Biology subjects

Serafini, M.

Publications and source records attributed to Serafini, M..

7 recordsLinked to original sources

bioRxiv2026

A humanized ossicle platform for real-time tracking and preclinical assessment of CAR-based immunotherapies in acute myeloid leukemia

Preclinical evaluation of chimeric antigen receptor (CAR)-T therapies for acute myeloid leukemia (AML) is limited by the lack of models that faithfully recapitulate the human bone marrow (BM) niche. Here, we implemented a humanized ossicle-based AML model that enables simultaneous engraftment of leukemic blasts and longitudinal assessment of responses to CAR-based immunotherapies. Intravenous or intra-ossicle injection of AML blasts produced robust, reproducible disease mimicking features of human AML within its microenvironment. To monitor tumor burden and immune effector cells in real-time, we developed a dual bioluminescence system using distinct luciferases in AML and CAR-T cells. This approach allowed non-invasive longitudinal tracking of CAR-T cell localization, expansion, persistence, and leukemic clearance within the ossicle. Overall, our model provides a powerful platform to study CAR-T cell behavior within a human BM niche and, for the first time, allows simultaneous longitudinal visualization of leukemic burden and CAR-T cell dynamics in a physiologically relevant ossicle-based AML model. TeaserHumanized ossicles combined with dual BLI enable tracking of AML progression and CAR-T cell dynamics in a human stromal niche.

Cell Biology↗

bioRxiv2026

Differential TIM-3 glycosylation enables specific dual targeting CAR-T therapy in acute myeloid leukemia

Chimeric Antigen Receptor (CAR) therapy for Acute Myeloid Leukemia (AML) is hindered by disease heterogeneity, antigen overlap with normal hematopoiesis, and the persistence of leukemic stem cells (LSCs). To overcome these barriers, we employed the IF-BETTER strategy to simultaneously target CD33 and the LSC-associated marker TIM-3 by pairing a second-generation CAR with a Cytokine-Costimulatory Receptor (CCR) in two dual CAR configurations (CD33.CAR/TIM-3.CCR and TIM-3.CAR/CD33.CCR). Both constructs displayed potent antigen-restricted cytotoxicity against AML cell lines and primary blasts, achieving leukemia clearance, while sparing normal immune and hematopoietic cells. Mechanistic studies revealed that the TIM-3.CAR single-chain fragment variable (scFv) recognizes a protein-proximal epitope whose interaction is selectively enhanced by AML-specific hyper-fucosylated and hyper-sialylated N-glycans. Fucosylation blockade reduced TIM-3.CAR avidity and cytotoxicity, confirming a glycosylation-modulated interaction. Integrating this glycosylation-tolerant TIM-3 scFv into a dual CAR framework enables selective targeting of AML cells, providing a rational strategy for safer and more effective AML-directed immunotherapy.

immunology↗

bioRxiv2026

Error-free and efficient prime editing in absence of maternal Polθ

Prime editing is a versatile genome-editing technology developed to make small edits while minimising the unwanted mutations associated with approaches based on double-strand breaks. In zebrafish embryos, however, prime editing is highly error-prone, typically yielding more unwanted mutations than precise edits. Here, we show that most unwanted mutations are deletions flanked by microhomologies and insertions templated from neighbouring sequences, both hallmarks of microhomology-mediated end-joining of double-strand breaks. These breaks are unlikely to result from direct cleavage of both genomic strands by the prime editor. Instead, we propose that the rapid cell divisions of early zebrafish embryos promote the conversion of nicks into double-strand breaks by replication forks. Loss of maternally deposited DNA polymerase {theta} (Pol{theta}), the core factor in microhomology-mediated end-joining, abolishes unwanted mutations and can increase edit rates above 50%, resulting in error-free and efficient prime editing. Our work identifies the source of unwanted prime-editing mutations in zebrafish and establishes a practical strategy to achieve high edit rates and near-perfect precision.

genetics↗

bioRxiv2025

Effects of anti-RANKL, Zoledronate or combination therapy in a mouse model of Fibrous Dysplasia: a preclinical study

Bone fragility and pain are major clinical issues in fibrous dysplasia (FD) of bone, a genetic disorder characterized by increased bone resorption and lytic lesions. Both bisphosphonates (BPs) and denosumab are currently used to treat FD patients, although important concerns remain unsolved. BPs downregulate bone remodeling but their effects on FD lesions and pain are variable. Contrarily, denosumab converts FD tissue into mineralized bone and prevents disease progression, but disease rebound occurs upon treatment withdrawal. The combination of these two drugs may represent an effective and safe strategy for FD treatment. We used a FD mouse model (EF1-GsR201C) to assess whether zoledronate (ZOL) addition to anti-RANKL antibody (RANKL) treatment could preserve the effects of RANKL inhibition after treatment discontinuation. We show that RANKL treatment rapidly reduced bone turnover markers (BTMs) and increased bone mass in affected skeletal segments, but FD lesions recurred shortly after discontinuation. Importantly, RANKL+ZOL combination delayed disease rebound after RANKL withdrawal, as bone density was preserved, BTMs rise was prevented, and no new lesions were observed. ZOL monotreatment increased bone density and reduced BTMs but did not fully halt disease progression. Finally, both RANKL and RANKL+ZOL treatments reduced fracture incidence and ameliorated pain-like behavior in FD mice. These results demonstrate that combining zoledronate with denosumab may effectively treat FD. This strategy could particularly benefit patients with severe, rapidly progressive disease, in which RANKL inhibition would block lesion expansion and reduce bone turnover, while zoledronate would slow down the resumption of the disease and the rebound effect.

pathology↗

bioRxiv2025

Exploring the mechanism and pattern of bone formation during RANKL inhibition in a mouse model of fibrous dysplasia

Fibrous dysplasia (FD) of bone is a genetic fibro-osseous disorder with increased bone remodeling activity. Inhibition of RANKL modifies FD lesions by inducing the replacement of the fibrous tissue with bone. This effect was observed in FD murine models receiving anti-mouse RANKL antibodies or small molecule RANKL inhibitors and in FD patients treated with denosumab. However, in neither case the mechanism and pattern of deposition of the newly formed bone were clarified. We performed radiographic, morphological and molecular analyses on EF1-GsR201C (FD) mice receiving an anti-mouse RANKL antibody. We observed that RANKL inhibition caused a decrease in the expression of genes involved in osteogenesis, osteoclastogenesis, matrix remodeling and osteoblast-osteoclast cross-talk in affected skeletal segments. Nonetheless, intra-lesional bone surfaces were covered by a continuous layer of osteoid, indicating that bone formation was actively ongoing in the pathological tissue in spite of the treatment. Accordingly, all bone surfaces within FD lesions showed calcein labeling which was never detected in the fibrous tissue far from bone. These results indicate that in the absence of RANKL activity, bone formation in FD tissue does not occur diffusely or stochastically. In contrast, it is restricted to bone surfaces where osteoprogenitor cells are orderly recruited from the adjacent fibrosis, progressively converting it into bone. Clinically, this suggests that the volume of pre-treatment bone in FD lesions may be a determinant of the skeletal improvement that individual patients may achieve during the same denosumab treatment course. As a consequence, it may also be a determinant of the severity of the rebound effect that they can experience upon treatment discontinuation.

pathology↗

bioRxiv2024

Orthogonal IMiD-Degron Pairs Induce Selective Protein Degradation in Cells

Immunomodulatory imide drugs (IMiDs) including thalidomide, lenalidomide, and pomalidomide, can be used to induce degradation of a protein of interest that is fused to a short zinc finger (ZF) degron motif. These IMiDs, however, also induce degradation of endogenous neosubstrates, including IKZF1 and IKZF3. To improve degradation selectivity, we took a bump-and-hole approach to design and screen bumped IMiD analogs against 8380 ZF mutants. This yielded a bumped IMiD analog that induces efficient degradation of a mutant ZF degron, while not affecting other cellular proteins, including IKZF1 and IKZF3. In proof-of-concept studies, this system was applied to induce efficient degradation of TRIM28, a disease-relevant protein with no known small molecule binders. We anticipate that this system will make a valuable addition to the current arsenal of degron systems for use in target validation. One-Sentence SummaryEngineered zinc-finger-based degrons enable targeted protein degradation induced by selective molecular glues.

biochemistry↗

bioRxiv2021

Disease modeling by efficient genome editing using a near PAM-less base editor in vivo.

Base Editors are emerging as an innovative technology to introduce point mutations in complex genomes. So far, the requirement of an NGG Protospacer Adjacent Motif (PAM) at a suitable position often limits the editing possibility to model human pathological mutations in animals. Here we show that, using the CBE4max-SpRY variant recognizing the NRN PAM sequence, we could introduce point mutations for the first time in an animal model and achieved up to 100% efficiency, thus drastically increasing the base editing possibilities. With this near PAM-less base editor we could simultaneously mutate several genes and developed a co-selection method to identify the most edited embryos based on a simple visual screening. Finally, we applied our method to create a new zebrafish model for melanoma predisposition based on the simultaneous editing of multiple genes. Altogether, our results considerably expand the Base Editor application to introduce human disease-causing mutations in zebrafish.

developmental biology↗