Functional Blood‐Brain Barrier Crossing by Biomimetic M13 Phage Vectors for Targeted Neuronal Delivery Authors: Silvia Vercellino, Lucia Pappagallo, Francesca De Chirico, Luca Boselli, Hazem Ahmed, Luca Spagnoletti, Giovanni Pittaluga, Paolo Emidio Costantini, Matteo Calvaresi, Pier Paolo Pompa, Alberto Danielli, Fabio Benfenati, Valentina Castagnola Published: 2026 ABSTRACT ABSTRACTThe blood‐brain barrier (BBB) is a highly selective gatekeeper. While providing essential protection against pathogens, BBB also poses a major challenge by restricting the delivery of systemically administered therapeutics. Consequently, there is a pressing need for innovative strategies that can efficiently and selectively target neuronal populations to treat diverse brain disorders. Nanomaterials enable transcellular transport across the endothelial layer; however, given the biomolecular corona paradigm, designing biomimetic hybrid nanoconstructs that preserve and leverage biological functionality within a complex physiological environment is more promising. Here, we investigated the potential of M13 bacteriophages as biocompatible and versatile nanovectors across the BBB. M13 phages exhibited: (i) a high cargo capacity within the main protein capsomer (here exploited for fluorescent labeling); (ii) offered genetic flexibility for ligand display; (iii) showed a natural propensity to efficiently cross the endothelial layer via specific intracellular pathways; (iv) maintained their structure and functionality intact. We also demonstrate that M13 phages engineered to target specific neuronal populations via single‐domain antibody display retain their specificity after BBB translocation and in a protein‐ and cell‐crowded environment. Overall, our results provide mechanistic insights emphasizing an unprecedented field of application for (engineered) M13 phages as biomimetic nanotools for targeting and delivery across the BBB. Introduction The BBB, situated at the interface between the circulation and brain parenchyma, plays an essential role in protecting the brain from the ever‐changing periphery, regulating its homeostasis, and excluding the entry of pathogens [1, 2], also by virtue of the expression of specific tight junction (TJ) proteins on the adjacent membranes of brain endothelial cells, which seal the paracellular spaces [3, 4]. The BBB strictly regulates the passage of ions and substances [5], therefore limiting the targeting of systemically administered molecular treatments to specific brain regions. In particular, BBB hinders the delivery of about 98% of conventional therapeutic molecules to the brain [5]. By virtue of their size and controllable surface properties, nanosized materials offer a promising approach to overcoming physiological barriers through active biological trafficking. Most strategies use proteins (e.g., antibodies, transferrin, lactoferrin, apolipoprotein E), peptides (e.g., the Trans‐Activator of Transcription—Tat), or glucose [6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17] to [6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17] decorate the NP surface and promote receptor‐mediated transcytosis across the barrier. However, the NP approaches for BBB crossing still present limitations mainly related to the in vitro‐to‐in vivo translation [18]. The in vivo delivery of NPs across the intact BBB is, on average, below 0.15% of the injected dose, with some exceptions up to 1% [14, 17, 19, 20]. This entails that off‐target dispersion phenomena strongly limit the efficacy of the nano‐formulations. In a peripheral administration scenario, this is mainly due to the formation of a biomolecular corona, a well‐known damper for the development of nanomedicine and its translation into clinical settings [21, 22, 23, 24, 25, 26, 27, 28]. Upon NP introduction into the systemic circulation, plasma proteins adsorb onto their surface, leading to opsonization and subsequent sequestration by the mononuclear phagocyte system (MPS), ultimately resulting in increased liver uptake. In addition, the initial composition of the corona undergoes a drastic evolution during BBB crossing and may not predict NP fate or performance once it reaches the brain [29]. A promising strategy to break this paradigm is the development of biomimetic systems that, thanks to their biological origin, can “cheat” or leverage cellular machinery while maintaining their programmed biological functionality [30, 31]. Bacteriophages have attracted growing interest in this area. Phages are the most abundant entities on Earth (estimated as 1031 phage particles) [32], ubiquitous across environments and in the human body, especially in the gut, where they are intrinsic components of the microbiome [33, 34, 35], and the main viral community in the human oral cavity [36]. M13 F‐specific filamentous phages (Ff) are viruses naturally programmed to infect Escherichia coli using the F pilus as a primary receptor, but they are totally harmless to eukaryotic cells, like the vast majority of other phage types [37]. M13 phages exhibit a nanofiber‐like morphology, with a length determined by the packaging of a single‐stranded DNA (ssDNA) genome encoding 11 proteins [38, 39]. Among them, the most commonly exploited for peptide display are pVIII and pIII [40], as they can be modified independently (orthogonally) to display biochemical ligands through genetic engineering [41]. M13 phages can be produced in an efficient and cost‐effective process [39, 41, 42, 43, 44], which makes them advantageous over antibodies. Moreover, due to their high aspect ratio (approximately 900 nm long and 6 nm wide), they exhibit excellent surface cargo capacity. So far, phage display technology has been employed to discover novel peptide ligands specific to brain vascular receptors [45, 46, 47, 48, 49, 50, 51]. In parallel, M13 has been proposed as a vector for targeting protein amyloids in neurodegenerative diseases [52, 53] or glioblastoma [54, 55]. Isolated pVIII‐peptide fusions have previously been leveraged as targeting moieties to generate targeted phage‐liposome nanoparticles, able to deliver doxorubicin or siRNA payloads to tumor cells both in vitro and in vivo. In this application, the hydrophobic pVIII protein core facilitates liposomal insertion of the fusion, resulting in liposomal display of the targeting ligand for cell‐specific targeting [56, 57, 58, 59]. This approach has also been followed to obtain dual‐targeted vesicle‐like carriers functionalized with an MCF‐7‐specific phage protein and a BBB‐specific peptide, allowing the system to cross a BBB model, target breast cancer cells, and deliver its payload [60]. Interestingly, in targeted phage‐liposomes, the payload is carried inside the nanoparticle, rather than on its surface, as for M13 phages. Also, the aspect ratios of these two nanoparticles are very different (string/spaghetti vs. sphere). The specific interactions between M13 phages and endothelial cells, and their natural propensity to overcome barriers, remain unclear. In the current study, BBB in vitro models of increasing complexity (2‐ to 3D) [61, 62] were used to evaluate the crossing efficacy of both wild‐type (WT) and re‐targeted M13 phages. Special attention was paid to phage characterization and interactions with biomolecules, and to gauging the influence of payloads (e.g., fluorescent labels) on bio‐nano interactions. We followed the journey of wild‐type M13 phages (M13WT) across the endothelial layer, investigating internalization, intracellular trafficking pathways, and transcytosis, and discovering peculiar interactions with endothelial cells, possibly derived from their evolutionarily honed nature, that might represent an asset in the development of bio‐efficient nanohybrid formulations. In addition, we propose a proof‐of‐concept model system for targeting neuronal subpopulations using M13‐engineered phages with tropism for ALFA‐Tag‐expressing neurons (M13NbALFA). Collectively, our data support M13 phages as a suitable orthogonal platform for targeted delivery across biological barriers, defining a novel paradigm for the application of these bionanovectors in brain pathologies. Results and Discussion M13 Phages are Engineered, Produced, Purified, and Characterized by a Panel of Complementary Techniques Wild‐type M13 phages (M13WT) were prepared following the superinfection method using M13KO7 phage on E. coli TG1 host cultures [63, 64] as described in the Methods section. M13 phages with targeted tropism against ALFA Tag [65] (M13NbALFA) were prepared in E. coli DH5α cells through the co‐transformation method, using a phagemid designed to display an anti‐ALFA‐Tag nanobody on the pIII protein and a helper phage vector (HyperΔori, see Methods and Figure S1 for more details). The complete characterization of the M13WT and M13NbALFA is depicted in Figure 1. TEM imaging (see also Figure S2) and SDS‐PAGE were used to confirm the filamentous structure and to analyze the capsid composition, excluding the presence of contaminants (Figure 1A,B). UV–vis absorption spectra (Figure 1C,E) showed the characteristic 269‐nm peak due to the combined absorption of DNA and protein [66]. A higher absorption region is visible below 240 nm (minimum at 245 nm), which is in good agreement with previous reports [67]. Absorption spectra were used to estimate virion concentration and sample purity, as detailed in the Methods. Characterization of M13WT and M13NbALFA. (A) Representative TEM micrographs, with high magnification insets, illustrating the structure of M13WT and M13NbALFA, respectively. (B) The SDS‐PAGE electrophoresis analysis of M13WT after production and purification shows two predominant bands corresponding to the pIII (≈ 45 KDa) and pVIII (≈ 6 KDa) coat proteins. (C) Absorption (left) and emission (right) spectra of M13WT before (dotted line) and after conjugation with CF594 (full red line). (D) Agarose gel electrophoresis of M13WT phage after conjugation with CF594, visualized through the fluorescent proteins (left) and through DNA staining with SYBR Gold (right). (E) Absorption (left) and emission (right) spectra of M13NbALFA before (dotted line) and after conjugation with CF488 (full green line). (F) Agarose gel electrophoresis of M13NbALFA after conjugation with CF488, visualized through the fluorescent proteins (left) and through DNA staining with SYBR Gold (right). The conjugation of fluorescent dyes to the viral capsid proteins was performed via an amine coupling reaction with an NHS ester (see Methods). After conjugation, the purified virions were analyzed by absorption and emission spectroscopy to assess fluorophore functionalization (Figure 1C,E). Finally, 2.5% agarose gel electrophoresis of labeled M13 allowed us to exclude the presence of free fluorophore molecules (see Figure S3A and Figure 1D,F). Since intact virions can align with the electric field and penetrate the agarose gel, we could use their electrophoretic mobility distribution as a proxy for phage homogeneity in the batch. It can be observed that while the M13WT presents a single band in the gel, indicating a monomodal size distribution, engineered phages (i.e., M13NbALFA) often exhibit multiple bands with faster electrophoretic mobility, suggesting a population of phages with smaller sizes, in agreement with phagemid and not helper phage packaging. Agarose gels also allowed us to assess the colloidal stability of the phages, as disassembled, degraded, or aggregated phage proteins would not enter the gel (see Figure S3B,C). For certain phage concentrations, a small fraction of reversibly aggregated virions could be observed in the well. Nevertheless, an appropriate dilution was sufficient to reverse the aggregation (Figure S3B). Once we assessed the quality and the reproducibility of the M13 preparations (see Figure S4) and before initiating in vitro investigation, we quantified residual endotoxin levels (lipopolysaccharide—LPS) on the M13 surface using the LAL chromogenic test. LPS amount found in the stock solution [1 × 1014 v/mL] was 5–8 EU/mL. Considering the M13 concentrations used in cell tests [5 × 1011 v/mL], the LPS concentration under our phage exposure condition was no more than 0.04 EU/mL, well below the FDA limit for contamination (0.5 EU/mL). M13 Phages are Internalized by Brain Endothelial Cells Through Energy‐Dependent Pathways After a preliminary biocompatibility assessment (Figure S5), we followed the journey of M13WT‐594 across the brain endothelium. First, the M13 uptake in bEnd.3 cells was measured over time through flow cytometry. The internalization kinetics, followed over 72 h, showed a progressively increasing amount of M13WT inside the cells (Figure 2A), which was further confirmed by confocal microscopy (Figure S6). Endothelial cells present contact inhibition that allows them to remain at 100% confluence for a few days once the barrier is formed [61, 68]. Therefore, when cells were continuously exposed to colloidally stable M13WT, cellular uptake increased over time due to the absence of cell division and the splitting of the material load between daughter cells [69]. To ensure that the fluorescent signal observed was indeed related to the labeled capsid, confocal imaging following immunostaining with Ab anti‐pVIII was performed, confirming full co‐localization of the two signals (Figure S7). M13 uptake in endothelial cells. (A) Kinetics of internalization of M13WT‐488 and M13 NbALFA‐488 in bEnd.3 cells treated for 24, 48, and 72 h (n =3 independent experiments; means ± SEM). (B) M13WT‐594 (working concentration = 5 × 1011 v/mL) uptake in bEND.3 cells as measured by flow cytometry in the presence or absence (untreated, UT) of NaN3 (1%), indicating a nearly suppressed uptake of the virus when mitochondrial energy production is inhibited. (C) M13WT‐594 uptake in bEND.3 cells as measured by flow cytometry in the presence of different inhibitors: Nocodazole (Noc, 320 nm), Cytochalasin D (CytD, 1 µm), and Chlorpromazine (CPZ, 30 µm). M13WT‐594 uptake decreases significantly in the presence of inhibitors of clathrin‐mediated uptake and phagocytosis. All data are presented as means ± SEM of n =3 independent experiments. Statistical analysis: one‐way ANOVA/Brown‐Forsythe/Bartlett's test: ****= p<0.0001. In addition, the excellent colloidal stability of both M13WT and M13NbALFA in the biological milieu was confirmed by agarose gel electrophoresis experiments reported in Figure S8, showing an invariant electrophoretic mobility of the fluorescent band related to the main coat proteins of M13WT incubated in increasing concentrations of fetal bovine serum (FBS; up to 80% v/v in phosphate buffer saline—PBS) for up to 72 h. The uptake of M13NbALFA‐488 was significantly lower at 48 and 72 h (Figure 2A), an effect possibly due to the modified structure of the pIII minor coat protein, devoted to the recognition of the bacterial receptor in M13WT. Recalling the engineering of phage reported in Materials and Methods, a possible explanation could entail the absence of pIII N1 or N2 domains in M13NbALFA. In fact, N1 and N2, which have been reported to interact with hydrophobic regions of amyloid‐β peptides [70], may favor interactions between M13 and the cell membranes in the BBB crossing model due to hydrophobic interactions. Another explanation could involve the NbALFA peptide. While sterically smaller than the sum of N1+N2 domains, its surface charge or biochemical/mechanical properties may weaken favorable interactions of the virions with the membranes to be crossed. It has been shown that mammalian cells can internalize phages through distinct endocytic pathways, highly dependent on the phage type and cellular model [71]. This idea has been known for decades [72, 73, 74] and clathrin‐ and caveolin‐dependent endocytosis, especially in the presence of engineered ligands that recognize cellular receptors, has been reported [75, 76]. T4 phages were found in endocytic vesicles of Madin–Darby canine kidney (MDCK) cells; however, experiments with a chloroquine inhibitor tended to exclude clathrin‐dependent endocytosis mechanisms in these settings [77]. Clathrin‐mediated endocytosis and macro‐pinocytosis (the most likely mechanism for micrometric objects) were also observed for M13 in HeLa and MCF‐7 cancer cells, and caveolin‐mediated endocytosis in human dermal microvascular endothelial (HDMEC) cells [78]. Finally, the PK1A2 bacteriophage was found to interact with cell surface polysialic acid, which shares structural similarity with the bacterial phage receptor in neuroblastoma cells in vitro [79]. Here, we employed a validated platform of selective uptake pathway inhibitors (see cytotoxicity assessment and efficacy of inhibitors at the selected concentrations in Figures S9 and S10) [61], prior to incubation with M13WT. First, we demonstrated that the process of internalization is, in our case, totally energy‐driven, as the uptake was nearly suppressed when cells were pre‐treated with the metabolic inhibitor NaN3 (used at 1%; Figure 2B). Microtubule disruption by nocodazole did not affect the uptake, indicating that micro‐pinocytosis was not primarily involved in the uptake process. In contrast, both actin disruption by cytochalasin D, and inhibition of clathrin assembly by chlorpromazine significantly reduced M13WT uptake (Figure 2C). Therefore, under our conditions, the active M13WT uptake in endothelial cells is mainly driven by clathrin‐mediated endocytic pathways. However, it is important to acknowledge that multiple pathways might be affected by the same inhibitor, making it difficult to establish a unique interpretation. M13 Phages Elude the Endo‐Lysosomal Pathway Cellular uptake is only the first step for targeted nanoformulations, and the ensuing intracellular trafficking mechanism may strongly impact their efficacy; therefore, the knowledge of these mechanisms and their relationship with the endocytic pathways in vitro and in vivo is of pivotal importance when designing a nanoformulation, but, at the same time, extremely challenging to dissect. The majority of reports focusing on nanoparticles suggest that all endocytic vesicles, independently of the endocytic pathway, quickly deliver their content to early endosomes (EE, marker: Rab5/EEA), and from there to two main routes: either direct recycling to the cell surface (marker: Rab11) or conversion into late endosomal vesicles (marker: Rab7) and trafficking to the lysosomal compartment (marker: LAMP1), where entrapment and degradation may occur [80]. Other mechanisms, such as retrograde transport from the plasma membrane and early/late endosomes back to the trans‐Golgi network (marker: VSP35 ‐ retromer complex), have also been observed for proteins and viruses, especially in polarized cells, and possibly linked to clathrin‐independent endocytosis [81]. We first explored the co‐localization of M13WT‐594 intracellular signal with EEA1 and Rab 5 (early endosomes), Rab 7 (late endosomes), and Rab 11 (recycling endosomes) at short incubation times (10, 20, 30, and 120 min), and all the aforementioned markers (including LAMP1 and VSP35) over longer incubation times (2, 4, 24, 48, and 72 h). Representative images and heat maps of the resulting Manders’ colocalization coefficients (MCCs) are reported in Figure S11. Co‐localization with the antibody anti‐pVIII is also reported as positive control. After 2 h of incubation, only a weak co‐localization with clathrin and Rab7 (late endosomes) was observed, which decreased over time (Figure S11). EE (and recycling endosomes) appear scarcely involved in colocalization with M13WT at early time points (10 min of incubation); however, it should be noted that the phage signal in cells at early time points, in the absence of massive internalization, is less visible than for longer treatments. Interestingly, the number of EE per cell that have internalized M13WT is significantly lower than that in “empty” cells, both after 30 min and 24 h (Figure S12). EEs are a first‐line sorting compartment, but their regulation is not yet fully elucidated [82, 83]. Early endosomal convergence is common for all pathways, with subsequent trafficking to late endosomes and lysosomes or detours via recycling endosomes or the ER/Golgi network. If the EE number decreases, endocytic vesicles may accumulate or fuse incorrectly, and their cargo might be misdirected (i.e., degraded too early or not at all). Moreover, it has been shown that endocytic vesicle internalization is not essential for Rab5‐mediated endosome formation and transport from the endosome to the vacuole, and that EE formation depends on a process initiated at the trans‐Golgi network [84]. The M13WT‐594 signal is mainly found colocalized with the lysosomal compartment during the first 4 h of incubation. Over time (24–72 h), an emerging extralysosomal population became the dominant detectable signal, as shown in Figure 3A. Intracellular trafficking. Representative confocal images of bEND.3 cells incubated with (A) M13WT‐594 (5 × 1011 v/mL, red) and (B) PS NPs (100 µg/mL, green) at different time points (4, 24, and 72 h) and the relative Manders’ colocalization coefficient (MCC, M1) of M13WT‐594/PS NPs over lysosomal staining (blue). (C) Representative confocal images of RAW 264.7 cells incubated with M13WT‐594 (5 × 1011 v/mL, red) at different time points (4, 24, and 72 h) and relative MCC, M1 of M13WT‐594 over lysosomal staining (blue). (D) Lysotracker green fluorescence intensity for either untreated (UT) or bafilomycin‐treated (BAF +) bEnd.3 cells that were incubated with PS NPs (100 µg/mL) or M13WT (5 × 1011 v/mL) for 24 (left) and 72 (right) hours (n =3 independent experiments; means ± SEM). Statistical analysis: ** p < 0.01, **** p < 0.0001 (one‐way ANOVA/Tukey's tests). (E) Emission spectra for M13WT conjugated with RatioWorks PH165 NHS ester (M13WT‐RW) at pH 7.5 and 4, showing three maxima (EmA, EmB, and EmC). Spectra on the left are excited at 488 nm, spectra on the right at 561 nm. (F) Schematic illustrating the principle of the experiment using M13WT‐RW. If the phage is in an acidic environment (i.e., lysosomes), the emission ratio EmA/(EmB *EmC) will be higher than in a neutral environment. If the phage is accumulated in the lysosome, when acidification is inhibited by bafilomycin, the emission ratio will decrease. (G) Normalized emission ratio for bEnd.3 (left) and RAW 264.7 (right) cells that were exposed, with and without pretreatment with bafilomycin (BAF +), to M13WT‐RW for 72 h (> 30 fields, n =3 independent experiments; means ± SEM). Statistical analysis: **** p < 0.0001 (unpaired Student's t‐test). These observations contrast with what is classically observed for nanomaterials, shown in Figure 3B, where polystyrene nanoparticles (PS‐NPs) of 100 nm diameter were incubated with bEnd.3 cells. Notably, M13WT‐594 appears to be less effective at evading the endo‐lysosomal pathway in other cellular settings, such as macrophages (RAW 264.7 cells, as shown in Figure 3C and Figure S13), although it shows a similar trend. This can be attributed to the specific role of macrophages, which exhibit enhanced phagocytic activity. To confirm this observation, we also analyzed the lysotracker pH‐dependent fluorescence intensity upon uptake of M13WT‐594 and control PS NPs. Indeed, lysosomal acidification and swelling are considered an indirect proxy of lysosomal accumulation [85]. The V‐ATPase inhibitor bafilomycin (BAF) [86] was used as a negative control to suppress lysosomal acidification. It must be specified that, in this experiment, it was not possible to normalize the number of fluorophores present on M13WT and PS NPs, and the total number of virions employed per well (2.5 × 1011) was more than double the number of PS NP (9.0 × 1010). As can be seen from the fluorescence quantification in Figure 3D, and representative imaging in Figure S14, internalization of M13WT, in contrast with PS NPs, did not induce significant lysosomal acidification compared to the untreated condition after 24 or 72 h of incubation. This is compatible with either a lack of accumulation or an efficient degradation of M13WT in the lysosomal compartment. Both cases are compatible with the previously discussed intracellular distribution. In a final experiment, we used the pH‐sensitive fluorescent probe Ratiowork (RW), directly conjugated to the phage protein coat (conjugation characterization shown in Figure S15). In Figure 3E, the emission spectra of M13WT‐RW are depicted for different pH values (7.4 and 4) and excitation wavelengths (488 and 561 nm). Three emission peaks were observed; they were used to calculate the emission ratio that, as schematized in Figure 3F, allows us to define if the probe is surrounded by a more acidic or neutral pH. The emission ratio was then calculated from intracellular imaging of bEnd.3 and RAW 264.7 cells subjected to M13WT‐RW incubation for 72 h with and without BAF treatment (12 h), with results shown in Figure 3G. If the M13WT virions were accumulated in lysosomes, BAF‐mediated inhibition of lysosomal acidification should result in a decrease of the emission ratio, as schematized in Figure 3F and observed for RAW 264.7 cells (in which a certain extent of co‐localization of M13 and lysosomes was observed; Figure 3C). Interestingly, for bEnd.3 cells, the trend was completely opposite, suggesting that the main M13WT‐RW population is located outside the lysosomal compartment. The observed tendency for M13 to elude the canonical endo‐lysosomal pathway in endothelial cells could, on one hand, be linked to its viral nature [87], and has previously been observed for another phage type [88]. On the other hand, this could be an advantage if the nanovector M13 has to cross the endothelial layer. Bypassing the endo‐lysosomal accumulation to induce a more efficient release from cells has been previously shown for NPs [89]. Indeed, accumulation in the lysosomal compartment would lead to degradation of the viral capsid, as shown in the simulated lysosomal environment presented in Figure S16. However, it is not clear whether the observed endosomal escape results in cytosolic release or whether the M13 undergoes preferential export pathways. As shown in Figure S17, export mechanisms in this cellular setting also differ from those reported for classical nanomaterials. When cells were pretreated with Brefeldin A (BFA)—a potent inhibitor of intracellular vesicle formation and protein trafficking between the endoplasmic reticulum (ER) and the Golgi apparatus [90] ‐ or monensin (MON)—a protein transport inhibitor known to regulate exosome secretion—in a 48 h pulse‐chase experiment, the M13 export significantly decreased after 24 h of chasing. This reduction suggests the involvement of recycling pathways in M13 trafficking and export, in contrast to the behavior observed for PS NPs. Treatment with Nocodazole (NOC—microtubule depolymerization), did not affect the export rate, in contrast to what observed for PS NPs. Taken together, our results suggest that M13WT, after internalization via active processes (Figure 2), follows distinct intracellular trafficking routes. The majority of the internalized population appears to be rapidly redirected from early endosomes toward the ER–Golgi secretory pathway. Consistently, the extracellular export is significantly impaired in the presence of inhibitors that block vesicular transport either from the ER to the Golgi or from the Golgi to the plasma membrane (Figure S17). We cannot exclude the existence of a minor subpopulation routing through the endo‐lysosomal pathway, as indicated by its colocalization with late endosomes (Rab7) at 2 h and with lysosomes (LAMP1) at 4 h (Figure S11). However, since microtubule depolymerization does not affect the export rate, it is unlikely that long‐distance vesicular transport is required for M13 trafficking (Figure S17). This suggests that even this subpopulation may ultimately escape the endo‐lysosomal compartment through a mechanism that preserves cell viability (see Figure 3, Figures S5, S13, and S14). M13 Phages Maintain Their Structure and Infective Potential After Transcytosis Across the Brain Endothelial Layer To evaluate the capability of labeled‐M13 phages to translocate across a simple 2D in vitro model of the BBB, we cultured the murine brain endothelial monolayer formed by bEnd.3 cells on a Transwell membrane. This model was validated for barrier formation and tight junctions’ expression in our previous works and in Figure S18 [61, 68]. Transcytosis of bacteriophages (with preferential apical‐to‐basolateral transport) is considered a natural and ubiquitous process, which explains the relatively high numbers of bacteriophages in tissues and organs of the human body [77]. Again, the mechanisms and efficacy of transcytosis largely depend on the combination of phage type and cellular model. We accordingly gauged the amount of M13WT translocated across the Transwell membrane by collecting basolateral fractions at different time points and measuring their fluorescence. To evaluate the transcellular component of the crossing with respect to the paracellular one, we compared the translocation across a fully formed and functional barrier with the translocation across a non‐confluent layer of endothelial cells (see low density—LD—condition in Figure S18). This normalization allowed us to estimate the contribution of the Transwell membrane per se, which represents an intrinsic diffusion barrier and a limitation of the 2D model. From our results (see Figure 4A), both labeled M13WT and M13NbALFA showed a “transcellular crossing ability” close to 40% (see Methods for the calculation), and about 10% of the total initial mass was found in the basolateral compartment. We also addressed the possible influence of the specific fluorophore on phage biological interactions, given that the measured transcellular crossing ability does not differ significantly when different fluorophores are used (see Table S1 for comparative p‐values). Considering that the covalently bound fluorophore units are in the range of 84–154 over about 2700 pVIII proteins of the main coat (therefore 5%), the differences (e.g., changes to the phage surface charge) contributed to our systems appear to be marginal. M13WT crossing of a 2D BBB in vitro model. (A) Kinetics of M13 permeability through a 2D Transwell BBB model evaluated by fluorescent quantification in the basolateral fraction and expressed as a percentage of crossing through a fully formed barrier (transcellular crossing), normalized against a non‐confluent cell layer (paracellular crossing). bEnd.3 cells were treated with M13WT‐594, M13WT‐488 (n =6 independent experiments; mean ± SEM, p‐value ≥ 0.33), or M13NbALFA‐488 (n =3 independent experiments, mean ± SEM, p‐value ≥ 0.64) to assess the contribution of the phage genetic engineering. (B) Representative TEM micrographs showing the morphology of M13WT recovered from the basolateral compartment of the Transwell setup. Scale bar: 200 nm. (C) Immunoblot analysis of the phage‐containing basolateral compartment compared to different dilutions of the M13WT stock, showing bands related to the major pVIII coat protein (8 KDa). (D) Quantification of infective M13WT (plaque‐forming units—PFU) compared to the fluorometric quantification of translocated M13WT collected from the basolateral fractions (FL). PFU are reported as normalized based on the starting stock concentration (n =3/6 independent experiments, mean ± SEM). Statistical analysis: unpaired Student's t‐test. In addition, using the same fluorophore, we compared the phage translocation (M13WT‐568) with fluorescently labeled transferrin (Tf‐568), as a reference condition (Figure S19). Applying the same normalization strategy across low‐density cultures, we observed that, despite the 50% degree of colocalization with clathrin of M13 compared to Tf‐568, consistent with the subpopulation hypothesis outlined above, the phage translocation rate is nearly two‐fold higher than that of Tf‐568. BBB crossing based on phage fluorescent labeling does not exclude the possibility that the measured fluorescence belongs to single or aggregated pVIII units upon structural disassembly of the phage virions or to free label. For these reasons, we set up three complementary label‐free techniques to detect intact phages in the basolateral fraction of the Transwell. Figure 4B reports a representative TEM micrograph of the basolateral fractions subjected to multiple purification steps to isolate and enrich the phage component (see Methods for more details). Although some background staining was present, the filamentous M13WT structure was distinctly visible and consistent with the characterization shown in Figure 1A. We also performed immunoblot analysis (Figure 4C) of the basolateral fractions, confirming the presence of pVIII protein with an electrophoretic mobility comparable to the control conditions. In parallel with assessing the integrity of the structure, we investigated whether the phage infectivity was unaltered after crossing the BBB. To this end, we performed a plaque assay, a method widely used to biologically characterize and quantify intact, infectious virions [91]. In a bacterial culture grown on agar‐based media, plaques develop following phage incubation from a single phage‐infected bacterium, called plaque‐forming units (PFU), which forms a circular spot of reduced turbidity in the bacterial layer. By tuning extreme dilutions of the phage solution (to ensure a single phage per spot), it is possible to assess the infectivity and the number of infective virions in the starting solution. First, for a given batch of M13WT, we compared the quantification of the stock solution obtained from the absorption spectrum (9.0 × 1013 virions/mL) with that obtained from the plaque assay (5.5 × 1013 virions/mL). A discrepancy between the values obtained with the two techniques is to be expected: the first accounts for all the virion components (proteins and DNA) present in the solution, whereas the second quantifies only the infective ones. Using the same quantified batch of M13WT, we performed three independent infections using a single basolateral fraction of a Transwell incubated with the phages in the apical compartment. Figure S20 depicts representative plaque assay dishes of three serial dilutions of M13WT collected in the basolateral (BL) compartment after BBB crossing, and the relative PFU counts. The extracted values of total virions in the BL fractions (PFU) were then normalized (considering the variability observed in the quantification of the stock concentration) and compared with the total virions in the BL fractions obtained from fluorometric measurements (Figure 4D). Notably, we found that almost 95% of translocated phages were still infective. Collectively, the results confirm that fluorescence is a good proxy to study M13 translocation and exclude that the observed signal is merely due to structural modifications, pVIII protein disassembly, or detachment of the fluorescent probe from the pVIII capsomers. In addition, all data indicate that translocation across a cellular layer does not induce significant structural or functional modifications in the filamentous phage. To challenge our biological vectors, we moved to a more complex 3D BBB model that includes primary human astrocytes and pericytes spontaneously associated with human endothelial cells, forming an assembloid that recapitulates the supracellular physiological architecture of the BBB. This model ensures a higher biological relevance than the 2D Transwell monolayer and has been validated for barrier formation (dextran 4 KDa permeability) and tight‐junction expression in our previous work [61, 62]. The schematic in Figure 5A summarizes the experimental steps for forming BBB assembloids, exposing them to M13WT‐594 for 72 h, slicing them, and analyzing phage penetration. In this configuration, the labeled phage needs to be internalized by the peripheral endothelial layer (given that the paracellular space is sealed by tight junctions, as shown in Figure 5B) and then transcytosed and exchanged between cell types [92] to reach the assembloid core. Astrocytes and pericytes were pre‐stained to distinguish the assembloid core from the brain endothelial layer and to quantify the localization of M13WT‐594 in the two regions. Experimental data in Figure 5C show that M13WT‐594 is evenly distributed in the two areas, indicating an excellent propensity for penetration and intercellular exchange of material. M13WT interactions with a 3D multicellular assembloid model of the BBB. (A) Schematics illustrating the experimental approach: 3D multicellular assembloids were prepared using prelabeled human astrocytes and human pericytes together with unlabeled human brain endothelial cells. Assembloids were treated with M13WT‐594 for 72 h, fixed, cryo‐sliced, and observed under a confocal microscope. (B) Representative 3D reconstruction of an assembloid slice: nuclei are in cyan, prelabelled (cell tracker) astrocytes and pericytes are in white, zonula occludens 1 (ZO‐1) immunoreactivity is in pink. Scale bar: 50 µm. (C) Representative confocal images of assembloid slices and related quantification of M13WT‐594 (indicated in yellow) distribution in the endothelial layer (pink dotted line) or in the astrocyte/pericyte core (white dotted line, white cells). Nuclei are represented in cyan. The capacity of phages to cross or interact with eukaryotic cells and tissues is supported by literature. Fast and directional transcytosis of different bacteriophages across confluent cell layers originating from the gut, lung, liver, kidney, and brain have been reported, with the transport displaying a preferential apical‐to‐basolateral directionality thus providing a mechanistic explanation for the occurrence of phages within the body [77], including blood, lymph, and other organs, as well as the brain, that are generally considered sterile [33]. M13, in particular, because of its filamentous structure, seems to possess the unique capability to bypass anatomical and physiological barriers characterized by varying degrees of selectivity and permeability, including barriers that are commonly impermeable to much smaller therapeutic molecules or other phages [93]. In mouse models, for example, M13 has been shown to accumulate in the hippocampus and olfactory bulb shortly after intranasal administration [94]. Moreover, engineered M13 phage conjugates were recently shown to deeply penetrate multicellular colon cancer spheroids [95] or to target neuroblastoma xenografts in Zebrafish after merely adding the phages to the water [96]. Engineered M13 Phages can Target a Neuronal Subpopulation Following the Crossing of the Brain Endothelial Layer In Vitro In Figure 2A, we showed that M13 phages with targeted tropism for ALFA‐Tag, labeled with a 488‐fluorophore, can cross a 2D BBB model. Next, we assessed the targeting capability of the re‐targeted phage on a model system of primary neurons to simulate the hypothetical targeting of specific neuronal populations from the peripheral circulation. Primary cortical neurons from embryonic mice were transiently transfected at 7 days in vitro (DIV7) with the ALFATag_hCD4_mOrange construct (Figure S1) to generate a synthetic target and assess the specific targeting ability of M13NbALFA. The chimeric hCD4 transmembrane protein, expressed under the control of the cytomegalovirus (CMV) promoter, presented an epitope ALFA‐Tag peptide on the extracellular side and an mOrange fluorophore fused to the intracellular domain of hCD4. The correct expression of the construct was demonstrated by the full colocalization between the mOrange reporter and ALFA‐Tag immunoreactivity, as shown in Figure S22 in both primary neurons and HEK293T cells. As illustrated in the schematics in Figure 6A, we incubated primary neurons with M13NbALFA‐488 and observed the co‐localization between the mOrange reporter and the 488‐fluorescent phages. After 1 h of incubation, a strong and very specific targeting of the transfected neuronal subpopulation was observed (Figure 6B and Figure S23), which was completely lacking when M13WT‐488 was used as a control (Figure S24). Interestingly, targeting also occurred in complex dynamic biological environments with high protein concentration, as illustrated in Figures S25 and S26, indicating a negligible influence of coronation processes on the biological identity of our nanovector, as observed in our previous publication [97]. A chase experiment, in which we monitored binding up to 96 h after 1 h of incubation and a washing step, demonstrated that co‐localization remained unaltered over time (see representative images and MCC in Figure 6C,D). M13NbALFA targeting efficacy in murine primary neuronal cultures. (A) Schematics of the model used to assess the M13NbALFA efficiency in neuronal targeting: primary neurons are transfected to transiently express human CD4 transmembrane protein carrying a fluorescent reporter (m‐Orange, intracellular) and an extracellular ALFA‐Tag. After exposure to M13NbALFA‐488, efficient targeting is reflected by the co‐localization between the m‐Orange reporter (neurons) and the 488 label (phage). (B) Representative confocal image of primary cortical neurons expressing the ALFATag_hCD4_mOrange (magenta) after 1 h exposure to the M13NbALFA‐488 (yellow), showing a complete overlap of the two signals. No phage signal is visible in non‐transfected neurons (gray: tubulin‐β3, cyan: nuclei, scale bar: 20 µm). (C) Representative images of primary cortical neurons expressing ALFATag_hCD4_mOrange (magenta) exposed for 1 h to M13NbALFA‐488 (yellow), washed, and chased for 48, 72, and 96 h (n =3 independent experiments, mean ± SEM). (D) Stability of the hCD4‐ALFA Tag binding with M13NbALFA‐488, expressed as MCC over time. (E) Schematic of the “cross‐ting” experimental setup: bEnd.3 cells form a 2D barrier on the Transwell membrane, and primary neurons are placed in the basolateral compartment. M13NbALFA‐488 is added to the apical compartment, and the neurons are analyzed 72 h later using a confocal microscope. (F) Representative confocal images showing an unaltered targeting efficiency of the M13NbALFA‐488 after crossing of the 2D BBB layer (magenta: CD4‐AlfaTag, yellow: M13, gray: tubulin‐β3, cyan: nuclei, scale bar: 20 µm). Finally, instead of directly incubating the transfected neurons with M13NbALFA‐488, we placed a Transwell membrane with a confluent bEnd.3 cells layer on top of the neuronal culture and introduced the phages from the apical compartment in a so‐called “cross‐ting” (crossing + targeting) configuration (see schematic in Figure 6E). Remarkably, even under these conditions, an efficient and specific targeting was achieved, confirming that M13 can overcome the endothelial layer, not only preserving its structural integrity and infectivity, but also its specific tropism toward the ALFA‐epitope tag (Figure 6F). Conclusions Viruses are entities at the border between chemistry and life: they can be considered replicants only in the biological context in which they are supposed to function, while outside this context, they simply constitute an ordered protein assembly, packaging DNA or RNA; in other words, a biological nano‐object. A nanomaterial in a biological environment interacts with cells and tissues through its biological identity, defined by the layer of biomolecules (the corona) that spontaneously adsorb onto its surface. The capability to drive biological interactions begins, therefore, by designing nanomaterials with controlled biological features, often borrowed from nature (so‐called biomimetic systems). In this way, the nano‐object will be able to interact with biological machinery in a sophisticated manner, already honed by nature over years of evolution. In this study, we advanced the concept of biomimetics, exploring whether a bacteriophage, although naturally interacting with prokaryotic cells where it can replicate, can nonetheless exploit certain intrinsic features, such as its nanostructure, its morphology, and molecular details of the surface, to mediate and direct specific interactions outside of its native context, namely with eukaryotic systems. Here, the M13 bacteriophage is proposed as a biomimetic carrier for crossing the brain endothelial layer and targeting specific neuronal populations. The bacteriophage system offers multiple advantages for targeted delivery over, for instance, classical antibodies, such as cost‐effective, scalable production, high cargo capacity of the capsid coat, and high flexibility for orthogonal genetic modification (and post‐functionalization) of the major (pVIII) and minor (pIII) capsid proteins. Using in vitro models of the BBB of increasing complexity, from 2‐ to 3D, we demonstrated that this biocompatible vector has an intrinsic propensity to cross the endothelial layer effectively via transcellular pathways. By labeling phages with fluorescent tags, we were able to follow their journey across the endothelial cell layer, and, in addition, to assess their capability to transport small, conjugated molecules across the barrier. In all tested cases, M13 phages maintained their functional features after crossing the endothelial layer. After translocation, wild‐type M13 virions were still able to infect a bacterial culture, whereas M13NbALFA could target a neuronal subpopulation transfected to express a matching docking moiety (the ALFA epitope tag) in a very specific manner. Despite some difficulties that might arise in the translation of these biological tools into an in vivo setting (i.e., ensuring stability of colloidal suspension and surface chemistry), studying M13 trafficking mechanisms across the brain endothelium in vitro is pivotal to engineering nanoformulations with efficient delivery properties. Our observations suggest that M13WT is internalized by brain endothelium cells via active processes, but then behaves very differently from other nanomaterials. Although different phage subpopulations might initially follow distinct trafficking pathways, the majority are rapidly redirected from early endosomes toward the ER–Golgi secretory pathway. We observed no significant co‐localization with lipid vesicles or lysosomes at extended exposure times, nor any involvement of long‐distance vesicular transport, suggesting that M13 ultimately might elude the endo‐lysosomal compartment. The ability to avoid lysosomal degradation has been observed in viruses that either fuse with the cell membrane (especially in enveloped viruses) or employ membrane‐modifying proteins that physically pierce the endosomal membrane. The process is sometimes facilitated by cellular cues (receptor binding, pH, proteolysis, etc.) that can trigger conformational rearrangement of the capsid [87]. While a dedicated study will be necessary to establish a definitive escape and export mechanism in this context, our findings strongly suggest that deepening this aspect may lead to precise control over intracellular pathways—a topic that has thus far posed significant challenges for engineered nanomaterials. In conclusion, we demonstrated that M13 phages engineered to target specific neuronal subpopulations retain their tropism after crossing the BBB in vitro. Interestingly, they maintain stability and targeting capability even in complex biological environments, and their weak interactions with plasma proteins might make them superior biologically derived nanomaterials, helping mitigate the frequent issue of protein coronation in the systemic circulation. Based on all these considerations, we interpret our findings as an opportunity in the nano field, opening numerous possibilities for the development of phage‐based nanotools as delivery vehicles for the brain. Experimental Section/Methods M13WT Production and Characterization Phage Production E. coli (TG1) Bacteria Were Infected With M13KO7 Helper Phage (New England Labs, N0315S), Supplemented With 0.4 mm IPTG and Kanamycin (25 mg/mL). The Culture was Grown Overnight to Allow for Phage Proliferation The Bacteria‐phage Suspension Was Then Spun for 25 Min at 10 000 X G to Remove the Bacterial Cells. Phage Purification Phages were then isolated from the supernatant by adding 4% w/v polyethylene glycol (PEG) 6000 (Merck) and 3% w/v NaCl (Merck) to the phage supernatant to induce precipitation. After complete dissolution of the compounds, the supernatant was kept on ice for 1 h, followed by centrifugation for 25 min at 15 000 x g to precipitate the phages, which were eventually resuspended in Milli‐Q water and further purified via isoelectric point (IEP) precipitation, by lowering the pH to 4.2 through HCl addition [41, 67, 98, 99, 100]. The phages were then pelleted by centrifugation at 15 000 x g for 5 min and resuspended in phosphate‐buffered saline (PBS). The procedure was repeated twice to ensure full PEG removal. The stock solution was kept in PBS with 1% NaN3 to prevent the growth of contaminants. Absorption Spectroscopy Characterization The Phage Concentration [v/mL] Is Calculated from the Phage Suspension UV–vis Absorbance Spectrum Using the Following Equation: v/mL=(Abs@269nm−Abs@320nm)·6x1016Phagegenomebp TEM Imaging TEM grids (formvar‐coated copper 200 mesh, Ted Pella, Redding, CA, USA) were cleaned through plasma oxygen. Then, a 5 µL‐drop of M13WT or M13NbALFA (at the concentration of 1 × 1011 v/mL) was added to the grid, allowed to deposit for about 30 s, and then removed with filter paper and negatively stained using 1% uranyl acetate. TEM imaging was performed using a JEOL JEM 1400 microscope (JEOL, Tokyo, Japan) operating at 120 kV. LPS Quantification LPS amount in M13WT batches was quantified using Pierce LAL Chromogenic Endotoxin Quantitation Kit (Thermo) following the manufacturer's instructions. M13WT was diluted 1:10 from the stock solution. Gel Electrophoresis Agarose gel electrophoresis was performed by preparing a native 2% w/v agarose gel with or without adding SYBR Gold dye, which was diluted 1:10 000 in the melted agarose solution. The electrophoretic run was carried out for 50 min at 90 V. The gels were imaged with an iBright gel imager equipped with a fluorescence light source. SDS‐Page Phages were denatured by the addition of β‐mercaptoethanol‐containing loading buffer and boiled for 15 min at 98°C. Samples were then run in a 13% acrylamide gel for around 90 min at 130 V. Phage Labeling With NHS‐Fluorophores M13 phages were labeled through NHS reaction with different fluorophores. First, the phage buffer was exchanged with a Zeba 40 kDa column to replace PBS with bicarbonate buffer pH 8. The NHS dye (10 mm) was then added to the phage suspension in a ratio of 1 µL dye for 200 µL phage (1014 virions/mL), and the reaction was carried out for 1 h at room temperature under shaking. Excess free dye was removed by buffer exchange with two Zeba 40 kDa columns, and the samples were kept in PBS at 4°C. Fluorescence Spectroscopy The emission was measured with a TECAN Spark microplate reader, adding 50 µL of phage suspension in PBS per well. For the CF594 labelled phages, samples were excited at 594 ±10 nm, and their emission was collected from 600 to 800 nm. For the CF488 labelled phages, the samples were excited at 490 ±10 nm, and the emission was recorded between 600 and 750 nm. Targeting Platform Development Molecular Cloning The Nb‐ALFA phage was produced by cloning the coding region for the anti‐ALFA tag nanobody [65] from the pET51b (+) EGFP_NbALFA construct (kindly gifted from Kai Johnsson—Addgene plasmid # 136626; http://n2t.net/addgene:136626; RRID: Addgene_136626) into the pComb3XSS DsbA phagemid [101]. pET51b (+) _EGFP_NbALFA was digested with SacI/SpeI restriction enzymes (NEB, New England Biolabs), purified through gel extraction, and cloned directly into the SacI‐SpeI linearized pComb3XSS DsbA, to generate phagemids pComb3XSS Nb‐ALFA. Positive clones were verified by Sanger sequencing (Mix2Seq kit, Eurofins Genomics). The molecular cloning of the ALFA‐Tag hCD4_mOrange synthetic target was obtained through a mutagenic PCR. hCD4_mOrange (hCD4‐mOrange was a gift from Sergi Padilla Parra—Addgene plasmid # 110192; http://n2t.net/addgene:110192; RRID: Addgene_110192) was used as template together with phosphorylated primers AD0302 (GAGCTCAGAAGAAGACTGACCGAACCCAAAGTGGTGCGGGCAAAAAAG) and AD0303 (CTCTTCGAGACGGCTCCCGCCGCTGGGCTTTCCCTGAGTGGCTG) in a PCR reaction using Q5 high fidelity DNA polymerase (NEB, New England Biolabs). Amplicons were then ligated and transformed into rubidium chloride competent E. coli DH5α cells. Positive clones were verified by Sanger sequencing (Mix2Seq kit, Eurofins Genomics). M13NbALFA Production and Purification Engineered phages were produced by co‐transforming E. coli DH5α cells with pComb3XSS Nb‐ALFA and the hyper phage M13KO7Δg3pΔoriF1 [101]. Co‐transformants were cultured for 24 h into liquid LB supplemented with Ampicillin (100 mg/mL), Kanamycin (25 mg/mL), and 0.4 mM IPTG to induce the expression of the Nb ALFA‐pIII fusion protein. The helper phage carries a kanamycin resistance gene, which, together with the resistance of the phagemid (ampicillin resistance cassette), allows the selection of co‐transformed positive DH5α clones that harbor both vectors. Positive clones were re‐inoculated in the liquid Lysogeny Broth (LB) in the presence of antibiotics and Isopropyl β‐D‐1‐thiogalactopyranoside (IPTG) to stimulate the expression of the fusion protein on the phagemid under the control of the lac promoter. This preparation method could ensure a homogenous population of phages in which all five copies of the pIII proteins are modified, and only the phagemid genome, but not the helper genome, can be packaged into the modified phage virions. Phage purification and characterization were performed as previously described for M13WT. In Vitro BBB Models 2D Murine Transwell Model Cell Culture Immortalized murine brain endothelial cells (bEnd.3 [BEND3] (ATCC CRL‐2299) were cultured in Dulbecco's modified eagle medium (DMEM) complemented with 10% FBS, 1% Penicillin/Streptavidin, and 1% Glutamine (cDMEM). Cells were maintained in T75 culture flasks kept in an incubator with 5% CO2 and 90% humidity at 37°C. Cells were maintained between passages 25 and 30. The cell culture medium was replaced twice a week. All reagents were purchased from Thermo Fisher unless specified differently. TEER Measurements Trans‐endothelial electrical resistance (TEER) measurements were performed with a CellZscope+ instrument (NanoAnalytics, Münster, Germany) on bEnd.3 cells grown in monolayers onto collagen‐coated upper Transwell membranes (Greiner Bio‐One). Viability Assay bEnd.3 cells were exposed to 5 × 1911 v/mL M13WT phages for 24, 48, and 72 h. The cell medium was then replaced with a solution consisting of propidium iodide (PI, 1 µm), calcein‐AM (2 µm) and Hoechst 33342 (1 µm) and incubated at 37°C for 10 min. Cells were imaged with a Nikon Eclipse‐80i upright epifluorescence microscope equipped with a 10× (0.5 NA) air objective. At least 5 fields per sample were acquired and analyzed with the CellProfiler software. Immunofluorescence and Confocal Imaging Cells were fixed in 4% paraformaldehyde in PBS for 15 min at room temperature (RT). Cells were then permeabilized with 0.1% (v/v) Triton X‐100 for 10 min and kept in 2% BSA blocking solution (w/v) for 30 min at RT. The incubation with the primary antibody was performed either for 1 h at RT or overnight at 4°C, followed by PBS washes and incubation with the appropriate secondary antibody for 1 h at RT. The primary antibodies used were rabbit monoclonal anti‐zonula occludens (ZO1; #61‐7300, ThermoFisher Scientific; 1:200), rabbit polyclonal anti‐claudin‐5 (CLD5; # PA599415, ThermoFisher Scientific; 1:500), and rabbit monoclonal anti‐tubulin β3 (#T2200, Sigma–Aldrich; 1:500), anti‐ALFATag (#N1505, NanoTag Biotechnologies; 1:500). The secondary antibodies were purchased from Molecular Probes (ThermoFisher Scientific). Cells were then incubated for 5 min with Hoechst (1 µm, #3342, Sigma) and mounted with Vectashield antifade mounting medium (#H‐1000‐10, Vector Laboratories) on microscope slides. Image acquisitions were performed using a confocal microscope (Leica SP8 TCS, Leica Microsystems GmbH, Wetzlar, Germany) with a 63X (1.4 NA) oil immersion objective. Confocal images were acquired by using 1,024x1,024 scan format at 400 Hz speed. Uptake bEnd.3 cells were seeded 3 days prior to the experiment in 24‐well plates at a cell density of 60 000/mL (0.5 mL per well). Once confluent, cells were treated with M13WT‐594 (5 × 1011 virions/mL) for 24, 48, and 72 h. For the inhibition of uptake pathways, cells were pre‐treated for 30 min with NaN3 (#71289, Merk; 1%), nocodazole (Noc, #M1404, Merk; 320 nm), chlorpromazine (CPZ, #C8138, Merk; 30 µm), and cytochalasin D (CytD, #C8273, Merk; 1 µm), respectively. Cells were then washed, harvested by trypsinization, suspended in PBS, and fixed in 2% paraformaldehyde (PFA). Cells were then run in flow cytometry employing a CytoFLEX S instrument (Beckman) equipped with 4 laser lines. The phage fluorescence was detected after excitation with a 561 nm laser through a 610/20 nm band pass filter to quantify the internalization rate. Single cells were gated based on physical parameters (FSC/SSC), and the results were expressed as a fold increase in fluorescence intensity over the untreated control. Co‐Localization with Intracellular Compartments bEnd.3 cells and RAW264.7 cells were incubated for 4, 24, and 72 h with M13WT (5 × 1011 virions/mL). Cells were washed three times with fresh PBS and fixed with 4% paraformaldehyde (PFA) for 30 min at RT, washed, and stored in PBS. For immunostaining the following Ab were used: EEA1 (#MA5‐14794 ThermoFisher; 1:200), Anti‐LAMP1 (#L1418 Merk; 1:500), anti‐clathrin (#MA1065, ThermoFisher; 1:200), anti‐Rab5 (#R4654, Merk; 1:100), anti‐Rab7 (#SAB5701027, Merk, 1:200), anti‐Rab11 (#HPA036407, Merk, 1:50), anti‐VSP35 (#PA5‐18248, ThermoFisher; 1:200), anti‐M13 (#61097, Progen; 1:00). Images were acquired with a Leica SP8 confocal microscope, equipped with a HC PL APO CS2 63x/1.40 oil objective. The image pixel size was 0.180 um × 0.180 um for the xy‐dimension, and 0.300 um for the z‐dimension. The pinhole size was set at 1 AU (Airy Unit). Typically, at least 20 Z‐stacks were acquired to image an appropriate cell portion. For the colocalization analysis from confocal imaging a CellProfiler in‐house pipeline was developed. Early Endosomes Analysis CellProfiler was used with an in‐house built pipeline. Cells were first segmented based on the nuclei staining. The cytoplasm was then identified using a low‐intensity threshold for endosome staining. Cells were classified as phage‐rich or phage‐poor based on the intensity of the phage channel within the cytoplasm of each cell. The two distributions were then analyzed by counting the number of endosomal spots identified by intensity with a strict intensity threshold in the endosomes channel. Lysotracker bEND.3 cells were incubated for 24 and 72 h with M13WT (5 × 1011 virions/mL) or PS NPs (Fluorospheres Carboxylate‐Modified, 100 nm, ThermoFisher, 100 µg/mL), then washed three times with fresh PBS, incubated with Lysotracker Green (ThermoFisher, 50 nm) for 10 min, and imaged at the confocal microscope. Exposure for 16 h to bafilomycin A1 (SantaCruz biotechnology, 100 nm), followed by three washes with PBS and incubation with Lysotracker (same as above), was used as a control condition for lysosomes alkalinization. Ratiowork Probe M13WT was labelled with the ratiowork dye through NHS reaction as previously described for fluorophores. bEnd.3 and RAW264.7 cells were exposed to M13WT‐RW (5 × 1011 virions/mL) for 4, 24, and 72 h, then washed three times with PBS and imaged at the confocal microscope. Bafilomycin A (100 nm) was added for 12 h prior to imaging. For the analysis, a CellProfiler in‐house pipeline was developed. Pulse and Chase (Export) Pulse: bEnd.3 cells were treated with M13WT‐594 (5 × 1012) for 24 h, then washed three times with FluoroBrite DMEM supplemented with 10%FBS, 1% P/S, 1% glutamine. Chase: cells were then incubated with brefeldin A (#B6542, Merck; 36 µm), monesin (#PHR2775, Merck; 15 µm), nocodazole (Noc, #M1404, Merk; 320 nm) or left as a control condition. Supernatant medium was collected at different time points (0.5, 24, and 72 h), the fluorescence was measured using a TECAN Spark microplate reader (TECAN, US), and the M13WT‐594 was quantified through interpolation from a calibration line. BBB Crossing For all BBB crossing experiments, bEnd.3 cells were seeded on 150 µg/mL collagen‐coated upper Transwell membranes (12 mm Ø inserts, pore size 3.0 µm, growth area 1.12 cm2). The cell density was 90 000 cells/mL 3 days prior to the experiment, while the low‐density paracellular transport controls were seeded at 10 000 cells/mL. The barrier formation was monitored by TEER measurements, and 10 Ω*cm2 was considered the minimum TEER value for a completely formed barrier. Cells were treated by replacing the apical fraction with 0.5 mL 5 × 1011 v/mL phages suspension in cDMEM. The basolateral fraction consisted of 1.5 mL of fresh cDMEM, from which 100 µL were withdrawn after 2, 24, 48, and 72 h from the phage treatment. The virions translocated across the barrier were quantified by fluorescence quantification in a TECAN Spark microplate reader (TECAN, US), by interpolation from a calibration line. The phage virion number translocated across the low‐density cell control Transwell (without a formed barrier) was considered the 100 % passage through the paracellular transport and used to normalize the phage transcellular crossing. Transferrin was employed to validate the integrity and functionality of the receptor‐mediated transcytosis pathway, providing a standardized positive control for efficient transcellular transport. bEnd.3 cells were treated with 5 µg/mL human transferrin CF568 conjugate (Biotium, #00083) or 568‐labeled M13WT (5 × 1011 v/mL). The transferrin concentration was precisely determined via a fluorescence calibration curve to match the signal intensity of M13WT‐568, thereby ensuring uniform quantification parameters for both treatments. Phage Recovery From BL Fraction For TEM analysis of the phages in the basolateral fraction, phages were first preconcentrated by ultracentrifugation at 100 000 x g for 3 h and resuspended in Milli‐Q water. Next, phages were purified through IP precipitation at pH 4.2, as previously described, and resuspended in Milli‐Q water. Phages were then further purified by ultracentrifugation at 100 000 x g for 3 h and resuspended in 200 µL of Milli‐Q water. Phages were then stained and mounted on TEM grids as described above. For Immunoblot analysis, basolateral fractions were incubated with 4% PEG 8000 (Sigma–Aldrich) and 3% NaCl (Sigma–Aldrich) for 1 h on ice. The solution was then centrifuged at 10 000 x g for 25 min at 4°C to pellet the phages. Pellets were resuspended in 25 µL PBS and stored at 4°C. Plaque Assay E. coli TG01 bacteria were inoculated in liquid culture 4 h prior to the experiment. The OD was monitored until reaching mid‐logarithmic growth phase OD = 0.4. The phage stock, or the phages contained in the basolateral fraction, was serially diluted in LB until reaching 3 appropriate concentrations. The bacteria were infected by adding 20 µL of phage suspension to 400 µL of bacterial culture. The mixture was immediately diluted in 3.5 mL pre‐warmed soft agar (0.7 % agar w/v LB), and the mixture was carefully poured onto pre‐warmed LBA plates and incubated overnight at 37°C. On the following day, plates were analyzed with a gel imager, and plaques were counted. Immunoblotting Samples were prepared for immunoblotting by adding 6X loading buffer (containing 2.5% β‐mercaptoethanol and SDS) and boiling at 98°C for 15 min. Samples were then loaded onto a polyacrylamide gel (6% for the stacking gel and 12% for the resolving gel) together with molecular mass markers (Blue Prestained Protein Standard, Broad Range Neb), and the SDS‐PAGE run was performed at 120 V until the bottom of the gel was reached. Proteins were transferred to a PVDF membrane (Merck, 0.2 µm pore size) using a wet blotting chamber and transfer buffer (0.25X Tris‐glycine and 20% methanol). The transfer was performed at 120 V for 2 h, and the membrane was incubated with blocking solution (PBS containing 0.05% Tween and 5% dry milk) for 50 min at room temperature (RT) with stirring and eventually incubated with anti‐pVIII (diluted 1:10 000 in blocking solution) primary monoclonal antibody (B62‐FE2; ProGen), overnight at 4°C under stirring conditions. After the incubation, the membrane was washed three times with PBS containing 0.05% Tween and then incubated with HRP IgG anti‐mouse secondary antibody (Jackson ImmunoResearch, diluted 1:10 000 in blocking solution) for 1 h at RT. After three washes, the membrane was incubated with an ECL solution (Lightwave plus GVS filter technology) and developed using the ChemiDoc MP imaging system (Bio‐Rad). 3D Human Assembloid Model Assembloid Preparation Assembloids were prepared in single wells of a 96‐well plate. Wells were coated with 1% agarose in PBS to prevent cell attachment. Human brain vascular pericytes (hBVPs, ScienCell) and human astrocytes (NHA, Lonza) were pre‐labeled with CellTracker Green CMFDA Dye (Invitrogen, #C2925) for 1 h at 37°C. The two cell lines and the endothelial cells (hCMEC/D3) were washed, collected, and counted. Cells were seeded together at a cell density of 1,500 cells per well in a 1:1:1 ratio in 200 µL the EGM‐2 Endothelial medium bulletkit (Lonza), as previously described [102]. Assembloids were incubated at 37°C, 5% CO2, 95% humidity for 72 h prior to incubation for 72 h with 5 × 1011 v/mL M13WT‐CF594 in a 96‐well plate in a volume of 100 µL. Assembloid cryo‐sectioning. Assembloids were collected in a 1.5 mL tube, washed, and fixed in 4% PFA for 30 min at RT, washed, and stored in PBS. They were then embedded in O.C.T compound (Scigen Scientific Inc.) in a 1.5 mL tube and kept on dry ice until completely solidified. Next, the solid O.C.T. that had encapsulated the assembloid pellet on the tip was mounted onto a cryostat sample holder. Assembloids were sliced in 20 µm‐thick slices, collected onto PLL/Gelatin‐coated microscope glass slides and stored at −20°C until further processing. Assembloid immunostaining and imaging. The immunostaining was performed on 20 µm‐thick assembloid slices. First, samples were washed 3 times to remove the O.C.T. and then permeabilized and blocked for 1 h at RT with a 10% (w/v) BSA solution in 0.3% Triton‐X100. After 3 washes, samples were incubated with anti‐ZO1 and anti‐pVIII primary antibodies (5 µg/mL, #61‐7300, ThermoFisher Scientific and #61097 Progen respectively) in 2% (w/v) BSA solution in 0.1% Triton‐X100 overnight at 4°C. Secondary antibodies were incubated at RT for 1 h in 2% (w/v) BSA solution in 0.1% Triton‐X100, together with Hoechst. Slices were then imaged with a Leica SP8 confocal microscope equipped with a 63x oil immersion objective, and images were analyzed with Cellprofiler image analysis software. The phage distribution was evaluated by quantifying the fluorescence intensity within each mask (core ROI and endothelial ROI). Each Z‐stack was analyzed separately, then averaged to obtain a single image value. For each biological replicate, at least 3 spheroids were imaged after physical and optical slicing. As spheroids are rounded objects, the images representing the bottom of the spheroids show higher endothelial layer phage localization, while the central slices show the opposite trend. Neuronal Platform for Targeting Primary Neuron Dissection and Culture Cortical neurons were obtained from mouse dissection (E17, E18) and plated at a density of 40 000 cells/well on PLL‐coated 18 mm‐diameter glass coverslips and cultured in 2 mL of complete Neurobasal medium containing 2% B27, 1% Penicillin/Streptavidin, and 1% Glutamine. Primary Neuron Transfection Neuron transfection was carried out at DIV7 by treating the neurons with 1 µg/well of DNA hCD4‐mOrange ALFA Tag and 2.5 µg/well of Lipofectamine 2000 (Invitrogen) in plain Neurobasal medium. Neurons were incubated at 37°C for 30 min, in the mixture, washed twice, and kept for at least 48 h in their conditioned medium prior to further experiments. HEK293T cells were seeded in µ‐Plate 96‐well Square plates (Ibidi, #89626). For transfection, 1 µg/well of hCD4‐mOrange ALFA‐tag DNA was diluted in DMEM and complexed with 2 µL of Lipofectamine 2000 (Invitrogen) per well. Following a 4 h incubation at 37°C, the transfection medium was replaced with fresh culture medium. Experimental treatments or downstream analyses were performed 24 h post‐transfection. Primary Neuron Phage Targeting The targeting experiments were performed on hCD4‐mOrange ALFA‐Tag expressing cortical neurons. Briefly, the nanobody‐expressing M13 were diluted to 5 × 1011 v/mL in the neuron‐conditioned medium and incubated for 1 h at 37°C. Neurons were subsequently either fixed in 4 % PFA or, for pulse and chase experiments, washed and placed in phage‐free conditioned medium and kept in culture. Samples were imaged with a Leica SP8 confocal microscope equipped with a 63x oil immersion objective. The images were analyzed with the CellProfiler image analysis software. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting File: adhm71281‐sup‐0001‐SuppMat.pdf.