Titanate nanoribbon-based nanobiohybrid for potential applications in regenerative medicine

Nanoparticles capable of mimicking natural tissues represent a major technological advancement in regenerative medicine. In this pilot study, the development of a new nanohybrid composed of titanate nanoribbons to mimic the extracellular matrix is reported. During the first phase, nanoribbons were synthesized by hydrothermal treatment. Subsequently, titanate nanoribbons were functionalized by heterobifunctional polyethylene-glycol (PEG) to graft type I collagen on their surface. Biological properties of this new nanobiohybrid such as cytotoxicity to cardiac cells and platelet aggregation ability were evaluated. The so-formed nanobiohybrid permits cellular adhesion and proliferation favoring fine cardiac tissue healing and regeneration.


Introduction
Recently, the market for tissue engineering has reached great heights, with key applications being clinical therapies and tissue modeling. [1][2][3][4] Other applications of tissue engineering are personalized and regenerative medicine, 4 cell-based biosensors etc. 5,6 A market forecast of surpassing $4.8 billion is being predicted by the year 2028, for a total value of tissue-engineered products with upcoming technologies such as 3D bioprinting and articial intelligence, 7 decellularization and recellularization of organs and electrospinning. 8 Regenerative medicine and tissue engineering aim to develop biological substitutes for the reconstruction of damaged tissues and the development of new healthy tissues. Currently, the main agents used for tissue regeneration are composed of proteins from the extracellular matrix (ECM). 9,10 These molecules arrange themselves in a unique and complex 3D structure having structural and biological properties similar to tissue sources. Polymer networks are also used for tissue reconstruction and regenerative medicine. 1 These structures enable ber-based self-regeneration of tissue while limiting the risks of rejection and disease transmission strongly present in a transplant tissue from a healthy part of the body. Generally, these networks are made of poly nanobers(lactic-co-glycolic acid) (PLGA), 11 poly(L-lactic acid) (PLLA), poly(caprolactone) (PCL), 12 poly(ethylene oxide) (PEO) also known as polyethylene glycol (PEG), 13 poly(vinyl alcohol) (PVA), 14 poly(ester urethane) urea (PEUU), gelatin, collagen, protein or brinogen.
The use of products based on biological components of human or animal origin, or polymer-based regenerative medicine or tissue reconstruction is not without risk. Those bioproducts oen cause side reactions, embolisms, allergies, infections, damage to nerves and tissue, transmission of diseases and viruses, the formation of blood clots, toxic and anaphylactic shock, necrosis of certain tissues etc. 15,16 Commercial products made of polymers may generate, in turn, the compression of the nerves, tissue damage, inammation, kidney and neurological damage, cerebrospinal uids leaks, etc. 17,18 Another approach has also been proposed to use inorganic bers such as carbon nanotubes and nanobers, 19,20 TiO 2 nanobers or nanotubes 21,22 or gold nanowires 23,24 for tissue reconstruction and regenerative medicine. 25 However, such structures are cylindrical in shape, with dimensions (length and width) of relatively low values.
A need therefore exists for a new material for use in regenerative medicine and tissue engineering which does not have the drawbacks listed above.
Since the observation of carbon nanotubes by Iijima in 1991, 26 the tubular morphology has been thoroughly studied, 27 especially those derived from titanium oxides. [28][29][30] On the other hand, alternative bidimensional morphologies such as nanoribbons were rarely studied and were most oen considered as undesired by-products of synthesis or as an intermediate step of the nanotubes formation. 31 Nevertheless, for about eenyears, their specic properties have enabled titanate nanoribbons to become a full-edged recognized nanostructure more and more studied by scientists. 32,33 Titanate materials are of great interest in regenerative medicine due to their mechanical properties and high resistance to corrosion. Their surfaces are covered with hydroxyl groups that offers the possibility, compared to polymer-based products or biological components of human or animal origin, of functionalizing them with active molecules to couple, for instance, therapeutic and diagnostic effects 34 for tissue regeneration. Furthermore, thanks to their radiosensitizing effect, titanates can also enhance any radiotherapeutic treatment. 35 Elongated shapes of TiO 2 nanomaterials such as titanate nanotubes (TiONTs) showed promising healing properties supported by a promotion of cell growth and proliferation thanks to TiONTs matrix. 36 Furthermore, in implants, TiONTs or TiO 2 are also providing photocatalytic activities enhancing their antibacterial properties. 37,38 These properties can also reduce low prolonged inammation reactions once implanted especially when coated with biopolymer such as collagen. 39 Collagen has signicant applications in tissue engineering. Owing to its excellent biocompatibility, biodegradability, facile extraction process, weak antigenicity and purication, scientic exploration concerning collagen have inspired the eld of tissue engineering. 40 In this pilot study, TiO 2 -based materials for regenerative medicine and tissue engineering were studied. Most particularly titanate nanoribbons (TiONRs) nanostructure with low cytotoxicity, functionalized with biocompatible active polymers and structural adhesion proteins (type I collagen) was developed. In this study, with a focus on cardiac damages, preliminary aggregation and adhesion measurements on broblasts demonstrated the particular interest of functionalized TiONRs to promote healing processes and regeneration of damaged heart's tissues.

Chemicals
All chemicals and reagents were of analytical grade and used without further purication. Methoxy polyethylene glycol 5000 g mol À1 (mPEG 5000 from Sigma Aldrich) and NHS-PEG 5000 -OH (JenKem Technology) were silanized to obtain mPEG 5000 -Si (see protocol SI_1) and NHS-PEG 5000 -Si respectively.

Synthesis of TiONRs
The synthesis of TiONRs has been previously described. 41 Briey, titanate nanoribbons were synthesized by a hydrothermal treatment in strongly basic conditions. For this reaction, 110 mL of NaOH aqueous solution at 10 mol L À1 was prepared and introduced into a sealed Teon reactor. Then 440 mg of TiO 2 precursor (P25 Degussa) was added to the solution and the mixture underwent pulsed ultrasound treatment for 30 min at a power of 375 W (Sonics Vibra-Cells). The hydrothermal treatment took place at 180 C with an autogenic pressure (7 bar), for 20 hours and under low mechanical stirring (150 rpm). The precipitate obtained at the end of the reaction was separated from the synthesis supernatant by a centrifugation cycle of 10 min at 11 000 Â g. Finally, in order to wash the powder and to reach a neutral pH, the precipitate was dialyzed against water (at 3.5 kDa MWCO) for several days before being freeze-dried prior to characterization.

PEGylation of TiONRs
TiONRs' surfaces were functionalized with a mixture of mPEG 5000 -Si and NHS-PEG 5000 -Si (see ESI and Fig. SI_1 † for the silanization of PEG derivatives). 10.6 mg of naked TiONRs were dispersed under manual agitation in dichloromethane. Then 32 mg (mass ratio PEG : TiONRs ¼ 3 : 1) of a molar ratio of 95% of mPEG 5000 -Si (30 mg) and 5% of NHS-PEG 5000 -Si (2 mg) were added to the TiONRs suspension. The mixture was magnetically stirred (150 rpm) for 48 h at 20 C under inert atmosphere. The excess of PEGs was then washed with centrifugation cycles in dichloromethane and PEGylated-TiONRs were nally freezedried before further use and characterization. Hereaer, these nanohybrids are referred to as TiONRs-PEG-NHS.

Functionalization of TiONRs-PEG
Functionalization of TiONRs-PEG-NHS with collagen was performed via NHS ester-amine reaction. Briey, TiONRs-PEG-NHS was suspended in PBS at 50 mg mL À1 with type I collagen (Horm from Nycodem) at 10 mg mL À1 at 20-22 C for 30 minutes. The suspension was then washed twice by centrifugation in PBS (4000 Â g, 2 min) and the nanohybrids (TiONRs-PEG-Coll-I) were resuspended in PBS.

Characterizations of TiONRs nanohybrids
Transmission Electron Microscopy (TEM) characterization was performed using a JEOL JEM-2100F microscope operating at 200 kV (point to point resolution of 0.19 nm). One hundred nanoribbons were counted in order to calculate nanoribbons average dimensions.
Powders were analysed using a Discovery TGA-TA Instruments with an air ow rate of 25 mL min À1 . A temperature ramp of 5 C min À1 from 25 C to 800 C was applied.
Specic surface area (SSA) measurements were performed using a Micromeritics Tristar II apparatus. Samples were outgassed in situ at 100 C under a pressure of 26 mbar for 15 h and the measurements were performed at liquid N 2 temperature with N 2 adsorbing gas.
Considering mass losses at different temperatures coupled to the SSA measurements, graing rates of the two polymers were calculated.
Polymer silanization was followed by proton nuclear magnetic resonance ( 1 H-NMR). 1 H-NMR spectra of synthesized polymers were recorded on a Bruker AVANCE 300 spectrometer in deuterated chloroform (CDCl 3 ) at 300 MHz and 293 K (see ESI †).

Cytotoxicity evaluation
With a focus on cardiac damages, MTT cytotoxicity tests were performed on cardiomyocytes and broblasts (rat's primary culture) in contact with naked TiONRs or TiONTs. Dose effect of TiONRs was evaluated for 72 hours. On day 1, the cells in 24-well plates were incubated (37 C, 5% CO 2 ) with TiONRs suspensions at 2, 20 and 66 mg mL À1 for 24 h. This operation was repeated twice on day 2 and 3 to study the dose effect (48 h and 72 h). For a morphology comparison, titanate nanotubes (TiONTs: length: 150 nm and diameter: 10 nm) synthesized following protocols from previous studies 42, 43 were also incubated with broblasts in the same conditions and at the nal concentration of 66 mg mL À1 . Aer incubation, cells were rinsed twice with PBS at 37 C and incubated for 1 h with 500 mL of MTT at 2 mg mL À1 in PuCK G+ cell medium. Finally, MTT solution was replaced by 500 mL of isopropanol solution with 0.1 mol L -1 HCl for 45 min at 37 C before optical analysis at 570 nm. The experiments were run in independent triplicate to perform statistical analyses.

Platelet aggregation
Platelet aggregation tests were performed on platelet-rich plasma (PRP) incubated with naked TiONRs. Blood samples were collected from volunteer donors into citrate 3.2% collector tubes (BD Vacutainer France). Tubes were centrifuged at 150 Â g for 10 min to obtain PRP. The residual blood was further centrifuged at 2500 Â g for 15 min to obtain plasma poor plasma (PPP). The chosen PRP chosen was coming from voluntary donors and was selected to have at least 350 Â 10 6 platelets per mL of plasma. 290 mL of PRP (>300 g L À1 ) were mixed with 10 mL of naked TiONRs to obtain nal concentrations of 1, 10, 25, 50, 100 and 250 mg mL À1 . The suspensions of TiONRs in PRP were preincubated for 20 min at 37 C before the measurements. As a positive control, 10 mL adenosine diphosphate (ADP) at 5 mM was used to stimulate the aggregation of 290 mL of PRP. Aggregation was measured via thromboaggregometer (Ta8v from SD Medical). Another experiment was also performed to study the inuence of naked TiONRs on platelet aggregation. 290 mL of PRP were mixed with 10 mL of naked TiONRs at nal concentrations of 100 and 250 mg mL À1 . 11 minutes aer reactions in the thrombo-aggregometer, 10 mL of ADP were added to induce the aggregation. Intensities of aggregation were then measured with light transmissions set up at 0% and 100% for PRP and PPP respectively.

Cell adhesion
Effects of TiONRs' functionalization with collagen on cell adhesion were measured. Type I collagen (Coll-I), naked TiONRs, TiONRs-PEG-NHS, naked TiONRs + 10 mg mL À1 of Coll-I and TiONRs-PEG-Coll-I were placed in wells of 96 wells plates (Maxisorp from Nunc) at 50 mg mL À1 of nanohybrids and 10 mg mL À1 of Coll-I for 2 h at 37 C. Then, wells were washed twice with PBS and passivated with bovine serum albumin (BSA) at 30 g L À1 in PBS to avoid unspecic binding. The excess of BSA was also washed twice with PBS. MRC-5 cells (human pulmonary broblasts) were incubated in the wells with 40 000 cells per well/100 mL at 37 C for 2 h. Then the wells were washed twice with PBS and the adhesive cells were xed with 100 mL per well of cold methanol (À20 C) for 10 min. The cells were coloured with 100 mL per well of crystal violet at 5 mg mL À1 in methanol for 10 min. The reactants were nally washed with water and the cells were dried overnight before dissolving the crystal violet in 100 mL per well of SDS solution (10 mg mL À1 ) and measuring the absorbance at 570 nm. The values of adhesion were compared to the cell adhesion on 10 mg mL -1 of Coll-I.

Characteristics of TiONRs
The TiONRs obtained from hydrothermal syntheses have a specic morphology (Fig. 1) with a length between 1 to 20 microns, and a width varying from 70 to 200 nm. These average dimensions were measured on 8 reproducible batches of TiONRs (Fig. SI_2 †). The average thickness of TiONRs is comprised between 3 and 40 nm; dimension's measurements was optimized in a previous publication. 41 The morphology of the TiONRs is particularly suitable to highly and easily cover surface for potential tissue regeneration demonstrated by a good biomimetics with the extracellular matrix. 44 By optimizing parameters, synthesis reaches 99% purity, that means less than 1% (in number) of by-products such as nanosheets, nanotubes or remaining TiO 2 precursor are mixed with nanoribbons. This synthesis is reproducible as there are only few variations in terms of structure, morphology, and chemical composition (type Na y H 2Ày Ti n O 2n+1 , xH 2 O) aer close to a dozen of syntheses (see ESI and Fig. SI_2 †). Functionalization of TiONRs with Si-PEG-NHS improved the colloidal stability of TiONRs ( Fig. 2-a). PEG polymers were chosen because of their high biocompatibility with many other nanomaterials. 42,45,46 The success of the silanization by Si-PEG-NHS was conrmed with NMR analyses (Fig. SI_3 †) proving a 90% yield of silanization and a nal 80% yield of active NHS aer functionalization. With a specic surface of 25 m 2 g À1 for naked TiONRs coupled to the two mass losses (from 100 to 450 C) of TiONRs-PEG-NHS (17.7% see Fig. SI_4 †), corresponding to the degradation of mPEG 5000 -Si and Si-PEG 5000 -NHS, the concentrations of these two PEG on TiONRS are 1 and 0.1 molecules per nm 2 , respectively.
Thus, TiONRs have 0.08 active NHS per nm 2 . All the physicochemical characteristics of the TiONRs are summarized in Table 1.

TiONRs as a potential nanohybrid for regenerative medicine
No signicant cytotoxicities of TiONRs were found on cardiomyocytes. The maximum toxicity on TiONRs on broblasts was below 20% (Fig. 2-b) for the highest concentrations (66 mg mL À1 ) and no signicant toxicity was found for the two other concentrations tested (2 and 20 mg mL À1 ). This value is more than twice lower than the cytotoxicity observed aer incubation of 1 to 3 doses of TiONTs at a similar concentration on both cell lines. This low toxicity of TiONRs could be attributed to their original morphology. In fact, elongated nanomaterials such as nanotubes or nanorods have usually more chance to be internalized by cells and then to inuence cell integrity and cause more damage. Regarding the TiONRs, because their lengths are higher than 1 mm and they are quite larger than classical nanotubes/rods (a few hundreds of nm compared to a few dozens of nm) they have less chance to be internalized and to promote cytotoxicity compared to the nanotube morphology that showed signicantly higher cell killing effect. 47,48 Furthermore, a dose effect was observed with TiONTs when none was observed aer 3 successive doses of TiONRs.
To conrm the hypothesis of lower cellular interactions of TiONRs compared to TiONTs, cytotoxicity measurements were correlated with optical microscopy pictures of broblasts ( Fig. 3-a) incubated with nanoparticles. Two similar doses of TiONRs at 66 mg mL À1 showed much more aggregates of particles on top of the cells than with 2 doses of TiONTs at 66 mg mL À1 conrming that cells might not have internalized many TiONRs compared to TiONTs.
For concentrations up to 250 mg mL À1 , TiONRs did not induce any spontaneous aggregation of the platelets in the PRP compared to the ADP which induced a rapid increase of aggregation within 3 min (Fig. 2-c). At a relatively high concentration (<100 mg mL À1 ), TiONRs do not have any side effect on platelet aggregation which allows their use as matrix for regenerative medicine. However, at 250 mg mL À1 , TiONRs seem to slow down the platelet activity, as evidenced by the 2-fold reduction of the measured aggregation compared to the control and the TiONRs at 100 mg mL À1 . A concentration of 250 mg mL À1 of TiONRs seems to be too high and might affect platelet aggregation.
Cellular adhesion efficiency of broblasts in the presence of TiONRs with or without PEG and functionalized or not with Coll-I was then quantied (Fig. 3-b). Surface covered with Type I collagen was used as a positive control. Coll-I is the most abundant collagen in human body made of bers used to heal  cardiomyocytes. * Significant differences compared to the control p < 0.05; $ significant differences compared to the TiONRs p < 0.05; (c) aggregation's activation of platelets in presence of TiONRs at 1, 10, 25, 50, 100 and 250 mg mL À1 and ADP at 5 mM. ADP was injected 11 minutes after incubation. Platelets aggregation's activation in presence of TiONRs at 100 and 250 mg mL À1 was also measured. ADP: adenosine diphosphate. wounds and already well used to improve cellular adhesion in regenerative medicine. 49,50 Among the non-toxic concentrations, a TiONRs' concentration of 50 mg mL À1 was chosen to cover the whole plate, thus limiting modulation of the platelet aggregation and forming a monolayer (the optimization of which was corroborated by optical microscopy: see Fig. SI_5 †). Surface coated with naked and PEGylated TiONRs do not allow cellular adhesion. 27% of adhesion were observed on naked TiONRs mixed with Coll-I. As TiONRs are completely covering the plate, this adhesion could certainly be explained by non-specic adsorption of the collagen on TiONRs surface allowing some cells to stick on the plate surface. Such unspe-cic adsorption is difficult to control and very common with proteins and nanomaterials. 51,52 For the chemically functionalized TiONRs-PEG-Coll-I, the percentage of adhesion is as good as the adhesion with only Coll-I alone, proving the efficacy of the graing of collagen on PEGylated TiONRs. As PEG is a wellknown polymer that prevents proteins adsorption 53,54 and that is also used for its antifouling properties preventing cellular adhesion, 55,56 the obtained rate of adhesion proved that PEG is not accessible on the surface of the plate. It also proved successful covalent functionalization of Coll-I on TiONRs-PEG-NHS that allowed successful adhesion of broblasts on wellplate. Besides, it can be noted that the Coll-I coating prevents the direct interaction of the TiONRs' scaffold with broblasts.

Conclusions
In this study, a new nanobiohybrid favoring cellular adhesion and proliferation has been developed for in ne tissue healing and regeneration, especially on cardiac damages. In a pilot study, we were able to synthesize in a reproducible manner titanate nanoribbons (TiONRs) with biocompatible functional polymers and type I collagen. TiONRs showed low cytotoxicity against cardiomyocytes and cardiac broblasts with a minimal dose effect compared to nanotube morphology. At concentrations below 100 mg mL À1 , TiONRs neither induced nor inhibit the platelet adhesion opening the way for their putative use as bandages to cover skin wounds. We demonstrated that well controlled functionalization of TiONRs could lead to future materials for regenerative medicine with potential theranostics' applications thanks to the interesting properties of titanate material.

Conflicts of interest
The authors declare no conict of interest.