Magnetic diatom shells: nature's blueprint for cellular transport

Abstract

The ability to move cells in space without impairing their behavior is a critical goal in the development of functional biomaterials and devices, with applications ranging from regenerative medicine to biosensing. In this study, we describe a novel approach for simultaneously displacing multiple cells using magnetized diatom shells. Highly porous biosilica shells from diatoms are functionalized through a multi-step decoration process involving ferromagnetic nanoparticles. Neuronal cells are then allowed to adhere for an appropriate duration before being moved magnetically using an external magnet. We demonstrate the safe transfer of neuron cell-loaded diatom shells by pipetting, as well as their controlled movement through twirling or along a simple fluidic channel. This proof-of-concept introduces a promising strategy for safely and efficiently relocating multiple cells simultaneously, paving the way for innovative applications in tissue engineering, biosensing, and beyond.

---

1. Introduction

The precise positioning of cells within a 3D scaffold, the promotion of 3D or multilayer aggregates, and the alignment of cells along elongated surface topographies are highly desirable for tissue engineering applications.^1–4 Such approaches better mimic the physiological microenvironment and enhance regeneration outcomes.^5–7 Notably, cells can be stretched,⁸ 3D printed and integrated in devices⁹ and even internally decorated with nanomotors.¹⁰ Locomotion and transfer of cells on specialized devices are critical in applications such as model diseases, drug delivery, and biosensing.^11–13 For instance, biohybrid devices, in which cells or their biomolecular machineries are directly positioned on electronic supports to form the electrical and electrochemical interface between the device and the extracellular environment, would significantly benefit from precise cell placement. Examples are photoresponsive proteins extracted from living bacteria in which the modulation of the voltage in the gate electrode in light-gated organic field effect transistors becomes crucial for applicative outcomes.¹⁴ Actually, the capability of positioning cells on microelectrodes or specific area of the devices would pave the way for advanced functional integration and improved device performance.^15–18 A widely adopted strategy to precisely move cells on devices involves internalizing or decorating the cell membranes with magnetic nanoparticles (MNPs) and subsequently moving the cells using an external magnet.^19,20 Superparamagnetic iron oxide-based nanoparticles MNPs (SPIONs) are particularly popular in both pre-clinical and clinical applications, such as magnetic hypothermia and magnetic resonance imaging, due to their favorable biocompatibility. Their ability to be magnetized with a low-field magnet and to lose magnetization when no longer need makes them a first choice in tissue engineering.²¹ The preemptive magnetization of cells with SPIONs, enabling their remote manipulation, was demonstrated across various cell types, including cardiomyocytes,²² smooth muscle cells,²³ T lymphocytes²⁴ and neurons.²⁵ However, recent studies have highlighted potential toxicity associated with the internalization or membrane decoration using magnetic nanoparticles, with the severity of these effects largely depending on nanoparticles’ size.^26,27 Critical questions remain regarding the fate of SPIONs after cellular internalization, the degradation process, impacts on cell metabolism and motility, and the effects of reactive oxygen species production.²⁸ These unresolved safety concerns are driving researchers to explore alternative, less invasive approaches to achieve similar outcomes. Diatom microalgae have long fascinated material scientists because they contain C- and Si-machinery for producing mesoporous, nanostructured biosilica shells, called frustules.²⁹ These outer membranes exhibit shapes, size and porous periodicity that vary depending on the microalgal species, offering numerous technological implications. These include natural photonic crystals and lenses for optics,^30,31 drug delivery cargoes,³² carriers for probiotics³³ and biocompatible platforms for tissue engineering.³⁴ Diatom shells can be obtained from living cells using various soft or harsh methods,³⁵ resulting in surfaces with different chemistry and topography that can be functionalized with organosilanes or other functional molecules. Interestingly, few recent works reported the magnetic functionalization of diatom shells with MNPs, for drug delivery^36,37 and exosome isolation³⁸ purpose. Depending on the species, diatoms can provide well-defined, naturally porous, ornate silica platforms suitable for creating living cells carriers characterized by hierarchical architecture of the nano-/microstructures and inherent biocompatibility.^34,39,40 Since it is widely accepted that rough and porous surfaces promote better cell attachment compared to flat or minimally textured surfaces,^7,41–44 we postulated that these natural, decorated structures could offer an ideal substrate for robust cell adhesion. We also sought to test, once cells adhered firmly to the diatom shells, it would be possible to move them spatially without significantly compromising their viability. In this study, we first describe the functionalization of the inorganic shells of Coscinodiscus wailesii diatoms with ferromagnetic nanoparticles, employing a combined electrostatic and covalent modification of the biosilica surface. We then assess the adhesion and viability of neuronal cells on the resulting magnetic diatom shells. Finally, as proof-of-concept of our hypothesis, we demonstrate the mobility of these cell-loaded magnetic shuttles, first using pipette suction, and subsequently through magnetic dragging within a simple microfluidic system.

2. Results & discussion

2.1. Magnetic functionalization of diatom shells

Diatoms are the most prolific source of biosilica in the world. Diatoms cultures easily lead to the production of silica microbeads or boxes with different sizes and shapes according to the species involved. Algal biosilica is organized in a double-valve system, held together by a nanostructured belt (namely girdle). The girdle is connected to the valves through organic material removable by acid, oxidative, acid-oxidative, or detergent treatments. After simple cleaning protocols of the organic part, biosilica can be obtained either as or separated in hemivalves starting from highly dense living giant diatom cells. In this study, we use the silica shells from Coscinodiscus wailesii (Cw) as cell shuttles (Fig. 1), leveraging their large size and distinctive hierarchical porous structure (Fig. 2), which will be discussed in greater detail below. The concave inner and the convex outer surfaces of the siliceous hemivalves of the Cw diatom (Fig. 2A) display exquisite multiscale porous architecture, detailed in Fig. 2B and C. Specifically, the inner side of the Cw diatom (Fig. 2B) features regular holes approximately 1–1.5 μm in diameter, overlaid on a secondary layer of smaller pores measuring few hundred nanometers.

Fig. 1 Sketch of the idea proposed in this work. Multiscale porous diatoms shells can be functionalized with magnetic moieties and used to move multiple cells in space at once.

Fig. 2 Morphology of Coscinodiscus wailesii (Cw) shells used in this study. SEM (A) and (B) and AFM (C) images of the inner (A, left and B) and outer (A, right and C) hemi-valves highlighting the multiscale porous structure of the Cw's shells.

In contrast, the convex outer surface (Fig. 2C) displays a striking star-like arrangement of nanopore clusters. This unique porous architecture, expected in nature to facilitate nutrient absorption, waste elimination and light focusing on photosystems, here attracted our attention since it is hypothesized to enhance neuron cell attachment for their multiscale porosity. Moreover, the large diameter of the diatom shells (100–120 μm) provides enough space to host multiple cells simultaneously. These features were key factors in selecting Cw diatoms for this proof-of-concept study. The separated yet intact diatom hemivalves were functionalized using a straightforward chemical approach. Specifically, ferromagnetic (Fe₃O₄) nanoparticles (MNPs) were synthesized to be used to decorate the diatom shells, thus making them magnetically active and moveable.⁴⁵ The synthesis of MNPs and the characterization of the decorated diatom shells are summarized in Fig. 3.

Fig. 3 Functionalization and characterization of magnetic shells with magnetic nanoparticles. (A) Average size and SEM image (right) of the as-synthesized magnetic NPs; (B) sketch of the functionalization process adopted; (C) optical images of the same diatom shells (top and bottom are different shells) before (left) and after (right) NPs-decoration (scale bar is 50 μm, top; 30 μm, bottom); (D) key frames extracted from the Video S1 (ESI†) to demonstrate the effectiveness of the magnetic functionalization process.

The synthesis method produced well-monodispersed MNPs with an average diameter of 150 ± 20 μm (Fig. 3A). In the synthesis procedure of ferromagnetic nanoparticles, the time of exposure to reducing agents and the time in which reaction quenching occurs are crucial aspects that influence the average size of nanomaterials. In our specific case, the reducing agent was in contact for a few minutes under strong stirring, preventing the formation of large nucleation centers, and the rapid quenching allowed further aggregation to be blocked. In addition, the choice of including a catechol-based capping and the interaction with diamine-impregnated biosilica allowed the nanoparticles to be stabilized, both in terms of size and oxidizability, at the biological interface. The multistep functionalization process used to obtain magnetic diatoms is outlined in Fig. 3B. Biosilica was functionalized with ethylenediamine (EDA) using a simple impregnation procedure already reported in literature for the synthesis of poly-amine/silica hybrid species.⁴⁶ The impregnation mechanism consists of a combination of physical confinement and electrostatic interaction between silica and the protruding amine functional groups. This method results convenient, for functionalization rate and stability, with respect to other multi-step covalent processings.⁴⁷ Amino-functionalized biosilica can be addressed by specifically capped magnetic nanoparticles. Here dopamine-capped nanosized ferromagnetite was prepared exploiting the direct –OH: iron oxide interaction, then the catechol-capped nanoparticles directly react with amine groups of biosilica.⁴⁸ Given the self-polymerization capacity of dopamine in basic aqueous environment, a strategy of closely controlling the exposure time of the frustules to the MNPs was used to obtain magnetized biosilica without the presence of undesired polydopamine aggregates. The different exposure times and qualitative criteria are shown on Table S1 (ESI†). Magnetic recovery and several washing steps were performed to purify the magnetic Cw shells. This procedure was completed prior to the initiation of dopamine polymerization. FTIR-ATR analysis was conducted to identify specific functional groups introduced during the chemical and magnetic decoration of the diatom shells (Fig. S1A, ESI†). The bare biosilica, both before and after functionalization, exhibited a strong Si–O signal peaked at 1185 cm⁻¹ with minimal contributions from organic functional groups. After EDA treatment and magnetic decoration, symmetric and asymmetric –CH stretching signals set at 2998–2985 cm⁻¹ increased, together with an –NH signal (shoulder like, weak, 3321 cm⁻¹) near the –OH peak, and the –NH bending signal set at 1581 cm⁻¹. Additionally, EDX analysis showed a significant increase in iron presence over the surface of the functionalized shells (Fig. S2, ESI†), further supporting the anchoring of the capped MNPs onto the amino-functionalized biosilica. Representative optical images of pristine and magnetically decorated diatoms are shown in Fig. 3C. The surface of the diatom shell decorated with magnetic nanoparticles can be better appreciated from SEM images as the one shown in Fig. S1B (ESI†) (brighter aggregates are the magnetic NPs). Finally, in Fig. 3D, key frames extracted from Video S1 (ESI†), illustrate the effectiveness of the functionalization approach, demonstrating that the magnetic diatom shells can be easily moved by applying a small external magnet.

2.2. Adhesion and viability of N2A cells on magnetic shells

Neuroblastoma N2A cell line was used in this work to demonstrate the possibility to use magnetic diatom shells as cell cargo. Neuroblastoma cells are widely used to study neuronal differentiation and axonal growth on novel biomaterials, signaling pathways and for toxicological studies.^7,49–52 SEM images of the N2A cells adhered to the magnetic shells at 24 h after seeding are shown in Fig. 4. In the very first experiment, magnetic diatom shells were randomly distributed in the cell culture medium, thus the deposed valves randomly exposed the concave or convex sides to the adhering neuron cells. However, as can be seen, cell attachment occurred indistinctly on both sides of the diatoms and even on the lateral edge. It can be noted that, given the large diameter of hemivalves, each shell can accommodate several (ca. 10–15) cells at a time. We believe that this excellent and intimate cell–material interaction can be primarily ascribed to the multiscale porosity of the diatom shells, which seems to enhance the development of protrusion and lamellipodia processes promoting the development of neurofilaments. In addition, positively charged chemical amines, used here as ethylenediamine to coat the surface of the frustules and widely recognized to promote neural cell adhesion, among other human cell types, may have further contributed to promoting cell adhesion to the silica shells.⁵³

Fig. 4 N2A adhesion to the magnetic diatom shells. SEM images aquired at different enlargements of N2A cells at 24 h from seeding adhering on the inner (A) and outer (B) side of the magnetized Cw shell.

The viability of cells cultured on magnetic diatoms was assessed by MTT assay (Fig. 5A). After 24 h, cells showed adequate viability (∼83%) compared with the culture dish control. At 72 h and 5 days, the viability decreased compared with the control, settling at values slightly above 50%. The diatom shells were obtained by extracting biosilica from live Coscinodiscus species cells, which is known to be biocompatible,³³ so we infer that the viability of N2A cells was not affected by the biosilica itself, but rather by the ferromagnetic nanoparticles used to decorate the shells.

Fig. 5 Magnetic diatoms as shuttles for N2A cells. (A) MTT viability assay, p-value = 0.05. (B) Schematic representation of the movement of magnetic diatoms with adhered cells by pipetting. Diatoms are drop-casted in the well (i) and (iv); then cells are seeded (ii) and (v). After 24 hours diatoms and cells growing on them are resuspended and transferred in a new well (iii) and (iv). Optical micrographs (40×) show diatom as drop casted in the well (iv); N2A cells adhered on a diatom shell and on the well surface after 24 hours of culture (v) and after being relocated in a new well by pipetting (vi). Scale bar: 50 μm (C) SEM images of N2A adhered on diatom shells after micropipette relocation at 24 hours. 4000×.

With the aim of using magnetic shells as cell shuttles, we first tested cell viability after moving the cells and diatoms from the original well to a new one (Fig. 5B). The diatoms were dropped into the well, then N2A cells were seeded and allowed to grow for the desired period of time, after which the diatoms were resuspended in the culture medium by gentle mechanical bumps. The medium containing the suspended diatoms was then aspired with a micropipette and moved to an adjacent empty well. In this way, only the cells attached to the diatoms were moved to the new well, while those adhering to the plastic surface of the well were not transferred. According to the results of MTT assay, relocating the diatoms and the cells growing on them by micropipette aspiration did not affect the cell viability if the latter were allowed to adhere for 24 hours (Fig. 5A).

It is noteworthy that, prior to pipetting, N2A cells could be observed (after 24 hours) both on the plastic surface of the well and on the diatom shells (Fig. 5B_v). However, following pipetting, cells were found exclusively adhered to the diatom shell (Fig. 5B_vi). This suggests that after 24 hours of growth on the diatom surface, cells could be mechanically relocated with minimal impact on their viability. SEM analysis confirmed that N2A cells remained tightly attached to the diatom surface even after being transferred to another well (Fig. 5C). Additionally, the cells maintained normal morphology, further supporting the idea that displacement did not negatively affect their viability. These results indicate that 24 hours is sufficient for the cells to establish intimate contact and strong adhesion to of the diatom shells, enabling them to be transferred along their diatom cargoes. In contrast, when cells were relocated after 72 hours or 5 days, optical absorption falls below the detection limit, indicating that few or no cells were successfully transferred to the new well. We hypothesize that, at longer adhesion time the cells form robust connections with both the silica shell and the underlying well surface, resulting in the diatoms becoming strongly anchored (Fig. S3 and S4, ESI†). Consequently, attempts to resuspend the diatoms likely disrupted the plasma membrane of the cells, leading to their death.

2.3. Movement of cell-loaded magnetic shells

Having established that 24 hours of adhesion is the optimal time to move diatoms without compromising cell viability, we further explored the feasibility of using the diatom's magnetic properties for cell transport. Initially a magnetic field was applied to the sample and diatom movement was observed under an optical microscope (Fig. 6A and Videos S2–S4, ESI†). N2A cells were cultured on diatoms for 24 hours and then resuspended as previously described. When a small NdFeB magnet (20 × 5 × 5 mm) was placed near the sample and moved around the edge of the culture dish, the diatoms were observed to twirl and follow the magnet's motion. Cells adhered to diatoms moved along with their cargoes, while, as expected, cells attached to the plastic substrate remained unaffected by the magnetic field. We further investigated the spatial control of cell-decorated magnetic diatoms using the simple fluidic shown in Fig. 6B, consisting of two identical wells connected by a channel (Fig. 6C_i). Each well has a diameter of 15 mm, while the channel is 10 mm long and 4 mm wide (Fig. 6C_ii). Diatoms and cells were initially seeded in the first well (reservoir, Fig. 6D_i) and subsequently dragged to the second well through the channel using magnetic fields (Fig. 6D_ii). N2A cells were seeded on diatoms and allowed to grow for 24 hours with a magnet placed under the reservoir well to prevent accidental flow of cell-decorated diatoms into the channel. To initiate the dragging process, the magnet was positioned beneath the channel, and it was later positioned beneath the second well. This setup allowed the magnetic diatom shuttles to move along the magnetic field gradient, reaching the collecting well (Fig. 6D_iii). As observed in the pipetting experiments, N2A cells in the reservoir well were found both on diatom shuttles and on the plastic surface of the culture dish (Fig. 6D_iv). However, after being dragged to the collecting well, cells were observed exclusively on the diatom surfaces, confirming that only cells adhered to magnetic diatoms were successfully transported (Fig. 6D_v and vi).

Fig. 6 Magnetic movement of the cell-loaded shells. (A) Key frames extracted from Video S4 (ESI†) demonstrate the ability to spin cell-laden shells under the effect of an external magnet. (B) Picture of the fluidic system. (C) 3D rendering of the fluidics device used for the experiment (i) and schematic of the wells and the channel (ii, quotes are given in cm). (D) Schematic of the fluidic experiment. Diatoms and cells are seeded in the reservoir (i) and then dragged to the channel (ii) to be finally collected in the collection well (iii). Optical images show cells growing both in diatoms and on the surface of the Petri dish in the reservoir (iv, 4×, scale bar: 200 μm) and cells growing solely on the shells in the collecting well (v and vi, 20×, scale bar: 50 μm).

These experimental results, albeit simple, demonstrate the potential of magnetic diatoms as shuttle platforms for spatially controlled cell transport, enabling the placement of cells at desired locations using an external magnetic field. This represents an important first step toward designing strategies for precise and simultaneous displacement of multiple cells leaving them intact and functional. In this approach, the size and, to some extent, the geometry of the chosen diatom are expected to influence the number of cells adhering to each shell. Larger valves can carry more cells but require stronger magnetic fields for movement, while smaller valves, though easier to manipulate, transport fewer cells at a time. For this preliminary study, we selected relatively large diatoms to demonstrate the ability to transport multiple cells in one shot. Depending on the intended application, different diatom shells could be chosen to balance these factors. For instance, Coscinodiscus radiatus frustules, a species with a much smaller average diameter than Coscinodiscus wailesii (in the 10–15 μm range), could accommodate 1 or 2 cells at a time, allowing the transition from a cell cluster paradigm to a single-cell shuttling paradigm. A major limitation of this method is that the silica shell cannot be easily removed from the system without compromising cell viability. As a result, shells are expected to remain in place for a long time and not degrade rapidly.^54–56 However, this characteristic makes the approach particularly suited for at least three main application areas. The first is regeneration of hard tissues, such as compact bone, where slowly or even not degrading scaffolds (made of Ta or Ti alloys, sintered hydroxyapatite or other crystalline calcium phosphates) but showing robust mechanical and load-bearing properties are advantageous.⁵⁷ In this case, compact constructs consisting of cell-loaded diatom shells can be envisioned, with the additional advantages of carrying cells ready to rapidly colonize adjacent areas, proliferate and/or differentiate in situ, and exhibiting a multiscale porous architecture that facilitates rapid vascularization of the material. The second potential application lies in the field of biosensors, where magnetic diatom shuttles enable the precise placement of cells onto electrodes for label free sensors or wells for optical and optoelectronic array sensors. In the case of neuronal cells, this could facilitate the development of biohybrid and neuromorphic devices^58,59 allowing cells to communicate with the environment through electrical and ionic signals while transferring this information to the underlying electronic system. Additionally, other potential applications could involve facilitating the study of single cells by enabling the spatial isolation of one or more supported cells on a single frustule using a magnetic field. This approach could be especially valuable if pluripotent cells are used instead of neurons, as they have the potential to initiate an entire cell culture in a designated region within a 3D culture. Furthermore, it could allow for targeted treatment, analysis, and study of single cells (or small clusters of cells), including advanced techniques such as DNA/RNA sequencing or Omics analysis, which remain significant challenges in cell research.^60,61 Finally, as a future prospect, a direct selection of less aggressive reducing agents, capping molecules belonging to biological redox chains (e.g., quinones, natural organic diacids), and non-synthetic amines from nature will be carried out. In this way, the tuning of nanoparticle size and chemistry will be carefully controlled in a completely sustainable manner.

3. Conclusions

In this study, we demonstrated the ad hoc displacement and placement of cells without compromising their viability through their adhesion to diatom hemivalves that are used as shuttles. Magnetized diatom shells derived from the uniquely shaped Coscinodiscus wailesii species were utilized, leveraging their inherent biocompatibility, high porosity (which promotes cell adhesion), and a large surface area (enabling the transport of multiple cells simultaneously). Through simple pilot experiments, we demonstrated the ability to translate and relocate multiple neuronal cells simultaneously without significantly impact on their viability. This proof-of-concept study introduces new possibilities for the development of advanced 3D biomaterials, biohybrid devices and platforms for studying cell–cell interaction and cellular network formation, where cells can be precisely and effortlessly positioned as needed.

4. Experimental section

4.1. Diatoms culture conditions

Coscinodiscus wailesii (CCAP strain 1013/9) cells were grown in F/2 Guillard prepared with sterile natural sea water, as unstirred suspension cultures in polystyrene flasks (250 mL). Natural sea water was sterilized in autoclave and filtered twice (0.22 μm ∅), beforehand. Culture medium was buffered with 1 mM NaCO₃ at pH 7.8. Cultures were grown until plateau phase was reached, under conditions optimized for the obtention of diatoms suited for technological applications (T: 18 ± 2 °C; relative humidity of 65%; light/dark cycle of 16/8 h, white light irradiation of 40 μmol m⁻² s⁻¹).⁶²

4.2. Cleaning procedure to obtain pristine biosilica shells

Cells from Cw cultures in plateau phase (reached ∼6 weeks after inoculation), were collected 24 hours before the cleaning procedures via natural settling in conic falcon (under dark conditions, overnight), avoiding detrimental physical stress that centrifugation or filtering might inflict on frustules. An extraction method, already reported in literature, was then performed: a high-density pellet (1 mL, ∼300 cells μL⁻¹) of settled living diatoms was treated with H₂SO₄ and H₂O₂ (final concentration of 0.25 M and 5 w/w%, respectively) at 70 °C for at least 2 h, until the disappearance of the characteristic greenish-brown coloration from photosynthetic pigments. Then silica shells were washed 3 times in deionized water (dH₂O) and collected via settling.³⁵

4.3. Synthesis of magnetic nanoparticles

The procedure for ferromagnetite nanoparticles has been adapted from literature.⁶³ An aqueous solution of FeCl₃·6H₂O (1.0 equiv.) was added swiftly to a pre-cooled (0 °C) EtOH solution of NaBH₄ (6.0 equiv.), and the solution was stirred at RT for at least 30 min. After the reaction, water was added in excess to quench the reducing agent, and the magnetic particles were recovered through magnetic separation. The black crude was washed several times with water, roughly separated via centrifugation to remove big sized particles and nanosized ferromagnetite was characterized using scan electron microscopy and dynamic light scattering spectroscopy.

4.4. Chemical decoration of diatom shells

Coscinodiscus wailesii, due to its regularity in the silica substructure and surface texture, together with its biological resistance, was chosen as silica source. Pristine Cw shells were collected from living Cw cells as explained above, prior to being magnetically decorated, following an adapted procedure from literature.⁶⁴ A simple impregnation and bridging protocol was performed to superficially decorate Cw silica shells with magnetic nanoparticles. Briefly, 10 μL of an EDA stock solution were added to a suspension of cleaned shells (2.0 mL; ∼30 shells μL⁻¹) in dH₂O (final EDA concentration of 0.075 M); the electrostatic bridging reaction was lightly stirred for 15 min under inert atmosphere, yielding suspension A. Separately, a previously sonicated magnetic nanoparticle suspension (0.5 mg MNPs in 100 μL; 30 min) was added dopamine powder to a final concentration of 65 mM, yielding suspension B. Then, suspension B was added to suspension A, bringing together the catechol-capped MNPs and the EDA-coated biosilica shells; allowing for the electrostatic bridging, and the conclusion of the dopamine polymerization to take place for another 30 min. Finally, the resulting magnetic shells were purified following several washing steps with dH₂O, using a NdFeB magnet.

4.5. Characterization of magnetic diatom shells

Fourier transform infrared-attenuated total reflectance (FTIR-ATR) spectroscopy has been used to investigate functional groups belonging to the chemical bulks of the different samples. A PerkinElmer two spectrophotometer equipped with A 2 × 2 mm diamond crystal has been exploited. After drying 0.2 g of each sample, 2.5 mg of sub-samples have been spread over diamond surface and analysed (4000–400 cm⁻¹ range, 4 cm⁻¹ resolution, 32 scan, 2 cm s⁻¹ rate). Each recorded scan set was averaged, corrected against ambient air as background and treated with ATR correction correlative function. To highlight the emerging bands, the smoothing function of the workstation was not needed. Dynamic light scattering (DLS) measurements on nanoparrticles were performed with a Nanosizer ZS (Malvern instruments) for the determination of the size distribution of particles suspended in T₂₅. DLS measurements were performed in backscattering mode at pre-fixed detector angle. Samples for SEM analysis were obtained by drop-casting 100 μl of diatoms suspension on a glass coverslip and allowing to dry overnight at room temperature. Samples were fixed on metal stubs and coated with a thin layer (∼10 nm) of gold. Gold was sputtered using a Q150RS DC magnetron sputter (Quorum Tech, London, UK) operated in rarefied argon atmosphere (10⁻¹ mbar, 18 mA, 1 minute). Scanning electron micrographs were collected using a ZEISS GeminiSEM 460 (Carl Zeiss AG, Oberkochen, Germany). Energy dispersive X-ray (EDX) spectroscopy was carried out to determine the presence of characteristic elements before and after the functionalization of the shells using a FEG SEM, Zeiss (3 keV, 0° tilting function, no mapping mode activated). Diatoms, both bare and decorated with cells, were investigates by atomic force microscopy (AFM) using a Park XE7 AFM system (Park System, Suwon, Republic of Korea) operated in air and at room temperature in tapping mode and using silicon cantilever with Al backside reflective coating; typical tip curvature radius ca. 7 nm, elastic constant ca. 26 N m⁻¹ and resonance frequency around 300 Hz (OMCL-AC160TS, Olympus Micro Cantilevers, Tokyo, Japan). AFM topography images were processed and analyzed using Park System XEI software (Park System, Suwon, Republic of Korea).

4.6. N2A cell culture

Mouse Neuro2A (N2A) neuroblastoma cell line (Tema Ricerca, Bologna, Italy) was used to assess the availability of diatoms as substrate to support cell growth and targeted displacement. Cells were cultured in Dulbecco's modified Eagle medium (DMEM, Fisher Scientific, MA, USA) supplemented with 10% festal bovine serum (FBS, Fisher Scientific, MA, USA), 200 μg mL⁻¹L-glutamine (Sigma Aldrich, MO, USA) and 200 μg mL⁻¹ penicillin/streptomycin (Sigma Aldrich, MO, USA), and were incubated at 37 °C with 5% CO₂ and ∼90% humidity. Depending on the purpose of the experiments, the desired amount of diatom suspension was drop-casted on the substrate (glass coverslip, 96-wells plate or Petri dish with integrated fluidic system) and then sterilized under UV-light (260–265 nm) for 30 minutes and then rinsed with sterile phosphate buffered saline (PBS) solution (pH 7.4) 1× (Sigma Aldrich, MO, USA). For all the experiments, cells were seeded at a density of 1 × 10⁴ cell per mL and cultured for 1, 3 and 5 days, refreshing cell culture medium every 24 hours. For SEM and AFM analysis cells were cultured on diatoms using a glass coverslip as support and fixed at different time points (24 hours, 72 hours and 5 days). Cells were fixed in a 4% glutaraldehyde solution in phosphate saline buffer (PBS) at pH 7.4 and then dehydrated with ascending concentrations of ethanol. Ethanol was allowed to evaporate overnight at room temperature.

4.7. N2A cell viability

Assessment of mitochondrial activity was used to evaluate viability of cells seeded on magnetic diatoms shuttles through the reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT, Thermo Fisher Scientific, MA, USA) to formazan salt (MTT assay). Living cells can convert MTT into formazan salt thanks to redox reactions happening in mitochondria, with the concentration of formazan salt being proportional to the mitochondrial activity. Formazan salt gives a purple shade to the medium that is quantified through absorption at 560 nm. Briefly, cells were seeded on diatoms at a density of 1 × 10⁴ cell per mL in a 96-wells culture plate and cultured for 24 hours, 72 hours and 5 days, refreshing the medium every 24 hours. Prior to seeding cells, diatoms were drop-casted in each well and sterilized as described above. In the same culture plate, an independent set of N2A cells were seeded in wells without diatoms at a density of 1 × 10⁴ cell per mL and used as control. At the end of the desired time point, 5 mg mL⁻¹ MTT solution in PBS was added to each well and allowed to incubate for 4 hours at 37 °C in a 5% CO₂ atmosphere. The medium was removed and 100 μl of dimethyl sulfoxide (DMSO, 99.99% pure, Sigma Aldrich, MO, USA) were added to each well to dissolve the formazan salt by shaking the plate for 45 minutes at 37 °C. The plate was read using a multimodal plate reader spectrophotometer (VICTOR 4×, PerkinElmer, MA, USA) at a wavelength of 560 nm. For each time point cell viability was evaluated before and after relocating diatoms and cells growing on them. To do so, cells and diatoms were resuspended by gentle mechanical bump. The medium containing the resuspended diatoms was aspired with a micropipette and deposited in an empty well of the 96-well cultured plate and then treated with the MTT solution as just described.

4.8. Displacement of cell-loaded magnetic diatoms

The ability of magnetic diatoms to act as a “shuttle” for cells growing on their surface was assessed by dragging the shells with N2As grown on them. Diatoms were drop casted in a 6-well culture plate and sterilized as described above, then 1 × 10⁴ cells per mL were seeded in the wells and allowed to grow for 24 hours and 72 hours, refreshing the medium every 24 hours. After the desired experimental time, diatoms and cells adhered on their surface were resuspended by gentle mechanical bumps. The medium with the resuspended diatoms was collected and transferred to a Petri dish (diameter: 35 mm). The sample was observed under an optical microscope while moving a NdFeB magnet (20 × 5 × 5 mm, attraction force: 24.5 N) around the edge of the Petri dish, paying attention not to touch the outer wall to avoid displacement of the diatoms ascribable to mechanical collision. Furthermore, the possibility to precisely control the displacement of the magnetic diatoms was preliminary investigated by a simple “fluidics” experiment. The fluidics was designed to feature two wells, one acting as cell reservoir (where cells and diatoms were seeded) and the other as collecting well, connected by a straight channel 1 cm long. The fluidic was fabricated by printing the mold with a 3D printer and by replicating it with a biphasic silicon rubber (Silicone RPRO 30) that was cured overnight at 60 °C. The silicon rubber mold was sterilized in 70% v/v ethanol/MilliQ water solution, dried under laminar flow hood and then placed into a 35 mm Petri dish. The silicon rubber mold adhered to the plastic surface of the dish, creating a simple fluidics device. A limited volume of diatoms suspension was drop casted in the seeding well (diameter: 15 mm) and sterilized as described above, then the fluidics was filled with culture medium and a defined volume of cell suspension was added in the seeding well (1 × 10⁴ cells per mL). The fluidics was incubated either for 24 hours or 72 hours (in this case the medium was refreshed every 24 hours), placing a NdFeB magnet (20 × 5 × 5 mm) in correspondence of the cell reservoir to gather most of the magnetic diatoms. After 24 hours in one case, and after 72 hours in the other case, the magnet was moved and placed beneath the channel to drag diatoms from the reservoir to the channel. After 5 hours the magnet was moved again and placed in correspondence of the collecting well where it stayed overnight. Each step was monitored by optical microscope imaging; the fluidics was incubated at 37 °C in 5% CO₂ atmosphere and 90% humidity for the entire duration of the experiment.

Author contributions

Conceptualization: DV and MB; data curation: AL, DV, CVG and MB; formal analysis and investigation: AL, DV, CVG, SRC; methodology: AL, MB and DV; supervision: DV and MB; funding acquisition: GF, FB; writing original draft: AL, DV, CG, MB, SRC; writing – review & editing: MB, DV, FB, GF.

Data availability

The data supporting this article has been included as part of the ESI.†

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

This work is related to Progetto “Hub scienze della vita della Regione Puglia” (codice progetto T4-AN-01 e codice CUP H93C22000560003) Azione 4.1, Piano Operativo Salute, 2014–2020, Ministero della Salute.

References

1. M. Peng, Q. Zhao, M. Wang and X. Du, Nanoscale, 2023, 15, 6105.
2. J. M. Lee and W. Y. Yeong, J. R. Soc., Interface, 2020, 17, 20200294.
3. A. Rossi, F. Furlani, G. Bassi, C. Cunha, A. Lunghi, F. Molinari, F. J. Teran, F. Lista, M. Bianchi, A. Piperno, M. Montesi and S. Panseri, Mater. Today Bio, 2024, 27, 101110.
4. F. Valle, B. Chelli, M. Bianchi, P. Greco, E. Bystrenova, I. Tonazzini and F. Biscarini, Adv. Eng. Mater., 2010, 12, B185.
5. N. Demri, S. Dumas, M. Nguyen, G. Gropplero, A. Abou-Hassan, S. Descroix and C. Wilhelm, Adv. Funct. Mater., 2022, 32, 2204850.
6. R. Plen, A. Smith, O. Blum, O. Aloni, U. Locker, Z. Shapira, S. Margel and O. Shefi, Adv. Funct. Mater., 2022, 32, 2270285.
7. M. Bianchi, S. Guzzo, A. Lunghi, P. Greco, A. Pisciotta, M. Murgia, G. Carnevale, L. Fadiga and F. Biscarini, ACS Appl. Mater. Interfaces, 2023, 15, 59224.
8. I. Constantinou and E. E. Bastounis, Trends Biotechnol., 2023, 41, 939.
9. A. Persaud, A. Maus, L. Strait and D. Zhu, Eng. Regener., 2022, 3, 292.
10. S. Liu, C. Gao and F. Peng, Mater. Today Adv., 2022, 16, 100281.
11. E. V. Batrakova, H. E. Gendelman and A. V. Kabanov, Expert Opin. Drug Delivery, 2011, 8, 415.
12. Y. Wu, Y. Liu, T. Wang, Q. Jiang, F. Xu and Z. Liu, Eng. Regener., 2022, 3, 131.
13. F. Milos, G. Tullii, F. Gobbo, F. Lodola, F. Galeotti, C. Verpelli, D. Mayer, V. Maybeck, A. Offenhäusser and M. R. Antognazza, ACS Appl. Mater. Interfaces, 2021, 13, 23438.
14. M. Di Lauro, S. La Gatta, C. A. Bortolotti, V. Beni, V. Parkula, S. Drakopoulou, M. Giordani, M. Berto, F. Milano, T. Cramer, M. Murgia, A. Agostiano, G. M. Farinola, M. Trotta and F. Biscarini, Adv. Electron. Mater., 2020, 6, 1900888.
15. J. A. Goding, A. D. Gilmour, U. A. Aregueta-Robles, E. A. Hasan and R. A. Green, Adv. Funct. Mater., 2018, 28, 1702969.
16. A. Lunghi, A. Mariano, M. Bianchi, N. B. Dinger, M. Murgia, E. Rondanina, A. Toma, P. Greco, M. Di Lauro, F. Santoro, L. Fadiga and F. Biscarini, Adv. Mater. Interfaces, 2022, 9, 2200709.
17. K. Czöndör, M. Garcia, A. Argento, A. Constals, C. Breillat, B. Tessier and O. Thoumine, Nat. Commun., 2013, 4, 2252.
18. M. Shein, A. Greenbaum, T. Gabay, R. Sorkin, M. David-Pur, E. Ben-Jacob and Y. Hanein, Biomed. Microdevices, 2009, 11, 495.
19. M. E. Gomes and R. M. A. Domingues, Adv. Funct. Mater., 2022, 32, 2211925.
20. Y. Pan, X. Du, F. Zhao and B. Xu, Chem. Soc. Rev., 2012, 41, 2912.
21. R. Lapusan, R. Borlan and M. Focsan, Nanoscale Adv., 2024, 6, 2234.
22. L. Zwi-Dantsis, B. Wang, C. Marijon, S. Zonetti, A. Ferrini, L. Massi, D. J. Stuckey, C. M. Terracciano and M. M. Stevens, Adv. Mater., 2020, 32, 1904598.
23. B. M. Mattix, T. R. Olsen, M. Casco, L. Reese, J. T. Poole, J. Zhang, R. P. Visconti, A. Simionescu, D. T. Simionescu and F. Alexis, Biomaterials, 2014, 35, 949.
24. L. Sanz-Ortega, J. M. Rojas, A. Marcos, Y. Portilla, J. V. Stein and D. F. Barber, J. Nanobiotechnol., 2019, 17, 14.
25. M. Marcus, M. Karni, K. Baranes, I. Levy, N. Alon, S. Margel and O. Shefi, J. Nanobiotechnol., 2016, 14, 37.
26. S. Chung, R. A. Revia and M. Zhang, Nanoscale Horiz., 2021, 6, 696.
27. J.-E. Bae, M.-I. Huh, B.-K. Ryu, J.-Y. Do, S.-U. Jin, M.-J. Moon, J.-C. Jung, Y. Chang, E. Kim, S.-G. Chi, G.-H. Lee and K.-S. Chae, Biomaterials, 2011, 32, 9401.
28. A. Van De Walle, J. E. Perez, A. Abou-Hassan, M. Hémadi, N. Luciani and C. Wilhelm, Mater. Today Nano, 2020, 11, 100084.
29. R. Ragni, S. R. Cicco, D. Vona and G. M. Farinola, Adv. Mater., 2018, 30, 1704289.
30. M. L. Presti, R. Ragni, D. Vona, G. Leone, S. Cicco and G. M. Farinola, MRS Adv., 2018, 3, 1509.
31. D. Vona, R. Ragni, E. Altamura, P. Albanese, M. M. Giangregorio, S. R. Cicco and G. M. Farinola, Appl. Sci., 2021, 11, 3327.
32. D. Vona, A. Flemma, F. Piccapane, P. Cotugno, S. R. Cicco, V. Armenise, C. Vicente-Garcia, M. M. Giangregorio, G. Procino and R. Ragni, Mar. Drugs, 2023, 21, 438.
33. D. Vona, S. R. Cicco, F. M. La Forgia, M. Vacca, A. Porrelli, G. Caggiano, M. De Angelis, L. Gesualdo and G. M. Farinola, Adv. Sustainable Syst., 2024, 8, 2400384.
34. S. R. Cicco, D. Vona, G. Leone, E. De Giglio, M. A. Bonifacio, S. Cometa, S. Fiore, F. Palumbo, R. Ragni and G. M. Farinola, Mater. Sci. Eng., C, 2019, 104, 109897.
35. D. Vona, L. Urbano, M. A. Bonifacio, E. De Giglio, S. Cometa, M. Mattioli-Belmonte, F. Palumbo, R. Ragni, S. R. Cicco and G. M. Farinola, Data Brief, 2016, 8, 312.
36. M. Li, J. Wu, D. Lin, J. Yang, N. Jiao, Y. Wang and L. Liu, Acta Biomater., 2022, 154, 443.
37. T. Todd, Z. Zhen, W. Tang, H. Chen, G. Wang, Y.-J. Chuang, K. Deaton, Z. Pan and J. Xie, Nanoscale, 2014, 6, 2073.
38. T. N. T. Dao, A. Seshadri Reddy, F. Zhao, H. Liu, B. Koo, M. Moniruzzaman, J. Kim and Y. Shin, ACS Sustainable Chem. Eng., 2021, 9, 3439.
39. G. Leone, D. Vona, E. De Giglio, M. A. Bonifacio, S. Cometa, S. Fiore, F. Palumbo, R. Ragni, G. M. Farinola and S. R. Cicco, Data Brief, 2019, 24, 103831.
40. S. R. Cicco, D. Vona, E. De Giglio, S. Cometa, M. Mattioli-Belmonte, F. Palumbo, R. Ragni and G. M. Farinola, ChemPlusChem, 2015, 80, 1104.
41. F. Gentile, L. Tirinato, E. Battista, F. Causa, C. Liberale, E. M. Di Fabrizio and P. Decuzzi, Biomaterials, 2010, 31, 7205.
42. B. Majhy, P. Priyadarshini and A. K. Sen, RSC Adv., 2021, 11, 15467.
43. G. Di Pompo, A. Liguori, M. Carlini, S. Avnet, M. Boi, N. Baldini, M. L. Focarete, M. Bianchi, C. Gualandi and G. Graziani, Biomater. Adv., 2023, 144, 213231.
44. C. Dionigi, M. Bianchi, P. D’Angelo, B. Chelli, P. Greco, A. Shehu, I. Tonazzini, A. N. Lazar and F. Biscarini, J. Mater. Chem., 2010, 20, 2213.
45. G. Buscemi, M. Trotta, D. Vona, G. M. Farinola, F. Milano and R. Ragni, Bioconjugate Chem., 2023, 34, 629.
46. K. Li, J. Jiang, S. Tian, F. Yan and X. Chen, J. Mater. Chem. A, 2015, 3, 2166.
47. S. Cohen, I. Chejanovsky and R. Y. Suckeveriene, Nanomaterials, 2022, 12, 2237.
48. S. Suárez-García, J. Sedó, J. Saiz-Poseu and D. Ruiz-Molina, Biomimetics, 2017, 2, 22.
49. L. Chen, P. Feng, X. Zhu, S. He, J. Duan and D. Zhou, J. Cell. Mol. Med., 2016, 20, 2102.
50. E. Aydın, H. Türkez and M. S. Keleş, Toxicol. Ind. Health, 2015, 31, 764.
51. H. Turkez, B. Togar, A. Tatar, F. Geyıkoglu and A. Hacımuftuoglu, Biologia, 2014, 69, 936.
52. S.-J. Park, M. L. Jin, H.-K. An, K.-S. Kim, M. J. Ko, C. M. Kim, Y. W. Choi and Y.-C. Lee, Neurosci. Lett., 2015, 588, 101.
53. P. Roach, T. Parker, N. Gadegaard and M. R. Alexander, Surf. Sci. Rep., 2010, 65, 145.
54. S. Kang, S. Hwang, H. Cheng, S. Yu, B. H. Kim, J. Kim, Y. Huang and J. A. Rogers, Adv. Funct. Mater., 2014, 24, 4427.
55. J.-K. Chang, H. Fang, C. A. Bower, E. Song, X. Yu and J. A. Rogers, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, E5522.
56. M. Schumacher, P. Habibovic and S. Van Rijt, Bioact. Mater., 2021, 6, 1921.
57. S. S. Lee, X. Du, I. Kim and S. J. Ferguson, Matter, 2022, 5, 2722.
58. C. Ausilio, C. Lubrano, D. Rana, G. M. Matrone, U. Bruno and F. Santoro, Adv. Sci., 2024, 11, 2305860.
59. K. Kim, M. Sung, H. Park and T. Lee, Adv. Electron. Mater., 2022, 8, 2100935.
60. Y. Ding, Y.-Y. Peng, S. Li, C. Tang, J. Gao, H.-Y. Wang, Z.-Y. Long, X.-M. Lu and Y.-T. Wang, Cell Biochem. Biophys., 2024, 82, 329.
61. C. Liu, Y. Chen, R. Tong, Z. Wang, D. Zhang, H. Chen and P. Zhang, Adv. Funct. Mater., 2024, 34, 2316357.
62. G. Della Rosa, D. Vona, A. Aloisi, R. Ragni, R. Di Corato, M. Lo Presti, S. R. Cicco, E. Altamura, A. Taurino, M. Catalano, G. M. Farinola and R. Rinaldi, ACS Sustainable Chem. Eng., 2019, 7, 2207.
63. S. Wang, J. Fu, K. Wang, M. Gao, J. Zhu, X. Wang and Q. Xu, Mater. Lett., 2017, 209, 576.
64. D. Losic, Y. Yu, M. S. Aw, S. Simovic, B. Thierry and J. Addai-Mensah, Chem. Commun., 2010, 46, 6323.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5tb00352k

---