3D Printing of large-scale and highly porous biodegradable tissue engineering scaffolds from poly(trimethylene-carbonate) using two-photon-polymerization

The introduction of two-photon polymerization (2PP) to the field of tissue engineering and regenerative medicine (TERM) has led to great expectations for the production of scaffolds with an unprecedented degree of complexity and tailorable architecture. Unfortunately, resolution and size are usually mutually exclusive when using 2PP, resulting in a lack of highly-detailed scaffolds with a relevant size for clinical application. Through the combination of using a highly reactive photopolymer and optimizing key printing parameters, we propose for the first time a biodegradable and biocompatible poly(trimethylene-carbonate) (PTMC)-based scaffold of large size (18 × 18 × 0.9 mm) with a volume of 292 mm3 produced using 2PP. This increase in size results in a significant volumetric increase by almost an order of magnitude compared to previously available large-scale structures (Stichel 2010 J. Laser Micro./Nanoeng. 5 209–12). The structure’s detailed design resulted in a highly porous scaffold (96%) with excellent cytocompatibility, supporting the attachment, proliferation and differentiation of human adipose-derived mesenchymal stem cells towards their osteogenic and chondrogenic lineages. This work strongly attests that 2PP is becoming a highly suitable technique for producing large-sized scaffolds with a complex architecture. We show as a proof-of-concept that an arrayed design of repetitive units can be produced, but a further perspective will be to print scaffolds with anisotropic features that are more representative of human tissues.


Introduction
The use of cell-laden three-dimensional (3D) scaffolds is a common strategy in the field of tissue engineering and regenerative medicine (TERM). When recreating the architectural features of tissues, several key factors, such as biocompatibility, mechanical support and porosity play an important role for the design of a scaffold. While traditional methods, such as salt leaching, gas foaming or freeze-drying follow these rules for scaffold fabrication, they lack the spatial control over the internal architecture of the scaffold. More specifically, they do not provide any control over the resulting pores size and overall porosity of the resulting structure. Both, however, are of vast importance to a scaffold's design as they directly affect nutritional diffusion, but also cellular proliferation and invasion of the scaffold [2].
Therefore, the field of TERM often relies on the use of additive manufacturing methods, often referred to as 3D printing. Most commonly, optical fabrication techniques based on 1-photon absorption are chosen, with stereolithography (SLA) being the most popular. SLA produces 3D objects in a layer-by-layer procedure by selectively curing a photosensitive resin. Freedom of design and a high degree of porosity are two highly-valued benefits which have propelled SLA in the field of TERM, this method is however limited to a feature resolution of around 20 µm [3].
This, in turn, imposes a lower thickness limit on the struts separating the pores, which in turn, restricts the overall volumetric porosity of a scaffold (i.e. pores can only be spaced so close to each other). Other methods, such as electrospinning, allow the fabrication of scaffolds with higher porosity of up to 97%. However, the tight packing of fibers on the micrometer scale has been reported to impede the process of cellularization [4,5]. While melt electrowriting allows the controlled stacking of thin fibers, this method cannot accurately reproduce intricate 3D structures based on CAD models, the same way photopolymerization-based techniques, such as SLA and two-photon polymerization (2PP) can.
2PP is a high-resolution 3D printing method, allowing to obtain volumetric elements of submicrometer resolution. Compared to SLA, 2PP permits to fabricate structures with a significantly higher spatial resolution. 2PP requires the combined energy of two photons to initiate photopolymerization, which in turn confines the phenomenon to the focal plane of the laser beam. This way, a high-resolution freeform 3D structures can be patterned into the resin by only moving the focal point of the laser [6].
Compared to other techniques, only a small number of biodegradable scaffolds have been reported with 2PP, even though it is a highly applicable technique. A major drawback to 2PP is that the increase of resolution is linked to comparatively low throughput. At laser writing speeds as low as 0.1 mm s −1 , which are not uncommon, the production of a cube of 1 cm 3 was estimated to take 300 years by Stichel et al [1] Even though faster speeds have been reported [7], scaffolds produced with 2PP are usually rather small to keep within a reasonable production time limit. So far, the largest reported scaffold had a side length of 5 mm, which underlines why 2PP faces a substantial acceptance barrier for TERM applications [1].
A further limiting factor for the scaffold production with 2PP is the availability of materials. Compared to other fabrication techniques, the number of reported biodegradable scaffolds produced with 2PP is very low. A possible reason for this is the lack of biodegradable and commercially available photopolymers suitable for 2PP. A point in favor of the development of more biodegradable resins is that the nonlinear absorption makes the resin preparation far less complex by not requiring additional light absorbers (e.g. orasol orange or titanium dioxide particles) [8]. The absence thereof further reduces the risk of toxicity of the final scaffold.
Polymers typically processed with SLA provide an excellent starting point for the use with 2PP. Typically, (meth-)acrylated poly(ester) polymers are used for SLA. As such, polylactic acid (PLA) has found use in both, SLA [9] and 2PP [10], for a range of applications, such as bone tissue formation. Another polymer, previously used for SLA, is poly(trimethylenecarbonate) (PTMC). Due to its peculiar degradation mechanism (surface erosion driven by enzymatic erosion), PTMC has shown to be a highly suitable polymer for the fabrication of scaffolds for bone tissue engineering and hence poses a great potential for the use in 2PP [11].
With all the recent developments in 2PP for TERM applications, some technical limitations remain. Currently, large scaffolds require a long time to produce and stitching adjacent field of views will leave a noticeable transition. Moreover, the limited availability of biocompatible and biodegradable photopolymers and suitable photoinitiators, as well as the high price point of 2PP systems is inhibiting the fast adoption of the technique for TERM applications.
In this work, we mitigate some of the mentioned bottlenecks of 2PP. First, we detail the composition of a highly photoreactive resin based on PTMC. Through a systematic screening study of process parameters for 2PP, we further show the production of novel biocompatible and biodegradable scaffolds with an unprecedented large size. The structures have a high volumetric porosity of 96% and complex microarchitectures. Moreover, these scaffolds proved excellent cellular invasion and differentiation. This shows the suitability of 2PP as an additive manufacturing technique to produce high-resolution scaffolds with highly relevant properties for TERM.

Photopolymer synthesis
3-armed PTMC, was synthesized by ring-opening polymerization (ROP) of trimethylene carbonate (TMC) in a three-necked round-bottomed flask with trimethylolpropane (initiator) under argon atmosphere at 130 • C for 3 d (50 g (0.49 mol) TMC and 1.48 g (5.0 mmol) trimethylolpropane with 0.065 g of Sn(Oct) 2 as catalyst). The 3-armed PTMC (PTMC oligomer) was dissolved in dichloromethane after cooling to room temperature. Methacrylatefunctionalized PTMC (PTMC-MA macromer) was synthesized by the following method. Under argon, 0.05 g hydroquinone, 10 ml triethylamine (0.072 mol) and 12 ml methacrylic anhydride (0.081 mol) were added. The mixed solution was reacted with stirring for 5 d in the dark at room temperature. The solution was precipitated and washed in cold methanol to remove unreacted monomer and other compounds. The PTMC-MA containing methanol was firstly dried in the dark at ambient conditions overnight and then dried in vacuum for another 7 d at room temperature (the synthesis route is shown in SD 1). The PTMC oligomer and the PTMC-MA macromer were analyzed by 1 H-NMR (Bruker Ascend 400/Avance III 400 MHz NMR spectrometer) to determine the average molecular mass (Mn) and the degree of functionalization of PTMC-MA. We obtained PTMC of 4300 g mol −1 with a degree of methacrylation for the PTMC-MA of 97%.

Resin preparation
PTMC resin was prepared by dissolving 3-armed PTMC-MA in tetrahydrofuran (THF) with 0.5 wt% M2CMK as a photoinitiator, providing a two-photon absorption cross section of around 100 Göppert-Mayer (GM) units [13]. After 24 h, the resin was heated to 150 • C to evaporate the THF. To eliminate air bubbles, the resin was placed in a vacuum for 10 min before dispensing into appropriate sample holders. The sample holders consisted of a 3 mm thick aluminum sheet the size of a microscope slide with holes, sealed on one side with cover slips using 2-component glue. Before printing with 2PP, the PTMC resin was heated to 50 • C.

Laser setup, 2PP printing and post-processing steps
A tunable femtosecond-pulsed laser (MaiTai eHP DeepSee, Spectra-Physics) was operated at 800 nm at a repetition rate of 80 MHz and a pulse duration of 70 fs after the microscope objective (UPlanS-Apo, 10x/0.4 NA, Olympus, Japan). Laser positioning was done by a dual-axis galvanometric scanner (Scanlabs, Germany) and sample positioning was done by a microscope stage (ScanPLUS IM, 120 × 180, Märzhäuser-Wetzlar, Germany). A writing speed of 1 000 mm s −1 was used and different laser intensities were screened from 20 to 380 mW. The schematics of the process is shown in figure  1. The laser was focused through a cover slip into the material. Due to the resin's inherent viscosity, the structures were printed top-down. Hence, the already polymerized part never obstructed the laser beam.
The buckyball structures were designed in Solid-Works (Dassault Systèmes, France). After structuring with 2PP, the sample holder was submerged in THF for 1 h to dissolve any uncrosslinked PTMC-MA. Two additional THF washing steps were carried out before storing the scaffold in 1-propanol.

Scaffold characterization
Optical analysis of the produced scaffolds was performed with a confocal microscope (LSM700, Zeiss, Germany), a scanning electron microscope (SEM, Philips XL Series 30) and and high resolution micro computed tomography (µCT). µCT was performed on a Scanco 50 scanner (Scanco Medical AG, Switzerland) at 70 kV, 57 µA with an 0.5 mm Al filter at 1000 ms integration time and 1500 Projections. Subsequently, the scaffold was skeletonized and strut elements were segmented by removing the intersecting nodes. Local strut thickness was measured and graphically overlayed as a representative sphere of corresponding diameter at the center of each strut for better visualization.

Cell viability
Viability of the cells was determined using a Live/Dead ® assay (Invitrogen, OR) using 0.2 µM calcein-AM (live stain) and 0.6 µM propidium iodide (dead stain) in serum-free medium for 60 min at 37 • C. The viability of the cells was monitored on day 14 and day 28 using a confocal laser scanning microscope LSM 700 (Zeiss, Germany).

Scanning electron microscopy (SEM)
SEM analysis of the constructs was performed after fixation in buffered paraformaldehyde at 4%, followed by gradual dehydration in ethanol and by immersion in hexamethyldisilazane (Sigma-Aldrich). After drying, the samples were sputter-coated with Au and investigated by SEM.

Histology
After 28 d of cultivation, cell-seeded scaffolds were washed in PBS and fixed overnight in  Roti ® Histofix 4% (Carl Roth, Germany) at 4 • C. Samples were embedded in paraffin blocks and processed at the Histopathology Department (Vienna BioCenter Core Facilities GmbH, Austria) for Hematoxylin-Eosin (H&E), Von Kossa, Alcian Blue staining, and for proliferative marker Ki67 immunostaining.  sample was quantified using the Quant-iT PicoGreen assay (Thermofisher). The content of sulfated glycosaminoglycans (sGAG) was quantified using the dimethylmethylene blue dye-binding assay (DMMB, Blyscan, Biocolor Ltd., United Kingdom), with a chondroitin sulphate standard.

Optimization of PTMC-MA printing using 2PP
In order to define an optimal 2PP processing window for the PTMC-MA, a laser intensity screening test was conducted. For this test, a fixed laser writing speed of 1000 mm s −1 was selected and an assay with a stepwise increase of the laser power from 20 to 380 mW was performed. The results presented in figure 2 revealed the 2PP processing window between 40 to 220 mW. Below and above this threshold, either no polymerization occurred or material degradation was visible, respectively. The first is seen by the absence of a printed structure at a laser intensity of 20 mW, whereas the latter was seen through the occurrence of bubbles due to thermal effects. The test cubes were printed at a fixed hatch value of 0.5 µm and a fixed dZ value of 1 µm, which describe the distance of two adjacent laser lines in xy-and z-direction, respectively. Consequently, the 2PP processing window was further optimized in a subsequent step to find optimal printing values for hatch and dZ. These values directly influence the quality of the printed parts and are important when printing more intricate structures, such as the architecture of the macroscaffolds. Therefore, we produced 3D microstructures consisting of hexagons and pentagons arranged to form a low-polygon sphere based on the fullerene structure. The microstructures, also referred to as buckyballs (BB) [12] were produced at varying hatch and dZ values ranging from 0.4 µm to 1.5 µm and at 100, 130 and 160 mW. The so obtained parametrically arrayed structures were then microscopically assessed whether they match their 3D model in terms of strut thickness and open pores.
We concluded that a laser intensity of 130 mW, a hatch value of 0.8 µm and a dZ value of 1 µm would yield structures that best resemble their 3D model. Using these parameters, printing of a single buckyball took around 20 s, which was considered reasonable for our application.
The 2PP fabrication of a single BB is shown in SD 2 (printing sped up 10 times, with a bottomup view). After dissolution of unpolymerized materials in THF and drying in HMDS, the BBs were characterized using SEM ( figure 3(a)). The measured diameter of the buckyball was around 300 µm with a strut width of around 20 µm and strut thickness of around 30 µm. The measured values correspond to the dimensions of the 3D model indicating that the printing parameters were chosen correctly.
Once the processing parameters of the PTMCbased resin were established, a porous scaffold was fabricated. The scaffold consisted of an array of 8 × 8 × 3 interconnected buckyballs forming a rectangular and highly porous structure ( figure 3(b)). It had a side length of 2.4 mm, a height of 0.9 mm and took around 60 min to produce (figures 4(a)-(c)). The structure was designed to feature a high volumetric porosity to support rapid cell invasion and nutrient exchange. SEM imaging confirmed that the delicate structures were preserved during post-processing of the sample.
In a last step, we further upscaled the scaffold to a side length of 18 × 18 × 0.9 mm (l, w, h), which, to the best of our knowledge, is the largest biodegradable   scaffold produced using 2PP to date ( figure 5(a)). It is composed of 3072 BB units, took 16 h to print and shows the potential of 2PP to produce highly-intricate scaffolds for TERM.
The scaffold was imaged using high-resolution micro-computed tomography, which showed its structural integrity (SD 3) and revealed a strut thickness ranging from 22 to 30 µm with a median thickness of 25 µm ( figure 5(b)).

2PP-produced PTMC-MA scaffolds support stem cells proliferation and differentiation
Scaffolds printed from PTMC-MA using 2PP supported the adhesion and proliferation of hASCs. Two days post-seeding, hASCs attached to the scaffold and exhibited elongated filopodia ( figure 6(a)). Over 28 d of culture, the hASCs colonized the entire scaffold and formed a dense cellular layer on the surface of the scaffold (figures 6(b) and (c)). Live/dead staining conducted at day 28 showed a high cellular viability (figure 6(d)). The proliferation of the hASCs was validated by the significant increase of DNA (figure 6(e)) and presence of ki67-positive cells (figure 6(f)). We then evaluated whether PTMC-MA scaffolds would support hASCs to differentiate towards chondrogenic and osteogenic phenotypes.
When cultivated in appropriate media, the seeded macroscaffolds supported hASCs chondrogenic and osteogenic differentiation. SEM cross section demonstrated that such 3D scaffolds, due to their high degree of porosity, were easily colonized by the cells (figure 7(a)). sGAG secreted by the hASCs seeded on the scaffolds was shown to be significantly higher for CM condition compared to control condition ( figure 7(b)), which was also shown by the positive Alcian Blue staining ( figure 7(c)). SEM and histology sections demonstrated that a large amount of ECM was deposited throughout the thickness of the structures. Finally, once cultivated in osteogenic condition, hASCs were able to differentiate toward osteoblasts and to deposit calcium precipitates as shown by the positive Von Kossa staining (figure 7(d)).

2PP is a relevant technique to print TERM scaffolds
SLA and digital light processing (DLP) are the most used lithography-based additive manufacturing processes for the production of scaffolds for TERM. They enable the rapid production of scaffolds of clinicallyrelevant sizes and porosities while also enabling rapid design iterations.
However, their minimal resolution is dictated by the diameter of the laser beam [26]. In contrast, printing processes based on two-photon absorption permit to achieve a significantly higher resolution, albeit at the expense of the printing time, which typically affects a scaffold's final size. This relation is further reflected in the available literature on scaffolds produced with 2PP for TERM.
While most of the available reports demonstrate a high resolution, the largest construct to date, to the best of our knowledge, featured a side length of only 5 mm (figure 8). Despite being highly porous (∼80%), this construct took nevertheless 4 h and 20 min to print and was produced from Ormocer ® , a commercial, non-biodegradable hybrid polymer [1].
In our work, we demonstrated that we were able to significantly increase the volume by almost an order of magnitude by producing scaffolds with a side length of 18 mm and a volume of 292 mm 3 .
On the hardware side, this was made possible by using a low magnification objective (x10, numerical aperture of 0.4) and a fast writing speed (1000 mm s −1 ), whereas optimizing the material-specific parameters and the printing settings was critical on the software side. Together, this Figure 8. Our work introduces a 2PP printed scaffold, novel not only regarding its composition, but also its unprecedented large size. Graphical representation of the volumes of scaffolds printed with 2PP presented in the literature, with the nature of the photopolymer, the objective used and the writing speed. ND indicates that either the microscope objective or the writing speed is not described in the experimental section of the reference. References using biodegradable biomaterials are highlighted in bold characters. List of the references is available here [1,7,[14][15][16][17][18][19][20][21][22][23][24][25]. enabled the fast writing speeds of 1 h for the smaller scaffold presented in figure 4 and 16 h for the larger scaffold presented in figure 5.
Achieving fast printing speeds is one the key factors to realizing constructs with a relevant size for TERM. In previous publications, printing speeds ranged from 5 mm s −1 [1,18] up to 30 mm s −1 [7]. However, writing as slow as 0.1 mm s −1 is not uncommon and at these speeds, the production of a 1 cm 3 block was estimated to take 300 years by Stichel et al [1] Working at fast writing speeds requires photoinitiators with a high two-photon absorption cross section (δ). Photoinitiators typically used in 1-photon absorption, such as different types of Irgacure have δ values in the order of around 20-30 Göppert Mayer (GM) units [27]. Photoinitiators that were specifically developed for two-photon absorption, such as M2CMK used in this work, have a δ value of around 150 GM, requiring a lower laser intensity for initiation [13]. This, in turn, allows faster writing speeds but also a lower initiator concentration, which is favorable for the design of biocompatible scaffolds. Lastly, the absorption maximum of the chosen photoinitiator is of importance, as light in the near infrared has a higher penetration depth than UV-light [14].
Apart from material properties and print settings, technical advances of the 2PP setup are also of importance. For instance, Trautmann et al developed a single-line single-pass process to increase the speed of scaffold fabrication [18]. This strategy takes advantage of true 3D writing and does not follow a layer-by-layer approach, enabling the fabrication of scaffolds with a surface larger than 9 mm 2 and a 140 times faster throughput. The major downside of this technique is the approximation of cubic structures by inlaying ellipses, as it lowers the accuracy and limits the type of structures that can be printed.
A different approach to upscaling 2PP production is the employment of multiple foci. By splitting the beam before the scanner, the foci can be moved in parallel. Using a spatial light modulator, a 16-foci system was demonstrated to produce a scaffold with a surface of 1.6 mm 2 tested with bovine endothelial cells [20]. Using the same approach, a 2-and 6-foci setup was shown for the fabrication of nichoids for stem cell expansion [28]. Parallel foci further require a higher laser output power, which may result in laser drifting over time and possibly thermal damage of the crystal screen. Another disadvantage of multiple foci is the requirement to lower the scanning speed which impacted the time gain significantly [20,29]. Zandrini et al showed the polymerization of SZ2080 resin with a laser energy of 260 mW and a translation speed of 3 mm s −1 . In comparison, our scaffold was written at 1 000 mm s −1 using half the laser intensity, which consequently increased our volumetric processing time by a factor of 166 and 55 compared to the 2-and 6-foci setup, respectively. The combination of fast writing speeds, a highly reactive photoinitiator and optimized process parameters have resulted in a scaffold for TERM applications with a previously unreported volume.
With regards to the microarchitecture of the scaffold, we aimed at producing a structure with a small pore size and high volumetric porosity at reasonable production times. Porosity is essential to create a microenvironment favorable for the cell proliferation, ECM deposition and suitable neo-tissue formation. While a high porosity (>90%) is generally regarded beneficial for osteogenic scaffolds [19,30] there has been a wide range of reported pore sizes (50-500 µm) suitable for osseous tissue further depending on what type of bone is mimicked [30][31][32]. Numerous publications have shown that osteogenesis is more favorable in scaffolds with pores around 300 µm and larger [30,[33][34][35][36][37]. Naturally, other factors, such as pore geometry, scaffold material, cell type, cell culture conditions and surface topology also play an important role in deciding a cell's fate.
In finding a compromise between these key parameters, we have created a scaffold with a side length of 18 mm with pore sizes of 300 µm and a volumetric porosity of 96% in around 16 h. Choosing such a highly-porous structure with narrow pore sizes, only leaves space for small scaffold struts. This stringent criterium highlights the need for 2PP as an appropriate fabrication technique for TERM, as many conventional photofabrication techniques based on 1-photon absorption would fall short of this task. Moreover, the increase in resolution provided by 2PP could further allow the addition of micro-patterns on a scaffold's surface to further accelerate the cellular fate towards osteogenesis [38].

Novel biocompatible PTMC-based resin for 2PP
The material shown in this work is another point of novelty and of high importance. As shown in figure  8, scaffolds produced by 2PP are usually composed of commercial resins, such as Ormocer ® , OrmoComp ® or SZ2080. While these were shown to be cytocompatible, they are not biodegradable [15,39,40]. In general, biocompatible and biodegradable materials are preferred for implantation, as they do not create potentially toxic by-products during degradation. Although there have been efforts to make commercially available and biodegradable materials, as seen with bioORMOCER ® , to the best of our knowledge, this material has not been demonstrated with 2PP.
More recently, reports on scaffolds produced from photopolymerizable polyesters, such as PLA and PCL have emerged. Interestingly, these materials were also shown to be polymerizable using 2PP after being functionalized by methacrylate moieties [10,17]. As an alternative to these polyesters, we introduced the photopolymerizable PTMC-MA. Grijpma et al have previously shown that this polymer can be processed with SLA to form scaffolds for numerous applications, such bone [41], vascular [42] or even meniscus repair [43]. To the best of our knowledge, this polycarbonate-based photopolymer has not been processed with 2PP yet. Due to its short size of only 4.3 kDa and the presence of the three MA-branches, the methacrylate double bond conversion and hence the crosslinking polymerization is statistically more efficient compared to longer and linear macromers. This is beneficial for the production of large scaffolds, as the resin can be processed with a fast writing speed and with a high laser power without burning the biomaterial (130 mW at 1000 mm s −1 ). In comparison, PLA-MA was processed with a writing speed of 1 to 3 mm s −1 at 15 mW [20] and with 50 µm s −1 at 30 mW [44].
The cytocompatibility study in this work showed great cellular acceptance of the PTMC-based scaffolds produced with 2PP. These results are in accordance with the observations by Guillaume et al on PTMC scaffolds produced by SLA, suggesting that 2PP has no adverse effect on the scaffold production [41,45]. Nevertheless, the main difference between the two structures is the higher porosity (i.e. 96%) that was reached with 2PP when compared to structures with a porosity of around 70% printed using SLA. 2PP allows the production of scaffolds with a very low material density while still providing substantial surface contact to encourage physical cell adhesion and proliferation. This permits a fast cell invasion and penetration and, in turn, results in a high ECM deposition. After seeding the scaffold with hASCs and cultivating them in chondrogenic conditions, homogenous sGAG distribution could be seen throughout the scaffold. Moreover, hASCs differentiated towards osteoblasts as demonstrated by the presence of calcium phosphate (CaP) mineralization. Contrary to the sGAG deposition, CaP precipitation was not observed uniformly throughout the cross section of the scaffold but was localized at the interface between scaffold struts and ECM. This is not an isolated finding as tissue mineralization is commonly seen to initiate at the interface with the implant [46,47]. Further tailoring the surface properties of the scaffold to provide CaP nucleation sites would help in increasing the scaffold's osteogenic potential. For instance, Werner et al have demonstrated that surface curvature impacts the fate of stem cell differentiation, with convex surfaces promoting osteogenesis [48]. Further, the inclusion of micropatterns on the scaffold surface has previously shown to impact cell-adhesion and differentiation of preosteoblast cells [38]. Such 3D geometric cues could easily be integrated in the 2PP manufacturing workflow of implants dedicated to bone regeneration.
Previous literature has shown cell acceptance of scaffolds produced with 2PP, but to the best of our knowledge, cell penetration, scaffold invasion and ECM production at this scale have never been reported.
Lastly, as mentioned in the introduction, this research represents a first proof-of-concept that 2PP can produce highly intricate structures of large-scale sizes with excellent cytocompatibility when optimal printing parameters are chosen. In the future, we plan to replace the repetitive subunits in this scaffold with similarly intricate structures that result in a more anisotropic architecture and hence better resemble the conditions found in most tissues. The opportunity offered now by 2PP has been the holy grail for many researchers questing to recreate artificial scaffolds biomimicking for instance the anisotropic architecture of osteochondral tissues, amongst others [49,50].

Conclusion
Due to the non-linear absorption of 2PP, a range of structures with highly complex architectures have been presented over the last decade. This high degree of detail is usually connected to a long printing time and therefore scaffolds are only produced at a small scale. To become a relevant technique for the production of TERM scaffolds, the constructs must be significantly scaled up while being printed in a shorter time. Here, we demonstrated that by: (i) formulating a specialized photosensitive resin, based on a highlyreactive three-armed methacrylated PTMC with a photoinitiator specifically developed for 2-photon absorption, and by (ii) optimizing the 2PP printing parameters, large structures were produced at an unprecedented speed of 1000 mm s −1 . The resulting constructs have a volume of an order of magnitude higher than previously reported constructs and have proven to be suitable for homogenous cell invasion and matrix deposition. This research showed advancements in two areas for 2PP, namely the processing of a new biomaterial and the production of highly porous and biodegradable scaffolds at a relevant scale for TERM.