CodonBert large language model for mRNA vaccines.

mRNA-based vaccines and therapeutics are gaining popularity and usage across a wide range of conditions. One of the critical issues when designing such mRNAs is sequence optimization. Even small proteins or peptides can be encoded by an enormously large number of mRNAs. The actual mRNA sequence can have a large impact on several properties including expression, stability, immunogenicity, and more. To enable the selection of an optimal sequence, we developed CodonBERT, a large language model (LLM) for mRNAs. Unlike prior models, CodonBERT uses codons as inputs which enables it to learn better representations. CodonBERT was trained using more than 10 million mRNA sequences from a diverse set of organisms. The resulting model captures important biological concepts. CodonBERT can also be extended to perform prediction tasks for various mRNA properties. CodonBERT outperforms previous mRNA prediction methods including on a new flu vaccine dataset.

In Vivo Two-photon Imaging of Megakaryocytes and Proplatelets in the Mouse Skull Bone Marrow.

Platelets are produced by megakaryocytes, specialized cells located in the bone marrow. The possibility to image megakaryocytes in real time and their native environment was described more than 10 years ago and sheds new light on the process of platelet formation. Megakaryocytes extend elongated protrusions, called proplatelets, through the endothelial lining of sinusoid vessels. This paper presents a protocol to simultaneously image in real time fluorescently labeled megakaryocytes in the skull bone marrow and sinusoid vessels. This technique relies on a minor surgery that keeps the skull intact to limit inflammatory reactions. The mouse head is immobilized with a ring glued to the skull to prevent movements from breathing. Using two-photon microscopy, megakaryocytes can be visualized for up to a few hours, enabling the observation of cell protrusions and proplatelets in the process of elongation inside sinusoid vessels. This allows the quantification of several parameters related to the morphology of the protrusions (width, length, presence of constriction areas) and their elongation behavior (velocity, regularity, or presence of pauses or retraction phases). This technique also allows simultaneous recording of circulating platelets in sinusoid vessels to determine platelet velocity and blood flow direction. This method is particularly useful to study the role of genes of interest in platelet formation using genetically modified mice and is also amenable to pharmacological testing (study the mechanisms, evaluating drugs in the treatment of platelet production disorders). It has become an invaluable tool, especially to complement in vitro studies as it is now known that in vivo and in vitro proplatelet formation rely on different mechanisms. It has been shown, for example, that in vitro microtubules are required for proplatelet elongation per se. However, in vivo, they rather serve as a scaffold, elongation being mainly promoted by blood flow forces.

High Resolution Episcopic Microscopy for Qualitative and Quantitative Data in Phenotyping Altered Embryos and Adult Mice Using the New "Histo3D" System.

3D imaging in animal models, during development or in adults, facilitates the identification of structural morphological changes that cannot be achieved with traditional 2D histological staining. Through the reconstruction of whole embryos or a region-of-interest, specific changes are better delimited and can be easily quantified. We focused here on high-resolution episcopic microscopy (HREM), and its potential for visualizing and quantifying the organ systems of normal and genetically altered embryos and adult organisms. Although the technique is based on episcopic images, these are of high resolution and are close to histological quality. The images reflect the tissue structure and densities revealed by histology, albeit in a grayscale color map. HREM technology permits researchers to take advantage of serial 2D aligned stacks of images to perform 3D reconstructions. Three-dimensional visualization allows for an appreciation of topology and morphology that is difficult to achieve with classical histological studies. The nature of the data lends itself to novel forms of computational analysis that permit the accurate quantitation and comparison of individual embryos in a manner that is impossible with histology. Here, we have developed a new HREM prototype consisting of the assembly of a Leica Biosystems Nanocut rotary microtome with optics and a camera. We describe some examples of applications in the prenatal and adult lifestage of the mouse to show the added value of HREM for phenotyping experimental cohorts to compare and quantify structure volumes. At prenatal stages, segmentations and 3D reconstructions allowed the quantification of neural tissue and ventricular system volumes of normal brains at E14.5 and E16.5 stages. 3D representations of normal cranial and peripheric nerves at E15.5 and of the normal urogenital system from stages E11.5 to E14.5 were also performed. We also present a methodology to quantify the volume of the atherosclerotic plaques of ApoEtm1Unc/tm1Unc mutant mice and illustrate a 3D reconstruction of knee ligaments in adult mice.

Cytoskeletal-based mechanisms differently regulate in vivo and in vitro proplatelet formation.

Platelets are produced by bone marrow megakaryocytes through cytoplasmic protrusions, named native proplatelets (nPPT), into blood vessels. Proplatelets also refer to protrusions observed in megakaryocyte culture (cPPT) that are morphologically different. Contrary to cPPT, the mechanisms of nPPT formation are poorly understood. We show here in living mice that nPPT elongation is in equilibrium between protrusive and retraction forces mediated by myosin-IIA. We also found, using WT and ß1-tubulin-deficient mice, that microtubule behavior differs between cPPT and nPPT, being absolutely required in vitro, while less critical in vivo. Remarkably, microtubule depolymerization in myosin-deficient mice did not affect nPPT elongation. We then calculated that blood Stokes'forces may be sufficient to promote nPPT extension, independently of myosin and microtubules. Together, we propose a new mechanism for nPPT extension that might explain contradictions between severely affected cPPT production and moderate platelet count defects in some patients and animal models.

Three-Dimensional Culture in a Methylcellulose-Based Hydrogel to Study the Impact of Stiffness on Megakaryocyte Differentiation.

The differentiation and maturation of megakaryocytes (MKs) occurs in a 3D environment where the cells must constantly adapt to the external physical and mechanical constraints during their development and migration to sinusoid vessels. In this chapter, we present a method for culture of mouse MKs from bone marrow hematopoietic progenitor cells in a methylcellulose 3D medium with a stiffness mimicking that of bone marrow. In addition, we describe how the MKs can be recovered to allow for analysis of their differentiation and maturation state by transmission electron microscopy, immunofluorescence or flow cytometry techniques and to evaluate their ability to form proplatelets. This approach allows (1) generation of MKs with a morphology that more closely resembles the MKs that differentiate in vivo, (2) recovery of megakaryocyte phenotypes sometimes observed in vivo but not found in classical liquid cultures, and (3) study of mechanotransduction pathways induced by the stiffness of the medium.

Importance of environmental stiffness for megakaryocyte differentiation and proplatelet formation.

Megakaryocyte (MK) differentiation occurs within the bone marrow (BM), a complex 3-dimensional (3D) environment of low stiffness exerting local external constraints. To evaluate the influence of the 3D mechanical constraints that MKs may encounter in vivo, we differentiated mouse BM progenitors in methylcellulose (MC) hydrogels tuned to mimic BM stiffness. We found that MKs grown in a medium of 30- to 60-Pa stiffness more closely resembled those in the BM in terms of demarcation membrane system (DMS) morphological aspect and exhibited higher ploidy levels, as compared with MKs in liquid culture. Following resuspension in a liquid medium, MC-grown MKs displayed twice as much proplatelet formation as cells grown in liquid culture. Thus, the MC gel, by mimicking external constraints, appeared to positively influence MK differentiation. To determine whether MKs adapt to extracellular stiffness through mechanotransduction involving actomyosin-based modulation of the intracellular tension, myosin-deficient (Myh9-/-) progenitors were grown in MC gels. Absence of myosin resulted in abnormal cell deformation and strongly decreased proplatelet formation, similarly to features observed for Myh9-/- MKs differentiated in situ but not in vitro. Moreover, megakaryoblastic leukemia 1 (MKL1), a well-known actor in mechanotransduction, was found to be preferentially relocated within the nucleus of MC-differentiated MKs, whereas its inhibition prevented MC-mediated increased proplatelet formation. Altogether, these data show that a 3D medium mimicking BM stiffness contributes, through the myosin IIA and MKL1 pathways, to a more favorable in vitro environment for MK differentiation, which ultimately translates into increased proplatelet production.

Myosin IIA is critical for organelle distribution and F-actin organization in megakaryocytes and platelets.

During proplatelet formation, a relatively homogeneous content of organelles is transported from the megakaryocyte (MK) to the nascent platelets along microtubule tracks. We found that platelets from Myh9(-/-) mice and a MYH9-RD patient were heterogeneous in their organelle content (granules and mitochondria). In addition, Myh9(-/-) MKs have an abnormal cytoplasmic clustering of organelles, suggesting that the platelet defect originates in the MKs. Myosin is not involved in the latest stage of organelle traffic along microtubular tracks in the proplatelet shafts as shown by confocal observations of proplatelet buds. By contrast, it is required for the earlier distribution of organelles within the large MK preplatelet fragments shed into the sinusoid circulation before terminal proplatelet remodeling. We show here that F-actin is abnormally clustered in the cytoplasm of Myh9(-/-) MKs and actin polymerization is impaired in platelets. Myosin IIA is required for normal granule motility and positioning within MKs, mechanisms that may be dependent on organelle traveling and tethering onto F-actin cytoskeleton tracks. Altogether, our results indicate that the distribution of organelles within platelets critically depends on a homogeneous organelle distribution within MKs and preplatelet fragments, which requires myosin IIA.

Biogenesis of the demarcation membrane system (DMS) in megakaryocytes.

The demarcation membrane system (DMS) in megakaryocytes forms the plasma membrane (PM) of future platelets. Using confocal microscopy, electron tomography, and large volume focused ion beam/scanning electron microscopy (FIB/SEM), we determined the sequential steps of DMS formation. We identified a pre-DMS that initiated at the cell periphery and was precisely located between the nuclear lobes. At all developmental stages, the DMS remained continuous with the cell surface. The number of these connections correlated well with the nuclear lobulation, suggesting a relationship with cleavage furrow formation and abortive cytokinesis. On DMS expansion, Golgi complexes assembled around the pre-DMS, and fusion profiles between trans-golgi network-derived vesicles and the DMS were observed. Brefeldin-A reduced DMS expansion, indicating that the exocytic pathway is essential for DMS biogenesis. Close contacts between the endoplasmic reticulum (ER) and the DMS were detected, suggesting physical interaction between the 2 membrane systems. FIB/SEM revealed that the DMS forms an intertwined tubular membrane network resembling the platelet open canalicular system. We thus propose the following steps in DMS biogenesis: (1) focal membrane assembly at the cell periphery; (2) PM invagination and formation of a perinuclear pre-DMS; (3) expansion through membrane delivery from Golgi complexes; and (4) ER-mediated lipid transfer.

Romiplostim administration shows reduced megakaryocyte response-capacity and increased myelofibrosis in a mouse model of MYH9-RD.

Macrothrombocytopenia in MYH9-related disease (MYH9-RD) results from defects in nonmuscular myosin-IIA function. Thrombopoietin receptor agonists (eltrombopag; romiplostim) seem to improve hemostasis, but little is known about their biologic effects in MYH9-RD. We administered romiplostim to Myh9(-/-) mice (100 µg/kg, every 3 days, during 1 month). MKs increased to similar numbers in Myh9(-/-) and wild-type (WT) mice (with an increase in immature MKs), but Myh9(-/-) platelet count response was much less (2.5-fold vs 8-fold increase). A strong increase in MK nuclei emboli in the lung, in WT and Myh9(-/-) mice, indicates increased transmigration of MKs from the BM. Prolonged (but not acute) treatment with romiplostim decreased expression of GPIb-IX-V complex and GPVI, but not of GPIIbIIIa, and bleeding time increased in WT mice. Microcirculation was not altered by the increased number of large platelets in any of the assessed organs, but in Myh9(-/-) mice a much stronger increase in BM reticulin fibers was present after 4 weeks of romiplostim treatment vs WT mice. These data further encourage short-term use of thrombopoietic agents in patients with MYH9-RDs; however, myelofibrosis has to be considered as a potential severe adverse effect during longer treatment. Reduction of GPIbIX/GPVI expression by romiplostim requires further studies.