduction of delivery vectors carrying CAR genes.^68,69 While effective, this strategy complicates the production process and prolongs timelines significantly. Moreover, the bead retrieval process leads to cellular yield losses. In addition, concerns arise regarding the intracellular uptake of beads by T cells, which cannot be mitigated by magnetic removal, prompting safety apprehensions post CAR-T cell infusion. In light of these challenges, Metzloff et al. proposed a concept wherein fragments of CD3 and CD28 antibodies are directly linked to the surface of LNPs mimicking APCs, obviating the cumbersome steps associated with beads activation, thus enabling streamlined mRNA CAR-T cell production in a single step⁷⁰ (Figure 3). The team utilized a combination of C14-4 ionizable lipid, cholesterol, DOPE, PEGylated lipid, and PEG lipid functionalized with maleimide groups to prepare LNPs, termed as mal-LNPs. Simultaneously, cleaved and reduced CD3 and CD28 antibody fragments (Fab) were incorporated into mal-LNPs, with surface coupling achieved via thiol-maleimide reaction to fabricate LNPs for T cell activation, referred to as aLNPs. They observed that, in the absence of activating magnetic beads, only aLNPs (bearing surface-conjugated maleimide groups + αCD3 antibody + αCD28 antibody) or αCD3-LNP + αCD28-LNP effectively delivered mRNA to primary human T cells, whereas mal-LNPs (solely conjugated with surface maleimide groups) or αCD3-LNP (surface-conjugated with maleimide groups + αCD3 antibody) or αCD28-LNP (surface-conjugated with maleimide groups + αCD28 antibody) failed to efficiently transfect T cells.⁷⁰ They further optimized the ratio of surface αCD3/αCD28 Fab on aLNPs and found that CAR-T cells based on aLNPs of 1:10 ratio exhibited both high transfection efficiency and potent tumor cell cytotoxicity. In a leukemia xenotransplantation model, anti-CD19 CAR T cells produced by 1:10 aLNPs effectively reduced tumor burden and extended survival in mice after three dosing cycles.⁷⁰

Figure 3.

Schematic illustration of ex vivo generation of CAR-T cells through dynabeads and aLNPs

The traditional dynabeads workflow entails adding activating beads, removing beads, and adding gene delivery vectors, taking a total of 48 h and consisting of 3 steps. However, the new workflow only needs to add aLNPs, merely taking 24 h and 1 step. Image reproduced with permission from Metzloff et al.⁷⁰ under a Creative Commons license.

Other mRNA-LNP-based CAR-engineered immune cells have recently received attention. Golubovskaya and his group introduced CAR mRNA-LNP technology designed for efficient transfection of NK cells expanded from primary peripheral blood mononuclear cells, resulting in the generation of functional CAR-NK cells.⁶² Incorporating CD19-CAR mRNA and BCMA-CAR mRNA into LNPs yielded remarkable CAR expression levels of 78% and 95% in NK cells, respectively. Subsequently, BCMA/CD19-CAR-NK cells effectively eliminated target cancer cell lines, accompanied by dose-dependent secretion of IFN-γ and granzyme B in vitro. Notably, both BCMA-CAR-NK and CD19-CAR-NK cells exhibited significantly heightened cytotoxicity, increased IFN-γ production, and enhanced granzyme B secretion compared with normal NK cells. Furthermore, CD19-CAR-NK cells demonstrated a substantial blockade of Nalm-6 tumor growth in vivo after four repeated infusions.⁶² Optimization or alteration of the components of LNPs may modulate the transfection efficiency of mRNA-LNP formulations. Shin et al. incorporated the cationic lipid 1,2-dioleoyl-3-trimethylammonium-propane into LNPs to generate bifunctional LNPs (DLNPs).⁷¹ In contrast to conventional LNPs, DLNPs primarily utilized lectin-mediated endocytosis to facilitate entry into NK cells, thereby enhancing mRNA delivery and overcoming resistance to genetic manipulation in these cells, resulting in a higher CAR transfection rate. In addition, DLNPs have been shown to ameliorate mitochondrial function in NK cells, leading to upregulation of mitochondrial fusion expression, enhanced mitochondrial respiration, and increased secretion levels of cytokines and granzyme B, consequently augmenting cytotoxicity against cancer cells.⁷¹

Another research team found that the use of LNPs 9322-O16B or 76-O17Se for delivering N1mψ-modified mRNA proved more efficient in primary macrophages and CD8+ T lymphocytes compared with Lipofectamine 2000 or MC3.⁷² Consequently, CD19-reactive CAR macrophages (CAR-MΦs) and CD8+ CAR-T lymphocytes exhibited significant efficacy in eliminating B lymphoma in vitro.

In clinical settings, an mRNA-LNP-based BCMA CAR-T cell injection (SYS6020) has been approved for clinical trials (NCT06359509) for the treatment of multiple myeloma, paving the way for the clinical translation of CAR-immune cells prepared ex vivo via mRNA-LNPs, and heralding a promising future.

In vivo manufacturing of CAR-immune cells through mRNA-LNPs

To generate CAR-immune cells in vivo, gene agents are directly injected into patients to target specific cell types and edit them in situ.⁷³ As an off-the-shelf immunotherapy, this advanced technology offers several distinct advantages. Firstly, this approach circumvents the need for labor-intensive and ex vivo manipulation of immune cells, streamlining the therapeutic process, and reducing the waiting time as well as enhancing accessibilities to cellular therapy for patients with advanced cancer. Furthermore, in vivo generation mitigates concerns related to the ex vivo expansion, manipulation, and re-infusion processes, reducing the risk of contamination. Overall, in vivo creation of CAR-engineered immune cells represents a promising and innovative paradigm in cancer immunotherapy. Recently, in situ generation of CAR-T cells by engineered viral vectors has led to interesting achievements in preclinical research.^14,73 The mRNA-LNP delivery system subjected to engineered modification and formulation optimization, is equally unparalleled in this regard and has its unique superiorities.

Rurik et al. were the first to investigate the feasibility of in situ generation of CAR-T cells by mRNA-LNPs (Figure 4A). They fabricated CD5-targeted mRNA-LNPs for the delivery of CAR elements, facilitating in vivo transduction of T cells and the generation of fibroblast activation protein (FAP)-targeting CAR-T cells specifically designed to target activated fibroblasts in a murine model of heart failure.⁷⁵ Their observations revealed that the antifibrotic CAR-T cells exhibited trogocytosis abilities, retaining the target antigen as they accumulated in the spleen. Remarkably, these cells demonstrated a significant reduction in fibrosis and restoration of cardiac function in mice. Notably, the activation of fibroblasts is common in various physiological and pathophysiological processes, such as tissue repair and wound healing.^76,77 This study ingeniously leveraged mRNA-mediated transient CAR expression, effectively circumventing the potential toxic risks associated with the prolonged presence of FAP-reactive CAR-T cells in vivo.

Figure 4.

Schematic illustration of GPC3 CAR and Siglec-GΔITIMs mRNA-laden LNPs enabling in situ generation of Siglec-GΔITIMs-expressing GPC3-targeted CAR-MFs for anti-hepatocellular carcinoma (HCC) immunotherapy

mRNA-LNP was prepared by adding ionizable lipid PPZ-A10, encapsulating GPC3 CAR and Siglec-GΔITIMs mRNA and delivered it to liver MΦs to generate HCC-targeted GPC3-specific CAR-MΦs with Siglec-G lacking ITIMs (Siglec GΔITIMs). Siglec GΔITIMs recognize and competitively bind to CD24 in HCC cells, but could not prime the activation of ITIMs, which overcame the inhibition of the “do not eat me” signaling pathway and significantly enhanced CAR-mediated cytophagocytosis. It also promoted the antigen presentation of CAR-MΦs and activated the adaptive immune system, rendering T cells to secrete killing cytokines, perforin, and granzyme, further enhancing the tumor clearance effect. Image reproduced with permission from Yang et al.⁷⁴ under a Creative Commons license.

Another research team devised a CD3-targeted LNP that encapsulated a plasmid containing a combination gene of interleukin-6 short hairpin RNA (IL-6 shRNA) and CD19-CAR.⁷⁸ Utilizing a CD3 antibody, this system specifically targeted T cells and efficiently transfected them to convert into CAR-T cells with IL-6 knockdown. This innovative approach resulted in the effective elimination of CD19-highly expressed leukemia tumor cells while concurrently mitigating CRS induced by IL-6.⁷⁸ While this study employed LNP delivery of plasmids rather than mRNA, it, in conjunction with the previous investigation, collectively unveils the common modus operandi of conventional in vivo LNP delivery. Most recently, another research team developed LNPs modified via antibody conjugation (Ab-LNPs) to target pan-T cell markers including CD3, CD5, and CD7 (Figure 4B), successfully achieving in situ generation of CAR-T cells.⁷⁹ Upon the delivery of CAR mRNA using these Ab-LNPs, antibody- and dose-dependent CAR expression, as well as cytokine secretion, were observed, with a remarkable B cell clearance efficacy of up to 90%. However, despite the surface antibody-modified LNPs demonstrating an affinity for T cells, a considerable portion, after intravenous injection, still exhibits a pronounced tendency to accumulate primarily in the liver,^78,80,81,82 leading to a remarkable decrease in overall utilization efficiency.

Accordingly, further development of novel and highly efficient mRNA-LNPs with enhanced targeting capabilities and tropism toward specific organs or target tissues can significantly improve their bioavailability. This advancement aims to further elevate the in vivo generation efficiency of edited CAR-immune cells, consequently enhancing therapeutic efficacy and minimizing non-specific biodistribution-induced toxic side effects to the greatest extent possible. Most recently, a novel approach known as selective organ targeting (SORT) has been developed,^83,84,85 enabling nucleic acid delivery to specific organs or tissues. The distinctive feature of SORT-LNP lies in the creative introduction of the “fifth element” into the LNP composition. Traditional LNPs typically consist of four components. SORT-LNPs, however, incorporate a fifth lipid component (SORT molecule), creating a unique lipid formulation design capable of targeted delivery of mRNA or gene editing systems to specific tissues^83,84 (Figure 4C). For instance, the incorporation of permanently cationic lipids results in predominant mRNA expression in the lungs, the addition of permanently anionic lipids facilitates delivery to the spleen, and the inclusion of ionizable amine lipids enhances delivery to the liver.⁸⁴ Correspondingly, SORT LNPs tailored for lung, spleen, and liver targeting were engineered to selectively edit therapeutically relevant cell types, including epithelial cells, endothelial cells, B cells, T cells, and hepatocytes.⁸³ Moreover, SORT LNPs tailored for targeted delivery to other organs such as brain,⁸⁶ pancreas,⁸⁷ lymph nodes,⁸⁸ bone and bone marrow,⁸⁹ and placenta^90,91 have also been meticulously designed and developed.

Álvarez-Benedicto et al. optimized SORT LNPs containing 10% 18:1 PA targeted to the spleen, achieving efficient transfection of T cells in wild-type mice.⁹² These spleen-targeted LNPs successfully delivered CAR-encoding mRNA to T cells in reporter mice and a lymphoreplete B cell lymphoma model, respectively, through intravenous injection without the need for surface modifying with active targeting ligands. Furthermore, in situ CAR-T cells improved overall survival in mice with B cell lymphoma. In addition, locally transfected CAR-T cells reduced tumor metastasis to the liver by enhancing tumor-infiltrating lymphocytes.⁹²

CAR-MΦs may offer advantages over CAR-T, as macrophages infiltrate the tumor microenvironment (TME), downregulating the proportion of tumor-associated macrophages (TAMs), activates adaptive immunity, and has lower non-tumor-targeted toxicity due to limited circulation time.⁹³ TAMs are key in the TME, influencing prognosis and treatment resistance.^94,95,96 Repurposing TAMs into CAR-MΦs enhances their phagocytic and cytotoxic functions against tumor cells, potentially improving outcomes.

Jiang and colleagues engineered SORT LNPs (CAR&Siglec-GΔITIMs LNP) for the co-delivery of CAR mRNA and immunoglobulin-like lectin-G mRNA with truncated sialic acid-binding immunoglobulin-like lectin-G (Siglec-GΔITIMs) to edit hepatic MΦs for hepatocellular carcinoma therapy⁷⁴ (Figure 5). Upon intravenous injection, these LNPs selectively adsorbed plasma proteins, utilizing endogenous targeting to selectively edit hepatic MΦs in the liver. The mRNA nucleic acid drugs were efficiently delivered to the cytoplasm of MΦs at the liver cancer site, enabling the in situ generation of MΦs expressing GPC3-specific CAR and Siglec-GΔITIMs. The combination of CAR-MΦ therapy and Siglec-GΔITIMs-mediated “don’t eat me” signal pathway blockade imparted precise and efficient targeted phagocytic capabilities to MΦs against liver cancer cells. Activated MΦs, acting as antigen-presenting cells, stimulated an adaptive immune response within the tumor, further enhancing the effectiveness of tumor clearance.⁷⁴ To further elucidate the liver MΦ-targeting mechanism of LNPs, the research team conducted a proteomic analysis to decipher the composition of the protein corona on the LNP surface. The results revealed that the enrichment of apolipoproteins (APOs) on the LNP surface imparts its liver-targeting characteristics, and differences in the abundance of APO A-I and E lead to varied uptake fates of LNPs in different liver cells. In addition, compared with DLin-MC3-DMA LNPs, the PPZ-A10 LNPs proposed in this study achieve MΦ targeting by adsorbing more vitronectin and complement components.⁷⁴ This team further explored the feasibility of in situ generation of CAR-MΦs to eliminate methicillin-resistant Staphylococcus aureus (MRSA) in a sepsis mouse model.⁹⁷ They co-encapsulated small interfering RNA (siRNA) targeting caspase-11/caspase-4 (siCASP11) with CAR mRNA targeting S. aureus surface protein A (SasA) into LNPs (CRV/LNP-RNA). CASP11 inhibition prevents mitochondrial reactive oxygen species from eliminating MRSA and mediates intracellular escape.⁹⁸ The MΦ-targeting peptide CRV was conjugated to the surface of the LNPs. Following intravenous injection, CRV/LNP-RNA was able to transfect MΦs in situ, inducing CAR expression and CASP11 knockdown. These engineered CAR-MΦs effectively phagocytosed extracellular MRSA and blocked its intracellular escape mechanisms, thereby eradicating the MRSA.⁹⁷

Figure 5.

Schematic representations of mRNA-LNP delivery mechanisms for CAR-T cell engineering and organ-specific mRNA delivery

(A) Schematic representation of the in vivo generation of transient FAP-CAR-Ts by anti-CD5 antibody-modified mRNA-LNP. Engineered mRNA-LNPs engage specifically with circulating T cells, and subsequently undergo cell endocytosis. Inside the cytoplasm, the LNPs escape from the endosome, undergo degradation, and release the mRNA encoding FAP-CAR proteins. Consequently, the T cells are reprogrammed into CAR-Ts specifically targeting the FAP marker, which is elevated on activated fibroblasts. Image reproduced with permission from Rurik et al.⁷⁵ under a Creative Commons license. (B) Schematic representation of delivery of CAR mRNA delivery via Ab-LNP. LNP constituents in an ethanol phase and CAR mRNA cargo in an aqueous phase are amalgamated through microfluidic blending to produce maleimide-functionalized LNPs (mal-LNPs) featuring maleimide-polyethylene glycol (mal-PEG) on their outer surface. These are then amalgamated with cleaved and reduced antibody fragments, which attach to the mal-LNPs, resulting in the formation of antibody-functionalized LNPs (Ab-LNPs). Image reproduced with permission from Billingsley et al.⁷⁹ under a Creative Commons license. (C) Schematic illustration of delivery of mRNA by SORT-LNPs to various organs. The introduction of a fifth lipid component, known as the SORT molecule, in the design of SORT-LNPs has enabled efficient mRNA delivery to various organs including the mouse brain, lung, liver, spleen, lymph node, pancreas, bone marrow, and placenta.

Regarding CAR-monocytes, engineering modifications are rare both in vitro and in vivo. The Mitchell lab pioneered attempts to develop in situ CAR monocytes.⁹⁹ Drawing inspiration from the affinity of myeloid cells for oxidized lipids, a high-throughput in vivo screening using ionizable lipids and DNA barcoding identified a novel class of oxidized LNPs with innate targeting capabilities for mRNA delivery to monocytes.⁹⁹ The study discovered that a specific ionizable lipid, C14-O2-based LNP, effectively delivered CD19-CAR mRNA to monocytes in situ. Approximately 2% of circulating monocytes expressed CD19-CAR, resulting in a 45% reduction in circulating B cells. Importantly, no CD19-CAR expression was detected in CD3+ T cells or CD19+ B cells.⁹⁹

Within the clinical landscape, regarding in vivo engineering viral vectors, Interius BioTherapeutics has engineered lentiviral vectors designed for in vivo targeting and transduction of CD7+ T cells and NK cells, facilitating the generation of CAR-T and CAR-NK cells to treat CD20+ B cell malignancies (NCT06539338). In addition, Umoja Biopharma announced that the FDA has approved their investigational new drug application for UB-VV111, an in situ generating CD19 CAR-T cell therapy, intended for the treatment of r/r large B cell lymphoma and chronic lymphocytic leukemia (NCT06528301). UB-VV111 features a surface-engineered lentiviral envelope and transgenes encoding CD19 CAR and rapamycin-activated cytokine receptor. With respect to mRNA-LNP in vivo delivery systems, Myeloid Therapeutics has advanced two in vivo CAR-immune cell therapies to phase 1 clinical trial. MT-302 has utilized mRNA-LNPs to express a trophoblast cell surface antigen 2 (TROP2)-targeting CAR specifically in myeloid cells, aimed at treating patients with advanced or metastatic TROP2+ epithelial tumors (NCT06478693). MT-303 has employed GPC3 CAR to arm myeloid cells, enabling them to kill hepatocellular carcinoma cells and elicit adaptive immune responses (NCT05969041). In summary, the successful initiation of these clinical trials underscores the promising potential of mRNA-LNP technology for in vivo CAR-immune cell therapy. This momentum bolsters our confidence in further advancing the clinical translation of these therapies, demonstrating their capability to address a diverse array of malignancies. The continued progress and expansion of these trials will be pivotal in establishing mRNA-LNPs as a cornerstone in the next generation of cancer immunotherapies, offering promise for more effective and targeted treatments.

Delivery of mRNA-LNP-mediated immunomodulatory factors to boost CAR-immune cells

The potential of mRNA-based cancer immunotherapy is immense, and designing mRNA to produce various immunomodulatory factors such as immune checkpoint inhibitors, cytokines, or other functional proteins has contributed to reshaping the TME and restoring immune adaptability.²⁶

Programmed death ligand 1 (PD-L1) has been found to be overexpressed in various cancer tumor cells^100,101,102 and is considered to play a major role in immune evasion and limitation of CAR-T cell efficacy.¹⁰³ Therefore, inhibiting the interaction between programmed cell death protein 1 (PD-1) and PD-L1 can enhance the effectiveness of CAR-T therapy in TMEs with PD-L1 overexpression.^104,105 While combining PD-1 blockade antibodies with CAR-T cell therapy is a feasible strategy, the use of free antibodies may lead to widespread inhibition of the PD-1 signaling pathway, resulting in autoimmune reactions.^106,107 Some studies have relied on viral gene delivery or clustered regularly interspaced short palindromic repeats-CRISPR-associated protein 9 genome editing to alter the genomic suppression of the PD-1 signaling pathway, but this can lead to permanent T cell desensitization.^108,109,110 To achieve a delicate balance between optimal therapeutic effects and minimal adverse reactions, timely and flexible control of the inhibition level of the PD1/PD-L1 pathway is crucial. Hamilton et al. engineered T cells in vitro by co-delivering mRNA encoding CAR and siRNA targeting PD-1 using LNPs, resulting in “super” T cells that exhibit short-term CAR expression while concurrently inhibiting the PD-1 signal, greatly restricting off-target effects¹¹¹ (Figure 6). These researchers indicated that super T cells, after completing their function, could revert to enhanced T cell functionality. The experiment result of PD-1 expression showed a robust and sustained knockdown of PD-1 mediated by siRNA, reaching approximately 60% one day post-transfection. The PD-1 gene knockdown was maintained up to 7 days post-transfection, closely aligning with the duration of CD19 CAR mRNA expression.¹¹¹ This synchronized CAR expression and PD-1 inhibition could achieve immune checkpoint desensitization during the critical window of CAR efficacy, simultaneously reducing the risk of autoimmunity after CAR expression faded. Researchers also discovered an interesting mutual enhancement between co-encapsulated mRNA and siRNA by LNPs. Increasing the siRNA content significantly enhanced mRNA expression to a certain level, with additional siRNA having minimal impact on mRNA expression.¹¹¹ Conversely, mRNA appears to moderately influence siRNA-mediated knockdown. Thus, within certain limits, co-delivery of mRNA and siRNA can mutually enhance the transfection and functionality of both. Moreover, the study suggests that, even in applications where gene silencing is not required, the inclusion of siRNA may benefit LNP-mediated mRNA delivery.¹¹¹ Future research should incorporate a broader range of immune checkpoints to modulate the activity of T cells, NK cells, and macrophages. These checkpoints may be targeted by LNPs to co-deliver corresponding inhibitors and enhance therapeutic efficacy. For instance, blocking cytotoxic T-lymphocyte-associated protein 4 (CTLA-4),¹¹² lymphocyte-activation gene 3 (LAG-3),¹¹³ T cell immunoreceptor with Ig and ITIM domains (TIGIT),¹¹⁴ and T cell immunoglobulin and mucin-domain containing-3 (TIM-3)¹¹⁵ could enhance CAR-T cell activity, and targeting natural killer group 2 member A (NKG2A),¹¹⁶ natural killer group 2 member D (NKG2D),¹¹⁷ and natural killer p46 (NKp46)¹¹⁸ holds promise for improving NK cell efficacy, while modulating the CD47-signal regulatory protein alpha (SIRPα)¹¹⁹ and Siglec10-CD24¹²⁰ axes can augment macrophage function.

Figure 6.

Shematic illustrations of echanisms for co-delivery of PD-1 siRNA by mRNA-LNP and tumor vaccine applications

(A) Schematic illustration of co-delivery of PD-1 siRNA and CD19 CAR mRNA by mRNA-LNP. After being taken into the cell, the siRNA and mRNA were released into the cytoplasm. The PD-1 siRNA formed a complex with RNA-induced silencing complex (RISC), which mediated targeted degradation of PD-1 mRNA. This complex recognized and bound to specific sites on PD-1 mRNA through sequence complementarity with the siRNA. Subsequently, the nuclease within RISC identified and cleaved the bound PD-1 mRNA, leading to its degradation. In addition, the CAR mRNA initiated translation within the ribosomes, resulting in transient expression of CD19 CAR on the cell membrane. (B) Schematic illustration of mRNA-LNP-based tumor vaccines boosting antitumor effects of CAR-T cells. The mRNA vaccine encoding TAAs or tumor epitopes was encapsulated in mRNA-LNP, leading host cells to produce these antigens. Once the mRNA vaccine enters host cells, the cellular biosynthetic machinery translates the mRNA, generating antigenic proteins. These antigenic proteins are recognized and processed into antigenic peptides by the intracellular antigen presentation machinery by APCs. Subsequently, these antigenic peptides are presented on the surface of CD4+ and CD8+ CAR-T cells, activating and priming a rapid and intense immune response against tumor cells.

Cytokines are a class of secreted proteins that can regulate the physiological and biochemical processes of immune cells, acting as either pro-inflammatory or anti-inflammatory factors to influence the function and activity of immune cells.¹²¹ Extensive explorations have validated the role of pro-inflammatory cytokines in enhancing immunotherapy.^122,123 However, systemic injection of cytokines can lead to severe systemic toxic reactions.^124,125 Therefore, local delivery or transient expression of cytokines may achieve therapeutic efficacy while minimizing toxic reactions. Olivera et al. explored a novel gene-engineered cell therapy by electroporating CAR-T cells with mRNA encoding IL-12 and IL-18.¹²⁶ This approach enables the CAR-T cells to autonomously synthesize these two factors and continuously release them into the vicinity of tumor cells, enhancing their anti-tumor capabilities for more effective cancer treatment. Upon transduction with single-chain IL-12 (scIL-12) and IL-18 decoy-resistant variant (DRIL18) mRNAs, the combination of these two cytokines mRNA-modified CARs synergized in ex vivo and in vivo toxicity assays, and further increased the secretion of IFN-γ, leading to the killing of B16-OVA target cells stably transfected to overexpress gp75 through intratumoral delivery.¹²⁶

Therefore, the mRNA-LNP delivery of pro-inflammatory cytokines to arm CAR-immune cells may significantly enhance the invasion and killing of effector cells to solid tumors, and improve the anti-tumor activity and persistence of CAR-immune cells, although there have been no reports of mRNA-LNP delivery cytokines modifying CAR-immune cells. Consequently, between stable expression of CAR and instantaneous and/or local secretion of cytokines, we can achieve a balance of efficacy and toxicity for maximum benefits.

Overexpression or knockdown/knockout of other regulatory factors in immune cells can also enhance their performance or reprogram their immunophenotype. One notable approach to improve T cell function and phenotype is through the use of mRNA-LNPs. For instance, a team has utilized LNP to effectively deliver forkhead box P3 (Foxp3) mRNA to CD4+ T cells, successfully reprogramming them into Tregs,¹²⁷ which facilitated broadening the cell sources of Tregs. Hence, in the future, co-delivery with FoxP3 and CAR mRNA via LNPs may enable the engineering of CAR-Tregs,^128,129 which have attracted much attention in autoimmune diseases.

mRNA-LNP-mediated vaccine delivery to enhance CAR immunotherapy

With the successful introduction of two mRNA-LNP-based COVID-19 vaccines, which have garnered immense success,^29,30,61 the prospect of utilizing mRNA-LNP delivery for enhancing CAR immunotherapy appears highly compelling. This innovative approach, inspired by the achievements in the field of vaccination, holds significant appeal for its potential application in augmenting the efficacy of CAR therapy against cancer. The proven success and safety profile of mRNA-LNP vaccines underscore the feasibility and attractiveness of leveraging similar technology to optimize and advance CAR-based immune interventions, paving the way for a synergistic and impactful fusion of mRNA-LNP and CAR-T technologies in the realm of immunotherapy.

In the treatment of solid tumors, inadequate CAR-T cell trafficking, the compact structure of tumors, and the tumor stromal cells form “physical barriers,” along with the “physiological barriers” created by the suppressive tumor immune microenvironment, significantly impede the infiltration and engraftment of CAR-T cells deep within the tumor.¹³⁰ In addition, the loss of tumor antigens and the heterogeneity of expression in solid tumors result in immune escape and subsequent recurrence of tumor cells in CAR-T therapy.^131,132,133 mRNA-based vaccine delivery technology may help improve the interaction of CAR-T cells with solid tumor cells. mRNA-based vaccine delivery technology may help improve the interaction of CAR-T cells with solid tumor cells. The initial step involves the delivery of mRNA into APCs, where tumor-associated antigens (TAAs) are synthesized and expressed. Subsequently, the target antigens are presented, along with major histocompatibility complexes, on the surface of APCs to CAR-T cells and endogenous lymphocytes. This activation process would finally enhance their killing capabilities to tumor cells^26,134 (Figure 6B).

Ma et al. were the first to evaluate vaccine-boosted CAR-T cells against solid tumors using amphiphile CAR-T ligands (amph-ligands) that trafficked to lymph nodes and anchored on APCs, enhancing CAR-T cell effectiveness.¹³⁵ Amph-ligand vaccines facilitated antigen presentation, activating and proliferating CAR-T cells with heightened tumor-killing abilities. They also explored vaccines to combat antigen heterogeneity, boosting CAR-T cell communication with the host immune system, preventing antigen-negative tumor evasion, and increasing IFN-γ secretion. This approach induced antigen spread, priming host T cells to eradicate tumors and prevent recurrence.¹³⁶ Ugur Sahin and his team developed an mRNA-LNP (termed as CARVac) encoding claudin 6 (CLDN6),¹³⁷ a TAA crucial in the tight junctions of tumor cells.¹³⁸ Upon injection, CARVac transfected DCs, prompting the cell membrane of DCs to express the CLDN6 antigen. Subsequently, DCs presented this antigen to CAR-T cells targeting CLDN6, further activating and promoting the engraftment and expansion of CAR-T cells. This approach enhanced the persistence and more favorable memory phenotype of CAR-T cells in combating solid tumors.¹³⁷ Building on this, they initiated a phase 1/2 first-in-human clinical trial to assess CLDN6 CAR-T cells, both alone and in combination with CARVac, in patients with r/r CLDN6+ solid tumors.¹³⁹ The outcomes revealed an overall response rate (ORR) of 33%, including one complete response; the disease control rate was 67%. Particularly noteworthy was that patients treated at higher dosage levels exhibited a higher response rate (ORR = 57%). Furthermore, the treatment demonstrated manageable toxicity, with 46% of patients experiencing cytokine release syndrome, mostly presenting mild symptoms. Dose-limiting toxicity occurred in two patients at higher dosage levels.¹³⁹ Although the clinical trial demonstrated promising results for CLDN6 CAR-T cells in terms of overall response rate and manageable toxicity, it is noteworthy that there was no significant improvement in antitumor efficacy between the vaccinated and non-vaccinated groups, regardless of dosage. Consequently, optimizing the dosing and administration frequency of CARVac will be crucial in future studies.

Limitations and challenges

Despite the promising potential of mRNA-LNP technology in CAR-engineered immune cells, several limitations and challenges must be addressed to optimize its clinical application, particularly for in vivo use of intravenously injected LNPs.

One significant issue is the generation of anti-polyethylene glycol (PEG) antibodies, which can lead to rapid clearance of LNPs and reduced efficacy. For example, a study assessed the impact of Pfizer’s Comirnaty LNP on inducing anti-PEG responses in a rat model at different times and doses post-administration.¹⁴⁰ The results indicated a dose-dependent induction of anti-PEG IgM after the initial injection of LNP, with higher doses of PEG-LNP inducing more persistent and higher levels of anti-PEG IgM.¹⁴⁰ In addition, in approximately 25% of healthy blood donors, detectable levels of anti-PEG antibodies are present in the bloodstream. Surprisingly, even among patients with no history of PEG therapy, up to 42% exhibit elevated concentrations of anti-PEG antibodies.¹⁴¹ This phenomenon is attributed to the widespread use of products containing PEG or PEG-conjugated compounds in daily life. Due to the transient expression of mRNA drugs, repeated administration appears to be essential to ensure definite curative effects. Therefore, a thorough investigation into the impact of PEG-based mRNA-LNPs on the immune system, and conversely, the potential effects of anti-PEG antibodies in the human immune system on the pharmacokinetics and targeted delivery of mRNA-LNP, may be a crucial topic for advancing the clinical translation of mRNA-LNPs in the future.

In addition, there are limitations in targeting specific cells and tissues, which affects the precision of CAR delivery. Studies have shown that, upon intravenous administration, mRNA-LNPs rapidly bind to plasma proteins, such as ApoE, forming a protein corona. The surface-bound ApoE can interact with the highly expressed low-density lipoprotein receptors on hepatocytes, thereby promoting the hepatic tropism and substantial off-target retention of LNPs.^142,143,144 Therefore, to improve the bioavailability of mRNA therapeutics, reduce their non-specific uptake and expression in the liver, and minimize the potential toxicity, it is crucial to develop a new generation of SORT-based and surface modified LNPs with enhanced targeting specificity.

Multiple dosing presents another challenge, as repeated administration can induce systemic inflammatory responses and cumulative toxicities. Aside from the effects of synthetic RNA nucleotides and lipids, which can act as adjuvants and promote inflammatory conditions,¹⁴⁵ this complicates the therapeutic regimen. The short-term expression of mRNA also poses a problem, necessitating frequent dosing to maintain therapeutic levels, which may not be feasible or safe for all patients. Systemic inflammatory responses following intravenous injections are particularly concerning,^146,147 as they can lead to severe adverse effects and limit the widespread use of this approach. For some settings, long-term persistence of CAR-immune cells seems to be required, which is not currently attainable with mRNA systems. To overcome these limitations, sustained-release mRNA delivery systems, such as microspheres and nanospheres,¹⁴⁸ hydrogels,¹⁴⁹ and polymeric micelles,¹⁵⁰ should be developed to extend the duration of mRNA expression, reducing the need for frequent dosing. Also, mRNA sequence optimization, including the use of modified nucleosides,¹⁵¹ has shown promise in increasing stability and translation efficiency, thereby enhancing therapeutic efficacy. Furthermore, self-amplifying RNA¹⁵² and trans-amplifying RNA¹⁵³ systems are being investigated for their ability to amplify the mRNA signal within the target cells, potentially allowing for lower doses and prolonged expression. Circular RNA is another innovative approach being studied,¹⁵⁴ as its closed-loop structure provides resistance to exonucleases, leading to greater stability and sustained protein production.

By addressing these challenges and continuing to refine mRNA-LNP technology, it is anticipated that the full potential of CAR-engineered immune cells can be realized, broadening their therapeutic application and improving patient outcomes.

Conclusions and future outlooks

This comprehensive review has navigated the cutting-edge landscape of mRNA-LNP applications in CAR-engineered immune cells, unraveling a multifaceted approach to revolutionize immunotherapy. Our exploration encompassed pivotal facets, beginning with the nuanced design and construction of mRNA-LNPs as sophisticated carriers, ensuring precise and efficient mRNA delivery to immune cells. The review delved into the ex vivo generation of CAR-immune cells through mRNA-LNPs, highlighting its pivotal role in engineering immune cells with a focus on therapeutic enhancement. Extending the discussion to in vivo manufacturing, the review showcased the innovative approach of generating CAR-immune cells within their native microenvironment, presenting a paradigm shift in immunotherapeutic strategies. Notably, we examined the versatility of mRNA-LNPs as a delivery vehicle for immunomodulatory factors, underscoring its potential to amplify CAR-immune cells and bolster their anti-tumor activities. In addition, the exploration extended to mRNA-LNP-mediated vaccine delivery, offering a promising avenue for enhancing CAR immunotherapy, drawing parallels with the successes observed in mRNA-based COVID-19 vaccines.

Looking forward, the future of mRNA-LNP applications in CAR immunotherapy holds exciting prospects. Ongoing efforts in optimizing mRNA-LNP formulations will focus on improving stability, targeting specificity, and minimizing immunogenicity. In addition, a series of ionizable cationic lipids suitable for other immune cells including NKTs, CIKs, neutrophils, MAITs, DCs, B cells, and more, should be screened out and the components suitable for the construction of LNPs should be optimized under continuous endeavors, which would further expand the range of cell types of LNP-based CAR-immune cells, thereby increasing the options available for future clinical treatments. Immune cells engineered by CAR technology have demonstrated significant efficacy in a variety of autoimmune diseases, including refractory systemic lupus erythematosus,^155,156,157 immune thrombocytopenia,¹⁵⁸ pemphigus vulgaris,¹⁵⁹ anti-synthetase syndrome, systemic sclerosis,¹⁶⁰ immune-mediated necrotizing myopathy,^160,161 refractory juvenile dermatomyositis,¹⁶² neuromyelitis optica spectrum disorder,¹⁶³ anti-N-methyl-D-aspartate receptor encephalitis,¹⁶⁴ multiple sclerosis,¹⁶⁵ myasthenia gravis (MG)^166,167 and Lambert-Eaton myasthenic syndrome,¹⁶⁶ stiff-person syndrome,¹⁶⁸ rheumatoid arthritis,¹⁶⁹ and type 1 diabetes mellitus,¹⁷⁰ etc. The transient expression of CAR molecules by mRNA-based CAR-immune cells, along with their short-lived nature, may limit their effectiveness in settings requiring long-term persistence of CAR-T cells, particularly in oncological applications where sustained CAR activity is often crucial for durable responses. This transient expression, however, may prove beneficial in non-malignant conditions such as autoimmune diseases and fibrosis, where short-term CAR activity could be sufficient to modulate the immune response or ameliorate tissue damage. For instance, in autoimmune diseases, transient CAR expression might reduce autoreactive immune cells without long-term immunosuppression, thereby minimizing potential side effects. Similarly, in fibrotic conditions, short-term CAR activity could target and remodel fibrotic tissue without prolonged immune activation. Thus, while mRNA-LNP technology may face challenges in tumor settings requiring persistent CAR activity, its application in non-malignant tumor indications represents a promising avenue where transient expression is not only adequate but potentially advantageous. Cartesian Therapeutics has developed an mRNA-based BCMA CAR-T therapy named Descartes-08 and advanced it into a clinical trial for treating MG.¹⁷¹ Fourteen patients with generalized MG received varying doses of Descartes-08. The results demonstrated that three patients achieved complete or near-complete symptom resolution after treatment, with this effect persisting 6 months post-treatment. In addition, two other patients no longer required long-term intravenous immunoglobulin therapy.¹⁷¹ As mentioned in this article, mRNA-LNP-induced CAR-T cells have been explored for treating heart injuries.⁷⁵ Apart from autoimmune diseases and fibrosis, the research of CAR technology in other non-malignant diseases is also high actively, exemplified as infectious diseases and senescence,^172,173,174 where the former includes human immunodeficiency virus,¹⁷⁵ hepatitis viruses,^176,177 human cytomegalovirus,¹⁷⁸ Epstein-Barr virus,¹⁷⁹ aspergillus fumigatus,¹⁸⁰ and S. aureus.¹⁸¹

Therefore, there is a need to further advance the development and application of mRNA-LNP-modified CAR-engineered immune cells in non-oncological diseases and various solid tumors, aiming to benefit a broader range of patients in the future. Also, further refinement of in vivo manufacturing strategies aims to harness the native microenvironments more effectively for CAR-immune cell generation. Expanding the spectrum of target specificities is a key consideration, enabling CAR-immune cells to combat a broader array of diseases with diverse antigen profiles. Ultimately, investigating synergistic combination therapies, such as integrating mRNA-LNP-mediated CAR immunotherapy with other modalities, holds potential to create more potent and versatile strategies for fighting diseases.

As these advancements progress, the translation into clinical settings and rigorous safety profiling become imperative. The marriage of mRNA-LNP technology with CAR-engineered immune cells not only enriches our understanding of immunotherapeutic strategies but also lays the foundation for innovative and personalized cancer treatments, promising a transformative impact on care paradigms for cancer and nonneoplastic diseases.

Acknowledgments

We thank all researchers for their contribution to this review. This work was supported by grants from the National Natural Science Foundation of China (no. 82330005, 82350103, and 82425003), Technology innovation plan key research and development projects of Hubei Province (no. 2023BCB019), the Fundamental Research Support Program of Huazhong University of Science and Technology (no. 5003530166), and the Fundamental Research Funds for the Central Universities, HUST (no. 5003530176). We would also like to acknowledge the group leader for the cited works, as the representative research combining mRNA-LNP technology with engineered immune cells. Some images in this review are reproduced from them, particularly from the Mitchell Lab. All the figures were created by using the website tool BioRender (https://www.biorender.com/, accessed on March 14, 2024).

Author contributions

Conceptualization, H.M.; writing – original draft, Z.C., Y.H., and J.S.; writing – review & editing, H.M. and Y.H.; visualization, Z.C. and J.S.; project administration, H.M.; funding acquisition, H.M. All authors have read and agreed to the published version of the manuscript.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used ChatGPT 3.5 to refine and polish the writing and enhance readability. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.