FAP-CAR-T cells reduce dystrophic muscle fibrosis, improving adeno-associated virus gene transfer efficacy

Abstract

Tissue fibrosis is a pathological feature of many diseases including muscular dystrophies such as Duchenne muscular dystrophy (DMD). Fibrosis may limit the effectiveness of gene therapy in muscle impacting on viral dosing but direct evidence is lacking. Strategies to reduce skeletal muscle fibrosis are limited. The fibrosis Fap gene is over-expressed in the skeletal muscles of a severe mouse model of DMD, suggesting that cells expressing membrane fibroblast activation protein (FAP) could be targeted by chimeric antigen receptor (CAR)-T cells. Two consecutive administrations of FAP-specific CAR-T cells in the severe DMD model reduced collagen deposits and fibrotic biomarkers and also reduced the number of FAP-positive cells in muscle. Single cell transcriptomics revealed that FAP-CAR-T cells triggered cellular interactions with otherwise inactive muscle resident macrophages and depleted specific subsets of FAP-highly-expressing fibro-adipogenic progenitor cells, pointing to their importance in the fibrosis process. Reducing fibrosis with FAP-CAR-T cells enhanced adeno-associated virus (AAV) microdystrophin gene transfer in the model by increasing vector copies, demonstrating that fibrosis is a restriction factor for AAV gene delivery in skeletal muscle. These results provide novel insights into therapeutic strategies for DMD or other fibrotic diseases.

Graphical abstract

Fibrosis is a pathological feature of many diseases, including Duchenne muscular dystrophy where skeletal muscles exhibit high levels of FAP. In mice, anti-FAP CAR-T cells reduced skeletal muscle fibrosis, eliminating fibrogenic cell subsets. Reducing muscle fibrosis increased the effectiveness of AAV gene therapy and may help to reduce viral dosing.

Introduction

Skeletal muscle fibrosis is a common hallmark of chronic progressive skeletal muscle degenerative disorders, most prominently associated with aging or with muscular dystrophies of genetic origin such as Duchenne muscular dystrophy (DMD).¹ DMD is an X-linked progressive muscular dystrophy caused by mutations in the dystrophin gene. In muscle, fibrosis manifests by the replacement of myofibers with fibroblasts and extracellular matrix (ECM) components, reducing muscle contractility and regeneration potential and function. In addition to worsening the general health of individuals, fibrosis may broadly impede the efficacy of gene therapy approaches in muscle.² In the fibrotic liver, hepatocyte gene delivery with adeno-associated virus (AAV) vectors or lentiviral vectors (LVs) is reduced.³ The effects of muscle fibrosis on AAV gene transfer have not yet been directly assessed.

Fibroblast activation protein (FAP), also known as seprase, is a homodimeric cell surface glycoprotein belonging to the serine protease family. FAP was initially identified as a target for epithelial cancer diagnostic and treatment being highly expressed in stromal fibroblasts of the majority of primary and metastatic epithelial tumors while being absent from normal adult tissues.⁴ Chimeric antigen receptor (CAR)-T cells specific for murine FAP were shown to deplete FAP-expressing stromal cells, reducing tumor growth in mice.⁵ More recently, FAP-specific CAR-T cells were shown to reduce cardiac fibrosis in phenylephrin/angiotensin II-induced model of cardiac fibrosis.^6,7 A CAR-T cell approach for reducing fibrosis in a skeletal muscle disease has never been reported before. The levels of FAP expression in various muscles during the progression of muscular dystrophies such as DMD have not been reported in detail. Whether or not FAP-specific CAR-T cells may function in DMD is unpredictable considering the large territory of the target tissue, and because DMD is associated with significant perturbations of the immune system such as local inflammatory infiltrates in muscle tissues as well as the presence of transforming growth factor β (TGF-β) and of regulatory T lymphocytes (TLs) which may hamper immunotherapeutic approaches.^8,9

This prompted us to test the effects of FAP-specific CAR T cell in a severe mouse model of DMD¹⁰ and to evaluate if the reduction of fibrosis by this approach could improve the efficacy of gene therapy using AAV-mediated transfer of microdystrophin.

Results

Among the existing murine models of DMD, we selected for our study D2.mdx mice bearing the mdx dystrophin mutation on the DBA2/J background carrying the LTBP4 genetic modifier increasing the phenotype severity.¹¹ This model exhibits severe histopathological features including fibrosis and calcifications in the diaphragm (DIA), as well as progressive collagen accumulation in limb muscles (Figures 1A–1C). Gene expression from digital droplet PCR (ddPCR) and RNA sequencing (RNA-seq) analyses revealed that limb muscles and DIA of D2.mdx mice expressed higher levels of fibrotic genes including murine Fap and Col3a compared to age-matched parental DBA2 mice (D2 mice) whereas little expression was found in the heart of D2.mdx mice or, as expected, in the muscles of D2 control mice (Figure 1D). Such differential expression results suggested that the cell surface FAP protein could represent a possible target of fibrosis for CAR T cells in D2.mdx mice, although variable effects might be observed depending on the levels of FAP in the different muscles.

Figure 1.

Fibrosis analysis in D2.mdx mice

(A) Hematoxylin and eosin staining in 3-month-old tibialis anterior (TA), extensor digitorum longus (EDL), diaphragm (DIA), and heart of D2 and D2.mdx mice. Representative images of 2 independent experiments (n = 6 to 7 independent animals). (B) Sirius red staining in TA, EDL, DIA, and heart from D2 and D2.mdx mice, showing collagen fibers in red. Representative images of 2 independent experiments (n = 6–7 independent animals). A–B Image size is shown by a scale bar. (C) Quantification of collagen deposits in DIA, TA, EDL, and heart from D2 and D2.mdx mice (n = 3 for D2 mice and n = 4 to 7 for D2.mdx mice). (D) Quantification of Fap (top) and Col3a (bottom) mRNA expression relative to the P0 gene in the TA and DIA from D2 and D2.mdx mice (n = 3 for D2 mice and n = 4 to 7 for D2.mdx mice). C–D indicate the ANOVA p values above the compared graphs, and error bars represent the standard deviation.

FAP-specific CAR-T cells were generated by transducing pre-activated spleen D2 T cells with a self-inactivated LV. The CAR construct encodes a murine FAP-specific single chain Fv (ScFv) antibody generated from clone 73.3¹² inserted into a third-generation CAR backbone containing CD28 and 41-BB co-stimulatory domains and previously used to treat plasmacytoid leukemia in mice¹³ (Figure 2A). The initial FAP-CAR construct also expressed a truncated CD19 tag, which was used to measure the CAR expression and this CD19 tag was removed in subsequent experiments to increase the vector titer (data not shown). The transduction of CD3 T cells from spleen of D2 mice cells generated an average vector-copy number (VCN) between 0.4 and 0.5 in the cell culture (Figure S1A). It produced about 95% of CD3⁺ T cells in 7 days of culture, with about 30% of CD8⁺ T cells expressing the CAR construct as well as CD62L, CD69, CD27, and CD44 suggesting a central memory phenotype for the majority of CD8⁺ T cells (Figures 2B and S1B). Such cultured cells were specifically cytotoxic in a dose-dependent manner, against murine FAP-expressing cells as shown by cell lysis (Figure 2C) and degranulation assays (Figure 2D) using Fap-transduced 3T3 cells or parental 3T3 cell controls as target cells.

Figure 2.

Generation of FAP-CAR-T cells and functional activity

(A) Schematic representation of the process used to generate FAP-CAR-T cells from D2 spleen T cells using a third-generation lentiviral vector pCCL-EF1a-scFvFAP-CD28-4.1BB-CD3ζ-T2A-ΔCD19-WPRE or pCCL-EF1a-scFvFAP-CD28-4.1BB-CD3ζ-WPRE in which the scFv against mFAP is fused to T cell signaling domains CD28, 4.1BB, and CD3ζ. (B) Flow cytometry assessment in cells cultured for 7 days. CD19 tag was used to identify CAR⁺ cells, and expressions of CD44, CD27, and CD62L to identify CD8 T cell subsets identified as Tn (naive T cells CD62L⁺⁺, CD44⁻), Tcm (central memory T cells, CD62L⁺⁺, CD44⁺⁺), Tem (effector memory T cells, CD62L⁻ CD44⁺⁺). p values were calculated with Student’s t test and are indicated above the compared conditions with error bars representing standard deviations (as in C–D). (C) Cytotoxic effects of cultured CAR-T cells (3 separate experiments) on 3T3-Luc2 cells expressing or not FAP. p values were calculated with two-way ANOVA multiple comparison statistics test. (D) Cytotoxic activity of FAP-CAR T cells was also measured by the detection of CD107a lysosomal-associated membrane protein on the surface of recently degranulated cells after coculture with 3T3-Luc2 cells FAP or not (2 experiments done in triplicate). p values were calculated with Student’s t test.

The intravenous infusion in 3-month-old D2.mdx mice, of 1 × 10⁶ FAP-specific CAR-T cell cultures as a single injection or as 2 injections of 5 × 10⁵ cells each infused 1 week apart, had limited effects on Fap expression in the tibialis anterior (TA) or DIA muscles, although a significant reduction of Col3a expression was observed in the TA but not in the DIA (Figure S2). However, two consecutive injections of 1 × 10⁶ FAP-CAR-T cells, infused one week apart, significantly reduced the expression of Fap and Col3a gene levels 2 weeks after the last CAR T cell injection in the TA (Figure 3A). These 2 injections of 1 × 10⁶ FAP-CAR-T cell each, markedly improved histological features in D2.mdx limb muscles TA, gastrocnemius (GA), and extensor digitorum longus (EDL) muscles as shown by reduced Sirius red stain of collagen fiber deposits compared to controls (e.g., mice treated with polyclonal non-transduced T cells or parental DBA2 mice) (Figures 3B, 3C, and S3B). This demonstrates for the first time that FAP-specific CAR-T cells can reduce fibrosis in dystrophic limb skeletal muscle. The anti-fibrotic effects of FAP-CAR-T cells varied in different muscles and were more pronounced when treatment was administered at an early age of life. In D2.mdx mice, the fibrotic phase begins at the end of the degeneration peak, around 1.5 months of age, rapidly reaching a first plateau at approximately 3 months, and then continues to progress more slowly until 8–9 months of age, when it reaches its maximum. In 2-month-old mice, fibrosis was reduced by FAP-CAR-T cells in the TA, GA, and EDL whereas no effect was seen in the DIA or heart (where no fibrosis was observed at that time) (Figures 3A–3C and S3). Between 3 and 6 months of age, FAP-CAR-T cells only reduced fibrosis in the TA with a marginal effect at 6 months. These kinetic results suggest that the fibrosis process cannot be reverted by the cytolytic effects of FAP-CAR-T cells once installed.

Figure 3.

FAP-CAR-T cells efficiency on muscular fibrosis

(A) mRNA expression of FAP (top) and Col3a (bottom) in the TA (left) and diaphragm (DIA) (right) from D2 (white bars) and D2.mdx mice treated with polyclonal T cells (orange bars) or with FAP-CAR-T cells (green bars) (n = 3 to 9 mice per group). Statistical analysis was done by one-way ANOVA statistic test and p values are represented above the compared conditions with error bars representing standard deviations, as in C–D. (B) Representative images of Sirius red histological sections of TA and DIA, from D2 mice, D2.mdx treated with polyclonal T cells or with FAP-CAR-T cells. The red intensity staining represents collagen fibers (10× objective). (C) Quantification of the collagen deposits in the TA and DIA of mice over the entire cross-sectional area of the muscle section. The results are shown as the average percentage of fibrotic positive regions over the total area of 2 sections per muscle. Each data point represents an individual mouse. (D) Histological staining of FAP⁺ cells in the TA. Image size is shown by a scale bar; (E) Quantification of FAP+ cells per milli Angström (mmA²). Statistical analysis: one-way ANOVA statistic test.

Immunostaining of TA muscle sections confirmed that FAP-CAR-T cells eliminated FAP-expressing cells present in this muscle (Figures 3D and 3E). To complete the analysis, we performed a single-cell RNA sequencing (scRNA-seq) on TA muscles five days after the last FAP-CAR-T cell injection (Figures 4A and 4B). After performing quality controls with Cell Ranger, more than 8,000 cells per condition with 1,817 gene transcripts per cell and 19,719 mean reads per cell were obtained on average. A dimensional reduction clustering analysis of these muscle mononuclear cells identified 20 cell subsets (Table 1; Figure 4A). Each cell subset was identified according to canonical gene expression profiles as described in published data.^{14,15,16,17,18,19,20,21,22,23,24,25} As expected from the dystrophic process ongoing in D2.mdx mice, and as reported in mdx mice,²⁵ fibro-adipogenic progenitor cells (FAP cells), fibroblasts, and immune cells including several populations of macrophages were increased in D2.mdx mice treated with control T cells compared to D2 mice (Figure 4B). The cluster of FAP cells was specifically reduced in the TA muscles of mice injected with FAP-CAR-T cells compared to muscles of mice treated with control T cells (Figures 4A and 4B). The cluster of endothelial cells, which was globally reduced in D2.mdx mice compared to healthy D2 mice, was further reduced by CAR-T cells, probably indirectly as endothelial cells do not express the Fap gene. All immune cell populations, that were already increased in D2.mdx mice compared to healthy controls, were additionally increased after FAP-CAR-T cell injections.

Figure 4.

Single-cell RNA-seq analysis

(A) UMAP (uniform manifold approximation and projection) representation of the final dataset from pooled data from TA and diaphragms of D2 mice or D2.mdx mice treated with polyclonal T cells or with FAP-CAR-T cells. Clusters were colored by manually assigned cell type identifiers based on skeletal muscle gene expression (see Table 1). RBC, red blood cells, aMAC, atypical macrophages. (B) Heatmap of percentage of each cluster from TA of D2 mice, D2.mdx mice treated with T cells and D2.mdx mice treated with FAP-CAR-T cells. (C) UMAP representation of re-clustering of Pdgfra expressing cells (i.e., the annotated FAP cells, fibroblasts, and aMAC clusters) from TAs of the different groups of mice. Clusters were colored by manually assigned cell type identifiers based on skeletal muscle gene expression and FAP cell gene expression. The upper left inset of the figure shows the parentage of Pdgfra-expressing cell subpopulations with FAP cells, aMAC, and fibroblasts origin-clusters. (D) Stacked bar plot showing the percentage of each cluster from the TA of the different groups of mice. (E) Circle plot showing the intercellular communication strength via secreted signaling interaction between major cell types for FAP cells, aMac in D2.mdx mice treated with control T cells or with FAP-CAR-T cells. The lines originating from a cell type indicate ligands connecting to the cell type where the receptors are expressed. The thickness of the line is proportional to the number of unique ligand-receptor interactions, with loops representing autocrine circuits. (F) Heatmap showing the dominant signaling pathways between source cells and target cells. As source cells, we tested FAP cells and aMAC and as target cells we tested neutrophils, M2 macrophages, M1 macrophages, T cells, B cells, dividing immune cells, Ly6c monocytes, Schwann cells, and aMAC cells in D2.mdx mice treated with FAP-CAR-T cells. Values represent the interaction probability between different receptor-ligand pairs within the same signaling pathway. n= 4 mice per group.

Table 1.

List of major genes expressed in different cells leading to identification of clusters

Cell name	Gene symbol
Fibro-adipogenic progenitor cells (FAP cells)	Pdgfra, Dcn, Fibn, Fap, Colec12, Col6a, Cxcl14, Cd248, Osr1, CD34
Atypical macrophages	C1qa, C1qb, C1qc, Smoc2, Cd68, Cd14, Pdgfra, Fap
Fibroblasts	Pdgfra, Dcn, Fibn, Col1a1, Col1a2
Ly6c monocytes	Ptprc, Lyz2, Cd68, Cd14, Spp1, Cd52, Ccr2, Tlr2
M2 macrophages	Ptprc, Lyz2, Cd68, Cd14, C1qc, C1qb, C1qa
M1 macrophages	Ptprc, Lyz2, Cd68, Cd14, Dock2, Pid1, Fcgr3, Csf1r
T cells	Ptprc, Itk, Skap1, Dock2, Cd28, Cd3d
Dividing immune cells	Mki67, Stmn1, Cd68, Cd74, Smc4
B cells	Ptprc, Cd79a, Cd19, IL21r, H2-Eb1
Neutrophils	S100a8, S100a9, Ccr1, Hcar2, Il1b, clec4e
Mesothelial cells	Gpm6a, Efemp1, Upk3b, Sox6, Wt1, Lrrn4
Muscle stem cells	Pax7, Myf5, Vcam1, Sytl2, Fgfr4, Chodl
Myocytes	Ckm, Myl1, Tnnc2, Tnni2, Tnnt3, Mylpf
Smooth muscle cells	Myh11, Acta2, Tagln, Synpo2, Ctnna3, Prkg1, Myl9
Pericytes	Rgs5, Abcc9, Kncj8, Trpc3, Prkg1, Gucy1a2
Nerve cells	Ccl21a, Mmrn1, Il7, Prox1, Pard6g
Schwann cells	Nkain2, Cdh19, S100b, Plp1, Dmd, Mbp
Platelets	F8, Ptprb, Dnase1l3
Red blood cells	Alas2, Hba-a2, Hbb-bt, Snca
Endothelial cells	Pecam1, Mecom, Dach1, Cdh5, Fabp4, Ablim3

The selection of canonical genes is based on literature. The Fap gene is indicated in bold.

To further specify the targeted cell subsets, a reclustering was performed on Pdgfra expressing cells based on uniform manifold approximation and projection (UMAP) projections showing a large cluster including fibroblasts and FAP cells as well as a small population of atypical macrophages (aMACs) sharing markers of resident macrophages as well as FAP cells (Table 1; Figure 4C). Within Pdgfra⁺ cells, the analysis of distinct expression patterns of various cellular marker genes resulted in the identification of 14 clusters (Figures S4A–S4D). This analysis segregated 8 subpopulations of FAP cells (clusters 0,1,2,4,5,6,7, 12, respectively) identified by the most highly expressed genes: Penk⁺, Cxcl14⁺, Thbs4⁺, Socs3⁺, Cmah⁺, Tnfrsf12a⁺, Fmo2⁺, and Kenk2⁺ FAP cells (Figures S4A and S4B). Two types of fibroblasts were identified (clusters 3 and 10), characterized by common Gphn, Cdk8, or Jarid2 expression but different levels of Fap, Pdgfra, and taco-1 (Figure S4C). The small population of aMACs that was observed initially remains as a unique cluster (cluster 13) and expresses markers of hematopoietic origin such as Ptprc (CD45), Mir142hg, Stat4, markers of macrophagic cells such as CD86, C1q, or ApoE as well as markers also found on FAP cells such as Fap, Pdgfra, Smoc2, Col5a1, or Jarid2 (Figures S4A and S4D). FAP-CAR-T cell treatment reduced only the first 2 clusters (0 and 1) which correspond to FAP cells characterized by the highest expression of the Fap gene (Figures 4C, 4D, S4A, and S4B). These 2 clusters of cells represent 43% of the Pdgfra⁺ FAP cells. This analysis confirmed that FAP-CAR-T cells eliminated cells expressing the target Fap gene and showed that the target cells were contained solely in the population of FAP cells.

Intercellular communication networks, through secretion, direct cell-to-cell and ECM contacts were inferred from scRNA-seq data using CellChat,²⁶ (Table 2) providing graphic representations of the effects of FAP-CAR-T cell treatment on the resident muscle cells (Figure 4E). FAP-CAR-T cell treatment modified the secreted signaling interactions. FAP cells engaged in communications with aMAC cells which in turn interacted with all resident muscle cell populations in stark contrast with the limited interactions that aMAC cells have with other cells in control conditions (Figure 4E). Data for direct cell-to-cell and ECM-cell contacts confirmed the existence of FAP cell-aMAC intercellular networks (Figure S5). FAP cells interacted with aMAC cells through periostin, macrophage inhibitory factor (MIF), complement and plasminogen activator (PLAU) while aMAC cells communicated through secreted phosphoprotein 1 (SPP1), CC motif chemokine ligand, MIF, annexin, galectin, and PLAU (Figure 4F). Of note, SPP1 also known as osteopontin, signals to most cell types, in particular, to neutrophils. Thus, FAP-CAR-T cells directly removed specific subsets of FAP⁺ cell subsets which are likely involved in the initiation of the fibrotic process of skeletal muscle, and this indirectly triggered the activation of macrophages and of immune cells which appear to be involved in the resolution of the fibrotic process as well as potentially in the remodeling of the tissue.

Table 2.

List of receptor-ligand genes and pathways tested with CellChat tool for the generation of circle plot cell communication

Ligands	Tgfb1, Tgfb2, Tgfb3, Bmp4, Bmp5, Bmp6, Hbegf, Fgf7, Pdgfa, Cxcl1, Vegfa, Vegfb, Igf1, Ccl5, Ccl3, Ccl9, Ccl7, Ccl8, Cxcl2, Cxcl12, Tnfsf9, Spp1, Mif, Il6, Il34, Fasl, Tnfsf12, Tnfsf14, Nampt, Angptl1, Ptn, Postn, Mdk, Ptn, Postn, C3, Lgals9, Pros1, Plau, Pros1, Sema3d, Sema3c, Sema3e, Sema3f, Anxa1, Gas6, Psap
Receptors	TGFbR1_R2, ACVR1_TGFbR, BMPR1A_ACVR2A, Ackr1, Cxcr2, Cxcr4, Egfr, Fgfr2, Pdgfra, Pdgfrb, Flt1, Kdr, Igf1r, Ccr1, Ccr2, Ccr5, Ackr3, ITGAV_ITGB3, ITGAV_ITGB5, CD74_CD44, Tnfrsf9, Cd44, CD74_CXCR4, IL6R_IL6ST, Csf1r, Fas, Tnfrsf12a, Ltbr, ITGA4_ITGB1, Tlr4, Pirb, Cdh5, Cdh11, Sdc1, Sdc2, Sdc4, Tek, Sdc3, Lrp1, Ncl, C3ar1, ITGAM_ITGB2, P4hb, Mertk, Plaur, P4hb, NRP2_PLXNA4, Fpr1, Axl, Gpr37l1, Havcr2
Pathways	TGFb, BMP, EGF, VGF, PDGF, VEGF, IGF, CCL, CXCL, MIF, IL6, CSF, LIGTH, TWEAK, CD137, SPP1, VISFATIN, NGPTL, ANGPT, MK, PTN, PERIOSTIN, COMPLEMENT, HGF, SEMA3, ANEXIN, GAS, GALECTIN, PROS, PSAP, PLAU

Microdystrophin gene transfer is currently one of the most promising therapeutic avenues in DMD based on using AAVs to transfer a shortened, but still functional, version of dystrophin to fit the AAV vector capacity.²⁷ To determine if the reduction of fibrosis by FAP-CAR-T had an impact on the efficacy of AAV-mediated microdystrophin gene transfer in muscle, these 2 treatments were combined either sequentially or concomitantly in D2.mdx mice. Two consecutive injections of FAP-CAR-T cells prior to the administration of a suboptimal dose (5 × 10¹² VG (viral genome)/kg) of an AAV 9 vector encoding a human microdystrophin under the control of the C5-12 promoter²⁸ led to significantly higher gene transfer in limb muscles compared to control mice treated with control polyclonal T cells (Figures 5A–5G). The level of AAV vector copies was increased by about 2.5-fold (Figure 5B) and levels of dystrophin gene (Figure 5C) and protein expression were increased by 2.5- to 3-fold by the mice FAP-CAR-T cell-treatment (Figures 5D and 5E). This augmentation was confirmed by immunostaining of dystrophin on muscle sections with 2–3 times more dystrophin positive fibers in mice treated with FAP-CAR-T cells than non-treated ones (Figures 5F and 5G). Higher dystrophin expression improved dystrophic features in muscles as shown by the reduction of inflammation, collagen deposits, calcified fibers, and the decrease of CD11b⁺ monocyte infiltrates (Figures 6A and 6B). In contrast, the concomitant treatment of AAV at the time of the second CAR-T cell injection did not produce the gene transfer enhancing effects (Figure S6) further strengthening the hypothesis that fibrosis must be reduced to permit AAV entry into the tissue. This also shows that CAR-T cells do not act through unspecific effects to enhance AAV gene transfer.

Figure 5.

Combined therapy: Gene transfer efficiency

(A) Schematic representation of the in vivo protocol. Two-month-old D2.Mdx mice were injected or not with two consecutive doses (1 × 10⁶) of control T cells or of FAP-CAR-T cells, and after 2 weeks of treatment the mice were injected with a single intravenous dose (5 × 10¹² VG/kg) of rAAV9-microdystrophin (μDystrophin or μDys). Tissues were collected 6 weeks after the AAV injection, 4 animals were included per group and statistical analyses used one-way ANOVA tests with p values indicated above the compared conditions. (B) Quantification of VGCN with ddPCR in TA, EDL, and DIA in DBA2-MDX mice. (C) Gene expression analysis of microdystrophin in TA, EDL, and DIA of DBA2-MDX mice. Results are shown as a histogram of relative abundance of gene expression over P0 normalizer gene. (D) Quantification of microdystrophin expression after capillary western blot with DYS-B antibody in TA, EDL, and DIA at 6 weeks post-injection in mice. Values are represented as normalized area under the curve (AUC). (E) Capillary western blot of protein lysates of TA at 6 weeks post-injection in mice. Microdystrophin are visible around 139 kDa. (F) Representative microdystrophin IHF pictures of the TA, EDL, and DIA at 6 weeks post-injection in mice. Images include a scale bar. (G) Quantification of dystrophin positive myofibers in TA, EDL, and DIA at 6 weeks post-injection in D2.mdx mice. The percentage of dystrophin+ fibers is represented of number of dystrophin positive fibers over fibers positive for laminin.

Figure 6.

Combined therapy: Histology effects

Mice were treated as in Figure 5A. (A) Representative images with scale bars, of hematoxylin and eosine, Sirius red, alizarine red coloration, and CD11b staining in the TA, EDL, and diaphragm of Dba2-mdx mice 6 weeks post-injection. For Sirius red coloration, red-colored regions are highly fibrotic and contain collagen deposits. For alizarine coloration, brown areas are indicated as a region containing calcium deposits. For CD11b staining, red signals indicate monocytes cells. (B) Quantification of the collagen, calcium deposits area and CD11b cells in the TA of mice. The results are shown as the percentage of fibrotic positive or calcium positive regions over the total area of the muscle cut for colorations. The results for CD11b staining are shown as the percentage of CD11b⁺ cells after nuclei segmentation on R software. Statistical p values are indicated above the compared conditions.

Discussion

Our results provide the first proof of concept that immunotherapy with FAP-specific CAR-T cells can reduce the fibrosis in skeletal limb muscle that results from a progressive pathological process like that of DMD. In D2.mdx mice, the fibrotic phase begins at around 1.5 months of age, rapidly reaching a first plateau at approximately 3 months and then continues to progress more slowly until 8–9 months of age, when it reaches its maximum. The anti-fibrosis effects were demonstrated at the biological, molecular, and histological level in several hind muscles; the TA, GA, and EDL of D2.mdx mice. In this model, the anti-fibrosis effects correlated with the dose of FAP-specific CAR T cells administered, and were more pronounced in young mice at 2 months of age compared to older mice. The mechanism of action of FAP-specific CAR-T cells is likely to be the killing of fibrogenic cells in muscles that secondarily prevented or delayed the accumulation of extracellular protein deposits in these tissues. This is supported by the removal of specific Fap-expressing cell subsets following CAR-T cell administration and is in agreement with the known mode of action of CAR- T cells in general, and with the cytotoxic specificity of FAP-specific CAR-T cells. The removal of extracellular protein deposits would not be expected from T cells that are generally not phagocytic cells, explaining why FAP-specific CAR-T cells would have no effect to reduce fibrosis when it is well established in older mice.

Fibrosis is a complex process involving multiple gene pathways and parameters²⁹ and the effects of FAP-CAR-T cells in our model provide novel cellular insights in this process through single-cell transcriptomics analyses. Whereas FAP cells are globally known to be pro-fibrotic, our data suggest that important functional subdivisions exist within these cells. The removal of Penk⁺ FAP and Cxcl14⁺ FAP cells suggests that they play a powerful role in the initiation of the fibrotic process, in agreement with their gene expression signature. The endogenous opioid precursor proenkephalin-A (Penk) gene has been involved in the regulation of hyperproliferative keloid fibroblasts³⁰ and is a biomarker of renal fibrosis.³¹ Cxcl14⁺ FAP cells are reportedly involved in muscle regeneration and enzymatic tissue remodeling.²⁰ The chemokine CXCL14 is also involved in the ATF3/Cxcl14/Jak2 signaling axis in hepatic stellate cells responsible for liver fibrosis.⁹ Other genes expressed by Penk⁺ Cxcl14⁺ FAP cells may also play important roles in fibrosis.

Cells other than FAP cells may also play a role in the fibrosis process, including macrophages. In muscular dystrophies, M1 macrophages are chronically activated and promote muscle injury while the reparative M2 macrophages functions are perturbed and pathologically repurposed to promote fibrosis through various pathways.^32,33,34 Interestingly, the administration of FAP-CAR-T cells specifically triggered an atypical population of macrophages (aMAC cells) interacting with FAP cells and with other immune cells by strongly engaging the periostin and complement pathways already reported in macrophages of mdx mice.²⁵ Such aMAC cells expressed markers of resident tissue macrophages such as Folr2 or C1q as well as low levels of FAP cell genes including Fap and Pdgfra. They interacted with all immune cells, particularly with neutrophils, through the Spp1 (osteopontin) pathway which is known to promote fibrosis in mdx mice.³⁵ These aMACs expressed low levels of the Lgals3 gene and may include or resemble a recently described population of Lgals3 expressing gal3+ macrophages found in dystrophic muscle and implicated in a temporal regulation of muscle fibrosis and regeneration through the Spp1 pathway.²⁵ In our study, the aMAC cells may simply respond to CAR-T cells by eliminating FAP cells or may have specific functions in the control of fibrosis in dystrophic muscle.

Altogether, our data provide cellular insights that may be useful to guide novel therapeutic strategies against fibrosis in addition to those currently being actively explored.¹⁹ They suggest that anti-fibrotic immunotherapy with FAP-specific CAR-T cells should be implemented early on to remove fibrogenic cells before extracellular protein content is too important. They also suggest that specific subsets of FAP cells and of macrophages may play an important role in the fibrogenic process. Further confirmation of these findings is warranted, but is beyond the scope of this study.

The anti-fibrotic effects of FAP-CAR-T cells were mainly observed in the TA, GA, and EDL hind muscles but not in the DIA, suggesting that fibrosis in this muscle is regulated by different pathophysiological mechanisms. Very high levels of Fap and Col3a genes are already expressed in the DIAs of 2 month-old D2.mdx mice and may be difficult to reduce with CAR-T cells. It is also possible that FAP-CAR-T cells do not reach the DIA efficiently linked either to this tissue’s properties (e.g., vascularization) or to the CAR T cell properties. Future modifications in the CAR-T cell administration protocol, including higher doses, repeated administration schedules or perhaps the inclusion of a lympho-depletion protocol as for human CAR-T cells, may increase the efficacy of CAR-T cells to reach a greater territory of muscles.

Microdystrophin gene transfer is currently an approved gene therapy for some DMD patients in several countries.³⁶ The high doses of AAV8 or AAV9 vector used pose a risk of potential occurrence of serious adverse events³⁷; therefore, strategies aiming to reduce AAV dosage would be important to increase the safety of systemically administered AAV gene therapy at high doses. By reducing fibrosis prior to administration of AAV, FAP-CAR-T cell treatment can relieve a block in AAV transduction and therefore could be used to reduce the vector dose. Our data suggest that fibrosis blocks the entry of rAAV into muscle because CAR-T cells increased the delivery of vector copies in muscle having a proportional effect to increase transgene expression. Presumably, the ECM and collagen deposits create a physical barrier limiting access of the particles to the tissue, by masking/perturbing the expression of vector attachment factors or entry receptors; or by sequestration of the particles away from myofibers. Coherent with our findings, hepatocyte gene delivery with systemically administered AAV or lentiviral vectors is also reduced in several models of liver fibrosis induced by various chemicals or by genetic mutations, although differential effects were observed depending on the models.³ Further studies are therefore needed to better understand the mechanisms of rAAV entry in fibrotic muscle tissues.

Overall, as shown here with CAR-T cells in a model of DMD, it appears that interfering with the fibrosis initiation steps may have significant benefits in gene therapy especially when high doses of rAAV vector are needed. We were not able to demonstrate a physiological benefit on muscle strength from the use of FAP-CAR-T cells either alone or in combination with low dose rAAV coding for microdystrophin (data not shown). Such dose of vector was 10 times lower than the dose of AAV9-μDys required to achieve a beneficial effect on muscle strength because our objective was to assess a potential enhancing effect of FAP-CAR-T treatment on AAV gene delivery, an effect that would not be detectable if 100% dystrophin-positive fibers were already obtained. Further study is therefore needed to design effective therapies of DMD combining AAV and FAP-CAR-T cells. However, reducing or avoiding fibrosis is probably a useful strategy to be able to reduce the vector dose in DMD gene therapy and to increase the safety of the approach based on reduced viral dosing effects. Our findings are likely to also improve the efficacy of gene therapy of other dystrophic pathologies involving fibrosis including limb-girdle muscular dystrophies. Preventing the induction of fibrosis by removing fibrogenic cells through CAR-T cells or by other means may also improve a broad range of pathological conditions or to reduce the effects of aging mediated by fibrosis.

Materials and methods

Plasmid constructions

Construction of the anti-FAP CAR plasmid: The pCCL-EF1a-scFvFAP-CD28-4.1BB-CD3ζ-T2A-ΔCD19-WPRE plasmid was constructed by cloning the VH-VL sequences of the anti-mouse FAP antibody 73.3¹² described in WO 2014/055442 in place of the VH-VL scFvCD123 sequences of the third generation CAR reported in a study by Bole-Richard e al.¹³ cloned into a pCCL-EF1a-scFvCD123-CD28-4.1BB-CD3ζ-T2A-ΔCD19-WPRE LV that was provided by A. Galy and F. Garnache-Ottou. A variant pCCL-EF1a-scFvFAP-CD28-4.1BB-CD3ζ-WPRE was constructed by excising the truncated CD19 (ΔCD19) tag from the initially constructed FAP-CAR lentiviral plasmid.

Construction of the FAP plasmid: The pCCL-SFFV-mFAP-WPRE plasmid was generated by inserting the mFAP cDNA (purchased from R&DSystems [RDC2905]) into a pCCL lentiviral plasmid using the SFFV retroviral promoter (kind gift from A. Thrasher, UCL London). For all plasmids, colonies were screened by digestion and correct clones were confirmed by sequencing.

Lentiviral production and generation of stably transduced target and control cell lines

LVs were produced by transient transfection of HEK293T cells using calcium phosphate and 4 plasmids including a transfer plasmid and 3 accessory plasmids (HIV-1 gagpol, HIV-Rev, and VSV-G) and titered as infectious genome (IG)/mL on HCT116 cells using the digital droplet PCR (ddPCR) provirus primers normalized to albumin as previously reported³⁸ (see primers and probes sequences in Table S1). The following transfer plasmids were used: pCCL-EF1a-scFvFAP-CD28-4.1BB-CD3ζ-T2A-ΔCD19-WPRE, pCCL-EF1a-scFvFAP-CD28-4.1BB-CD3ζ-WPRE, pCCL-SFFV-mFAP-WPRE, and pCCL-SFFV-Luc2-WPRE.

Stable cell lines expressing FAP, Luc2, or both transgenes, 3T3-FAP, 3T3-Luc2, and 3T3-FAP-Luc2, were generated by lentiviral transduction of the NIH-3T3 fibroblast cell line (ATCC) using the LV described previously. Transgene expression was confirmed by immunochemistry or bioluminescence assays.

FAP-CAR-T cells preparation and phenotyping

DBA2 primary murine splenic T cells were isolated using the “Pan T cell Negative Selection” kit (Miltenyi Biotec) and transduced with the indicated LV. The cells (1 × 10⁶ cells/well/mL) were cultured in RPMI-1640 supplemented with glutamine, penicillin/streptomycin (antibiotics), 10% fetal bovine serum (FBS), 50 μM beta-mercaptoethanol, in the presence of 50 μg/mL IL-2 + CD3/CD28 activation beads at a ratio of 1:1 (Gibco) (complete medium). After 48 h, the cells were transduced with the CAR-T LV (2 × 10⁷ IG/mL LV on 1 × 10⁶ cells/mL for a multiplicity of infection [MOI] of 20) in the presence of the poloxamer Lentiboost (0.5 mg/mL) (Revvity) to enhance transduction. After overnight incubation, cells were expanded in complete medium for another 4 days. T cells transduction efficacy was determined either by flow cytometry using an anti-human CD19 when using the LV bearing the truncated CD19 tag or by ddPCR to determine the VCN per cell using provirus-specific ddPCR primers in relation to copies of the murine Titin gene (Table S2). Control polyclonal T cells were generated in parallel with the same activation process but without LV transduction.

Phenotyping of CAR T cells was done after 7 days of culture using fluorescent immunostaining with a panel of anti-mouse antibodies against CD3, CD8, CD4, CD62L, CD44, CD69 (see supplemental Table S2), and the viability marker 7AAD, using the CytoFLEX S cytometer (Beckman Coulter). Data were analyzed with the Kaluza software (Beckman Coulter).

Luciferase assay for FAP-CAR-T cell recent cytotoxic activity

3T3-Luc2 control cells and 3T3-FAP-Luc2 target cells (5,000) were plated in a 96-well plate in 100 μL of DMEM medium supplemented with glutamine, antibiotics and 10% FBS. The following day, FAP-CAR-T cells or control non-specific T cells were added at 4 different concentrations in 100 μL of medium containing 5,000, 15,000, 37,500, or 75,000 cells, to co-culture cells and targets for 24 h. At the end of the co-culture, 100 μL of supernatant medium was removed from each well and mixed with 100 μL of luciferin solution (Bright-Glo Luciferase assay, Promega) to measure bioluminescence levels extemporaneously using a luminometer (560 nm) providing relative light units (RLU). Sixplicate wells were averaged and the percent specific lysis was calculated with the following equation: % specific lysis = 100 − (RLU-treated cells × 100)/mean RLU non-treated cells).

Degranulation assay to measure FAP-specific cytotoxicity

Control 3T3 cells and 3T3-FAP target cells (5,000) were plated in a 96-well plate in 100 μL of complete DMEM medium. The following day, 100 μL containing 75,000 FAP-CAR-T cells or control non-specific TL were added to the corresponding wells together with brefeldin A (1/2000) and 20 μL/mL of anti-CD107a. After 6 h incubation at 37°C, cells were washed with PBS-1X and stained with anti-CD8 antibody. After adding the viability marker 7AAD, cells were analyzed by flow cytometry.

Animal care and use

All animal procedures were approved by the National Ethical Committee, C2EA-51 (Evry-Courcouronnes, France), and the French Ministry of Research (MESRI) and received a national agreement number (APAFiS no. 33622 and APAFiS no. 45009). For this entire study, only males were used. The DBA2 (B6; 129S4-DBA2tm1Cpr/J, strain no. 000671) (D2) controls mice were purchased directly from Charles River Laboratories) while the DBA2/mdx (D2.B10-Dmdmdx/J, strain no. 013141) (D2.mdx) mice were bred and obtained from the Center d’Exploration et de Recherche Fonctionnelle Expérimentale (CERFE) in Evry, France. The D2. mdx strain originates from Jackson Laboratories (D2.B10-Dmd^mdx/J, strain no. 013141) and has been bred at CERFE for about 8–10 generations.

D2.mdx mice aged 2, 3, or 6 months were anesthetized and injected via the retro-orbital vein with 5 × 10⁵ or 1 × 10⁶ of control T cells or specific FAP-CAR-T cells resuspended in 100 μL PBS. Mice received a second injection of the same cells one week later. Two weeks after the second cell injection, D2.mdx mice and D2 age-matched control mice were euthanized to collect various muscles for molecular and histological analyses.

For rAAV treatment, D2.mdx mice received 2 consecutive injections of 1 × 10⁶ FAP-CAR-T cells as described previously, and unless indicated otherwise, were infused 2 weeks later with 5.10¹² VG (viral genome)/kg of rAAV9-microdystrophin vector. Control D2.mdx mice were treated similarly only with polyclonal control T cells instead of CAR-T cells. After 6 weeks, mice were euthanized to collect various muscles for molecular and histological analyses.

RNA extraction and gene expression analysis

TA, GA, EDL, DIA and heart collected from mice were directly in RNAlater for total RNA extraction using the RNeasy fibrous tissue kit (QIAGEN) following manufacturer’s recommendations. RNA (1,000 ng) was then reverse transcribed using the Verso cDNA Synthesis kit (Thermo Fisher Scientific, Waltham, MA, USA). For the ddPCR, 1× of ddPCR Supermix for Probes no dUTP (Bio-Rad), 16 ng of complementary DNA and the following primers sets (Table S1) were used: ddPCR Gene Expression Assay:Fap, Mouse (Bio-Rad, 10031252), ddPCR Gene Expression Assay:Col3a1, Mouse (Bio-Rad, 10031252), and ddPCR Gene Expression Assay:Mpz, Mouse (Bio-Rad, 10031255). PCR conditions were 95°C for 10 min + 40 times (94°C for 30 s, 60°C for 1 min) + 98°C for 10 min. Droplets were generated using the droplet generator QX200 (Bio-Rad) and results analyzed using QuantaLife software (Bio-Rad).

Histological and immunostaining analyses

Muscles collected from mice were immediately snap frozen in liquid nitrogen and processed for histological microscopy. Transverse cryosections (8–10 μm) from frozen muscles were air-dried, stained with Sirius red, red alizarin, and hematoxylin and eosin and examined with an Axioscan Z1 automated slide scanner (Carl Zeiss, Oberkochen, Germany) using a Plan APO 10X/0.45 NA objective.

Immunostainings were done by incubating muscle sections overnight at 4°C with the primary antibodies against mouse laminin, dystrophin, CD11b or FAP revealed by a goat secondary antibody conjugated with Alexa Fluor 594 dye (Table S2). The sections were then mounted using DAPI (4',6-diamidino-2-phenylindole) -Fluoromount-G (Southern Biotech, Birmingham, AL, USA) and visualized on an Axioscan Z1 automated slide scanner (Zeiss) with a Plan APO 10X/0.45 NA objective. Image processing and addition of scale bars was done with the Zen Lite software (Zeiss).

Automatic image analysis

From the scanned immunostained muscle sections, the percentage of dystrophin-positive myofibers and the fluorescence marking intensity in myofibers were quantified using automatic image analysis. Myofiber cytoplasmic regions enclosed within the membrane staining (Laminin) were segmented by morphological segmentation after contrast enhancement, artifact filtering (FiJi software 2.0.0-rc/1.52p, Morpholib plugin v 1.4.1), and image size adjustment to capture the membrane regions. Nuclei were detected from the DAPI channel using the local maxima detection. The fluorescence intensity in each object (fibers, fiber membranes, and nuclei) was measured for each channel together with fibers shape and size. Nuclei were associated to their parent fiber using the R software and all fluorescence and shape data were aggregated together. Non-fiber objects were filtered-out based on shape, size, and fluorescence criteria. Positive myofibers for any channel (laminin, microdystrophin, and nuclei) were detected based on the fluorescence distribution of negative control sections to determine the percentage of microdystrophin-positive fibers.

Collagen deposits and calcified fibers on muscle section were quantified using the open-source QuPath Software. Two pixel classifiers were created to train the software on 3 muscle sections to identify representative regions (collagen region), one to delineate the tissue to be analyzed and the other to identify the regions of collagen deposits. The surface area occupied by collagen was quantified in relation to the total muscle section surface area.

Genomic DNA extraction and viral genome copy number analysis

Genomic DNA was extracted from the muscles using a NucleoMag Pathogen kit (Macherey-Nagel, Allentown, PA, USA) according to the manufacturer’s instructions and was purified using the KingFisher Flex purification system (Thermo Fisher Scientific). Droplet digital PCR was performed for the detection of viral genome copy number (VGCN) per diploid genome using AAV_ITR specific primers and to the titin genomic region (Table S1) for normalization.

Capillary western

Muscle proteins were extracted in radio-immunoprecipitation assay (RIPA) buffer supplemented with Protease Inhibitor Cocktail EDTA-free (Roche), quantified by the bicinchoninic acid (BCA) method (Pierce BCA protein assay kit, Invitrogen) and deposited on the Simple Western Jess system (ProteinSimple, Bio-Techne, Minneapolis, MN, USA) using a 12–230 kDa separation module. Microdystrophin detection was performed using the antibody DysB (NCL-DYSB, Leica, 1:20) and chemiluminescence detection was quantified by the Compass software.

Single-cell RNA sequencing studies

Muscle cells were isolated from the TA and DIA tissues using enzymatic dissociation (dispase II [2.4 U/mL], collagenase A [2 μg/mL], Dnase1 [10 ng/mL] in HBSS + BSA 0.2%]) followed by successive filtrations through 100, 70, and 40 μm strainers to remove aggregates. Cell viability was assessed using Trypan blue exclusion, and only samples with a viability greater than 85% were used for sequencing. A target of 20,000 cells per sample was loaded into a microfluidics-based droplet system (10× Genomics Chromium, Pleasanton, CA, USA) to generate single-cell gel beads in emulsion using 10× Genomics Chromium Single Cell 3′ kit in which reverse transcription is performed to synthesize cDNA. This step was followed by amplification, fragmentation, and ligation of Illumina sequencing adaptors. The resulting libraries were then size-selected and quantified using an Agilent Bioanalyzer or similar platform. Libraries were sequenced on an Illumina NovaSeq 6000 platform using paired-end 150 bp reads (Illumina, Inc., San Diego, CA, USA). Sequencing was performed at a depth of around 20,000 reads per cell to achieve sufficient coverage for transcriptomic analysis. Raw data was demultiplexed and aligned to the reference genome (precompiled GRCm39-2024-A, 10× Genomics) using the Cell Ranger pipeline (10× Genomics, v.8.0.1). Sequencing data was processed using Seurat (v.4.0.0) in R software. Cells were filtered based on quality control metrics including the number of genes detected per cell (cells with fewer than 200 genes or more than 2,500 genes were excluded) and the proportion of mitochondrial genes (cells with >5% mitochondrial gene expression were excluded). Additional filtering included the exclusion of cells with low UMI (unique molecular identifier) counts and potential doublets.

Samples were integrated using the harmony package (v.1.2.1) cells were clustered using the FindClusters function, which employs a graph-based clustering approach with the Louvain algorithm. Clustering resolution was set to 0.6 to optimize cluster granularity. To visualize the clusters, UMAP was performed, and cells were projected into a two-dimensional space based on the top PCs. To identify differentially expressed genes between clusters, the FindAllMarkers function in Seurat was used. Genes expressed in more than 10% of any cluster’s cells were considered. Also, a minimal difference of 40% of expressing cells between each cluster and all the others was set to select interesting genes. Finally, a log2 fold change of at least 1 is used to filter-out genes with minor difference in expression. Marker genes for each cluster were annotated based on known biological functions.

Cell-cell communication analysis between ligand and receptor was performed using the CellChat package (V2.1.2) on the harmony integrated dataset.

The mouse CellChatDB was used to investigate the “secreted signaling,” “ECM-receptor,” and “cell-to-cell contact” categories of interaction across identified clusters.

Interactions supported by at least 10 cells are considered for plotting and analysis.²⁶

Statistical analysis

All data were analyzed using GraphPad Prism 10.2.2 software. Parametric tests such as t tests and ANOVA were used for statistical comparison. To compare the two groups, a statistical comparison was performed using an unpaired t test. To compare multiple groups, we used one-way ANOVA with Tukey’s correction for multiple comparison tests. Graphs were generated using Graphpad Prism v.10.2.2 or R version 3.6.2. The figures display the mean ± standard deviation.

Data availability

Requests can be sent to Anne Galy PhD, currently at ART-TG, Inserm 30 rue H. Desbruères, 91100 Corbeil-Essonnes, France (anne.galy@inserm.fr) and Isabelle Richard, Genethon, Integrare research unit, 1 bis rue de l’Internationale 91001 Evry-Courcouronnes, France (richard@genethon.fr).

Acknowledgments

This work was supported by funds from Genethon and from Inserm. We acknowledge the help with RNA-seq by the platform Single Cell Biomarkers UtechS and Florence Jagorel from the Biomics Platform, C2RT at Institut Pasteur (28 rue du docteur Roux, 75015 Paris, France) supported by France Génomique (ANR-10-INBS-09) and IBiSA. We are also grateful to Nathalie Bourg-Alibert and to the Bioexperimentation and Imaging platforms of Genethon for their help with breeding mice, histology, and microscopic image quantification.

Author contributions

M.F. and C.J.R. share first authorship. M.F. has contributed to methodology, investigation, validation, analysis, and writing. C.J.R. has contributed to the study conceptualization, methodology, investigation, and validation. G.C. contributed to methodology, investigation, software, and data analysis. S.F. and V.B. contributed to the investigations. F.G.-O. and E.B.-R. contributed to resources. S.A. contributed to methodology and resources. A.G. and I.R. are senior authors and co-corresponding authors, and both contributed to the study conceptualization, supervision, funding, writing, review and editing, as well as providing resources.

Declaration of interests

A.G., C.J.R., M.F., S.A., and I.R. are inventors of a patent entitled “Immunotherapy of skeletal myopathies using anti-FAP CAR-T cells” published on April 11, 2024, and referenced as WO/2024/074727.