Achieving biocompatibility and tailoring mechanical properties of SLA 3D printed devices for microfluidic and cell culture applications

Matt D. Nelson*^a, Patrick A. Tresco^a, Christian C. Yost^b and Bruce K. Gale^c
aBiomedical Engineering, University of Utah, 36 S. Wasatch Drive, SMBB 3100, Salt Lake City, Utah, USA. E-mail: matt.d.nelson@utah.edu
^bPediatrics, University of Utah Health, Salt Lake City, Utah, USA
^cMechanical Engineering, University of Utah, 50 S. Central Campus Dr., MEB, Rm 2110, Salt Lake City, Utah, USA

First published on 27th August 2024

---

---

Introduction

Microfluidic device fabrication in academic settings is dominated by soft lithography (specifically with polydimethylsiloxane, PDMS) largely because PDMS exhibits a host of beneficial properties for microfluidics and biomedical applications, including biocompatibility, mechanical compliance, and optical transparency. The overall benefits and drawbacks of soft lithography have been covered extensively in the literature over the past few decades.^1–3 One characteristic of particular relevance is that soft lithography is often limited to the production of 2D architectures, with the fabrication of more complex 3D microfluidic structures only possible by carefully aligning and bonding separate layers.

The use of additive manufacturing (3D printing) for the fabrication of microfluidic devices has burgeoned over the last decade^4–7 and has the potential to replace soft lithography in some microfluidic applications. A multitude of 3D printing techniques, namely stereolithography (SLA), fused deposition modeling (FDM), and multi jet modeling (MJM), among others, have enabled researchers to create three-dimensional microfluidic architectures while avoiding the need for traditional nanofabrication cleanrooms, which are expensive to operate and require advanced training. Over the years, these additive manufacturing techniques have seen steady advancements in resolution and material selection, bringing their capabilities closer to those offered by soft lithography.^8–13 Of these techniques, photopolymerization-based additive manufacturing techniques such as SLA, MJM, digital light processing (DLP), and masked stereolithography (MSLA) have become widely used, largely due to their combination of high resolution and optical clarity. However, despite their growing popularity, many devices printed with these techniques are cytotoxic without appropriate post-processing—even those printed with resins advertised as “biocompatible”. It is not always clear to researchers what the best biocompatible resin options available to them are outside of creating their own resin mixture, which lies outside the skillset of most microfluidic engineers and cell biologists. Applying UV or heat during post-processing can be used to improve biocompatibility,¹⁴ but some microfluidic components—namely, valves and actuators—require material compliance, which is challenging with resins that rigidify during post-processing with UV or heat. Thus, post-processing steps used to improve biocompatibility can be directly at odds with manufacturing steps needed for practical mechanical properties and device function in microfluidics. A major purpose of this paper is to explore ways to simultaneously enable tailored mechanical properties and reduce cytotoxicity in SLA 3D printed devices for microfluidic and other cell-based applications.

At a fundamental level, SLA resins are typically comprised of a mixture of monomers and/or oligomers with reactive vinyl groups, a photoinitiator, and occasionally other chemical species, such as UV absorbers to improve resolution. In these photopolymerization additive manufacturing techniques, the monomers/oligomers are selectively exposed to UV radiation comprising wavelengths between 365–405 nm, initiating localized free radical polymerization, which induces a liquid-to-solid phase transition by generating a cross-linked polymer structure before the bed is raised up one layer-height in preparation for the fabrication of the next layer (Fig. 1). Although a solid object is created during the 3D printing process, the three-dimensional mesh is not completely crosslinked, and residual monomer/oligomer remains in the structure.¹⁵ Depending on the monomers and photoinitiators used, these residual chemicals may be cytotoxic to cells,^16–19 rendering them incompatible for cell culture applications, including microphysiological systems. Most manufacturers recommend a UV post-cure following the 3D printing process to further polymerize the network. In some cases, this may be adequate to resolve cytotoxicity, but in many cases, it may not be sufficient. Moreover, this post-cure rigidifies the structure, which can disable valves and other actuating components.

Here, using a commercially available resin (Nextdent Ortho Flex, 3D Systems), we demonstrate that several post-processing techniques can be utilized to create non-cytotoxic devices and tailor their mechanical properties. While other researchers have studied post-processing of SLA 3D printed constructs,^20–22 to the best of our knowledge none have explored the relationship of these techniques to both cytotoxicity and the mechanical properties at a sufficient depth to guide the research community. In this work, in an effort to reduce the quantity of cytotoxic components, we employed baking at 120 °C, immersion in phosphate-buffered saline (PBS) with or without network swelling in ethanol, and UV radiation to remove residual methacrylate monomer/oligomer that was quantitatively evaluated through UV-vis spectrophotometry. To assess the impact of the removal of these cytotoxic components using the different methods on cell viability, human umbilical vein endothelial cells (HUVECs) were exposed to extraction medium and evaluated using an MTT assay. Finally, the Young's and flexural modulus were measured to examine whether mechanical properties change as a function of post-processing conditions. Our results will aid researchers using SLA/DLP/MSLA for microfluidics, microphysiological systems, and other cell-safe culture platforms to produce more biocompatible devices with the desired mechanical properties.

Results

Detection of residual monomer/oligomer in 3D-printed structures

Many commercially available resins used in SLA 3D printing have proprietary formulations, which increases the difficulty of identifying and eliminating cytotoxic chemical species. In the case of the resin used in this study (Nextdent Ortho Flex), the manufacturers indicated that the resin contains methacrylate ester monomer, acrylate ester, 2-phenoxyethyl acrylate, and diphenyl(2,4,6-trimethylbenzoyl) phosphine oxide (TPO, a common photoinitiator). Reports in the literature reveal that the vinylic groups of methyl methacrylate absorb UV light with a maximum absorption at 225 nm,^23,24 which suggested that the monomers in this resin mixture would exhibit a similar chemical signature. To verify this analysis, unaltered resin was diluted in ethanol to 1% (v/v), 0.5% (v/v), 0.1% (v/v), 0.05% (v/v), and 0.01% (v/v) and examined using UV-vis spectrophotometry alongside resin polymerized using either UV radiation or heat (Fig. 2). Strong absorbance peaks were observed between 200–300 nm, commensurate with concentration, and peaks representing solid structures presented a similar chemical signature with broader absorbance between 300–400 nm.

With a detection mechanism available, we sought to investigate which techniques were most effective at eliminating residual monomer/oligomer and how that correlates with cell viability. Outside of the manufacturer's recommendation of post-printing UV treatment (in this case, 30 minutes in a 405 nm UV chamber at 60 °C), reports in the literature have suggested that soaking in PBS and baking may also be effective.^20,21 As the resin mixture is readily soluble in ethanol and the polymerized structure expands in ethanol, we were also interested in the effects of a temporary ethanol immersion to swell the polymer chains, followed by soaking in PBS, with the rationale that diffusion of chemical species would be increased during chain expansion. Consequently, all four of these approaches were explored.

Small cylinders that approximate the shape/size of a standard microfluidic device were printed, cleaned, and dried before being grouped and post-processed. Following processing, groups were soaked in PBS overnight and fluid was collected for analysis. Time points of 1, 2, 4, 6, 24, and 48 hours were selected for samples analysis after baking at 120 °C. Soaking in PBS, with or without prior immersion in ethanol for 10 minutes, was carried out for 5 days, with time points and aliquots of extraction medium taken at 1, 3, and 5 days. UV curing was performed according to the manufacturer's instructions for 30, 60, or 90 minutes at 60 °C. Groups not undergoing soaking as a processing condition were soaked overnight in PBS to allow for extraction medium to be collected and analyzed. We found that baking at 120 °C was the most effective method for removing chemical species, with peaks decreasing in absorbance with each successive time interval and showing diminishing returns after 24 hours (Fig. 3). Polymerization with UV radiation quickly reduced the presence of methacrylate monomer but also exhibited diminishing returns beyond 60 minutes of treatment. Finally, temporary immersion in ethanol was shown to be more effective than soaking in PBS alone.

Cell viability

Microfluidic devices and cell culture platforms typically operate in the presence of warm cell culture media in a humidified environment. In this environment, residual monomer/oligomer from SLA 3D-printed devices may diffuse into cell culture medium and induce toxicity and cell death, initially observable as a contracted cytoplasm surrounding a shrunken nucleus (Fig. 4b).

Using UV-vis spectrophotometric data as a guide, we wanted to see how well the removal of these chemical species observed previously correlated with cell viability. Samples from each group were incubated in cell culture medium for 24 hours, and then this medium was extracted and deposited into wells (n = 4 for each group) in a 96-well plate. HUVECs were introduced into each well and exposed to the extraction media for a period of two days, and an MTT assay was subsequently used to quantify cytotoxicity (Fig. 5). Soaking the samples in PBS, even for a period of 5 days, was found to be insufficient to prevent cell death, which correlated well with UV-vis data that indicated that methacrylate monomer/oligomer was still detected after several days. Immersion in ethanol for just 10 minutes prior to soaking in PBS resulted in a minor improvement in cell viability. Baking at 120 °C and UV treatment with heat were the most effective techniques for creating a biocompatible surface for cell culture, matching closely with the corresponding UV spectra, which indicated that essentially no monomer/oligomer was detectable. Following this experiment, a subsequent MTT assay was run on selected groups for 10 days to determine whether any additional cytotoxicity resulted from longer-term exposure. Overall trends observed after 10 days of exposure were roughly similar to those found after 2 days.

Mechanical properties

The Young's and flexural modulus were evaluated to determine whether different post-processing strategies could be utilized to alter the mechanical properties of the material. Objects first printed using this method are only lightly crosslinked due to the relatively fast exposure time, resulting in a fairly flexible material. Upon UV post-curing, the material becomes rigid and brittle as crosslinking increases in the polymer network through polymerization of residual monomer/oligomer. The rationale postulated here was that removing monomer prior to post-curing would result in less polymerizable chemical species and thus a more compliant material. We were also interested in the relative compliance of the material polymerized through heat alone. The results (Fig. 6) align closely with the theoretical prediction. Namely, immediate UV polymerization when all the chemical species were present in the network resulted in the highest elastic and flexural modulus, and removing chemical species through soaking prior to post-curing the material resulted in commensurately lower rigidity. Immersion in ethanol prior to soaking in PBS caused the polymer network to swell, increasing diffusion of the chemical species out of the polymer network and resulting in higher compliance than soaking in PBS alone. Each successive time point at 120 °C exhibited a proportionately higher modulus, reaching a maximum after 24 hours, at which point the structure was nearly completely polymerized. Interestingly, while the flexural modulus values exhibited the same overall trends, the values for unpolymerized groups were far lower than those measured through tensile testing.

Metrology

Finally, the dimensions of samples processed in each condition were measured to see if they caused changes as a result of the various treatments (Fig. 7). Most groups varied less than 2% from untreated samples with the exception of baking for 24 hours at 120 °C, which exhibited minor shrinking.

Discussion

As photopolymerization-based additive manufacturing techniques become more commonly used for the fabrication of microfluidic devices and other cell culture platforms, it will be important for researchers to understand the cell–material interactions and how to tailor material properties for their project goals. Although this study focuses on just one of many resins, each with different chemical compositions and material properties, the approach should be applicable to most of the other materials due to the fact that most resins are comprised of polymerizable methacrylate or acrylate groups. Additionally, treatments utilizing heat or solvent immersion can be considered generally applicable for the removal of cytotoxic chemical species. Here we show that post-processing techniques can be utilized to effectively modulate the properties of SLA 3D-printed constructs, enabling cytocompatibility using a resin that was previously cytotoxic. Moreover, the approach can be used to control the mechanical properties that may be useful for the fabrication of microfluidic components such as valves. Overall, the theoretical predictions mapped well onto the UV-vis spectral data, which subsequently correlated with cell viability and the mechanical properties. We hypothesized that swelling the polymer network in a solvent with high affinity to the monomer would increase the diffusion of chemical species out of the polymer network, which was born out in both a lower UV chemical signature and an increased mechanical compliance. These observations will be useful for researchers who want to decrease the stiffness of their 3D printed constructs. The UV spectra representing samples baked at 120 °C suggested that it would be an effective method for the removal of toxic chemical species, which correlated with the higher cell viability observed. Moreover, we demonstrated that the degree of crosslinking (and thus the mechanical rigidity) can be tailored with the amount of time samples are exposed to heat. While researchers may be tempted to follow manufacturer's instructions when fabricating devices, our study shows that there is additional methodological flexibility that can be exploited for particular use-cases. Instead of abandoning a resin that may have other desirable characteristics, researchers can utilize these techniques to modify their current systems accordingly. We hope this work will aid researchers in understanding the mechanisms that guide the material properties of SLA 3D printed constructs and how to tailor those properties for their individual needs.

Methods

MSLA 3D printing

All objects described in this study were designed using Fusion 360 (Autodesk) and sliced using Chitubox with a layer height of 50 μm, an exposure time of 9.5 seconds, and a bottom exposure time of 60 seconds. Objects were created with a Sonic Mini 8k (Phrozen) MSLA 3D printer at 405 nm with a build plate modified by the addition of a 0.25′′ glass slab for smoother object surfaces, adhered using double-sided tape.

UV-vis spectrophotometry

Cylindrical constructs (diameter = 10 mm, height = 5 mm) were fabricated, cleaned by immersion in ethanol for 2 minutes, and allowed to dry for 10 minutes before being post-processed in the following conditions: (1) UV radiation, carried out using a Form Cure (Formlabs) at a wavelength of 405 nm and 60 °C; (2) baking at 120 °C using a standard laboratory oven; (3) immersion in ethanol for 10 minutes, followed by soaking at 4 mL PBS/construct (fluid changed daily); (4) soaking in PBS at 4 mL/construct (fluid changed daily). PBS was used as an extraction medium for methacrylate monomer/oligomer and collected for analysis with UV-vis spectrophotometry. UV radiation and baking groups were immersed in 4 mL PBS/construct overnight, while PBS was collected directly from groups that were soaked at each time point. A Lambda 950 UV-vis-NIR spectrophotometer (Perkin Elmer) equipped with a 2D detector was used to collect spectra from 185–450 nm.

Preparation of leachate extracts

Cylindrical constructs (n = 4 for each group) were fabricated and post-processed using the conditions described above. Constructs were sterilized via immersion in 100% ethanol for 2 minutes and cleaned using two subsequent PBS baths for 3 minutes each before being submerged in 1 mL Vascular Cell Basal Medium (PCS-100-030™, ATCC) and placed in a standard cell culture incubator. Extraction medium for each group was collected after 24 hours.

Cell culture

Primary human umbilical vein endothelial cells (PCS-100-010™, ATCC) were cultured in Vascular Cell Basal Medium (PCS-100-030™, ATCC) supplemented with Endothelial Cell Growth Kit-VEGF (PCS-100-041™, ATCC) and penicillin–streptomycin (5000 U mL⁻¹). Cells were introduced into a 96-well plate at a concentration of 1 × 10⁴ cells per well and cultured at 37 °C, 100% humidity, and 5% CO₂ using 150 μL of the extraction medium from each treatment group. After 48 hours, cell culture media was aspirated and 50 μL of the MTT reagent (MTT Assay Kit, Abcam, ab211091) and 50 μL of serum-free medium (Vascular Cell Basal Medium described above without the addition of FBS or ascorbic acid) was introduced into each well. Cells were incubated for 3 hours, fluid was removed from each well before adding the MTT solvent and then placed on an orbital shaker for 15 minutes. Absorbance was read using a plate reader at 590 nm, and values were converted to percent cytotoxicity following manufacturer's instructions from the assay kit using control values as a baseline for 0% cytotoxicity. For MTT data collected over 10 days, cylinders were exposed to fresh cell culture media every 48 hours to extract chemical species and the 150 μL of this extraction medium was used to change the media on the respective groups. Data are presented as mean ± standard error of the mean (SEM).

Mechanical properties

To measure Young's modulus, dogbone structures (6.5 mm width, 10 mm length, 0.75 mm thickness) were printed and processed according to the conditions described above (n = 4 samples for each group). Tensile testing was carried out using an Instron 5969 equipped with a 1 kN load cell and samples were stretched at a rate of 30 mm min⁻¹. To measure flexural modulus, rectangular slabs (10 mm × 40 mm × 1.5 mm) were printed as described above and bent at 2 mm min⁻¹ using a 3-point bend assembly. Data are expressed as mean ± standard deviation (SD).

Metrology

10 mm cubes were 3D printed, processed in selected conditions (n = 3 for each group), and measured with calipers. Data are expressed as mean ± standard deviation (SD).

Data availability

Data for this article are available at Open Science Framework at https://osf.io/pa5f4/files/osfstorage.

Conflicts of interest

There are no conflicts of interest to declare.

References

1. G. M. Whitesides, E. Ostuni, S. Takayama, X. Jiang and D. E. Ingber, Soft lithography in biology and biochemistry, Annu. Rev. Biomed. Eng., 2001, 3(1), 335–373 CrossRef CAS PubMed .
2. Y. Xia and G. M. Whitesides, Soft lithography, Annu. Rev. Mater. Sci., 1998, 28(1), 153–184 CrossRef CAS .
3. M. W. Toepke and D. J. Beebe, PDMS absorption of small molecules and consequences in microfluidic applications, Lab Chip, 2006, 6(12), 1484–1486 RSC .
4. A. V. Nielsen, M. J. Beauchamp, G. P. Nordin and A. T. Woolley, 3D Printed Microfluidics, Annu. Rev. Anal. Chem., 2019, 13(1), 1–21 Search PubMed .
5. R. D. Sochol, E. Sweet, C. C. Glick, S. Y. Wu, C. Yang and M. Restaino, et al., 3D printed microfluidics and microelectronics, Microelectron. Eng., 2018, 189, 52–68 CrossRef CAS .
6. A. K. Au, W. Huynh, L. F. Horowitz and A. Folch, 3D-Printed Microfluidics, Angew. Chem., Int. Ed., 2016, 55(12), 3862–3881 CrossRef CAS PubMed .
7. C. M. B. Ho, S. H. Ng, K. H. H. Li and Y. J. Yoon, 3D printed microfluidics for biological applications, Lab Chip, 2015, 15(18), 3627–3637 RSC .
8. H. Gong, B. P. Bickham, A. T. Woolley and G. P. Nordin, Custom 3D printer and resin for 18 μm × 20 μm microfluidic flow channels, Lab Chip, 2017, 17(17), 2899–2909 RSC .
9. H. Gong, A. T. Woolley and G. P. Nordin, High density 3D printed microfluidic valves, pumps, and multiplexers, Lab Chip, 2016, 16(13), 2450–2458 RSC .
10. M. S. Viglione, A. Saxton, D. Downs, A. T. Woolley, K. A. Christensen, P. M. Van Ry and G. P. Nordin, Integrated biocompatible 3D printed isoporous membranes with 7 μm pores, Lab Chip, 2024, 24(8), 2202–2207 RSC .
11. Y. S. Lee, N. Bhattacharjee and A. Folch, 3D-printed Quake-style microvalves and micropumps, Lab Chip, 2018, 18(8), 1207–1214 RSC .
12. N. Bhattacharjee, C. Parra-Cabrera, Y. T. Kim, A. P. Kuo and A. Folch, Desktop-Stereolithography 3D-Printing of a Poly(dimethylsiloxane)-Based Material with Sylgard-184 Properties, Adv. Mater., 2018, 30(22), e1800001 CrossRef PubMed .
13. M. D. Nelson, N. Ramkumar and B. K. Gale, Flexible, transparent, sub-100 μm microfluidic channels with fused deposition modeling 3D-printed thermoplastic polyurethane, J. Micromech. Microeng., 2019, 29(9), 095010 CrossRef CAS .
14. L. A. Milton, M. S. Viglione, L. J. Ong, G. P. Nordin and Y. C. Toh, Vat photopolymerization 3D printed microfluidic devices for organ-on-a-chip applications, Lab Chip, 2023, 23(16), 3537–3560 RSC .
15. G. Taormina, C. Sciancalepore, M. Messori and F. Bondioli, 3D printing processes for photocurable polymeric materials: technologies, materials, and future trends, J. Appl. Biomater. Funct. Mater., 2018, 16(3), 151–160 CAS .
16. E. Yoshii, Cytotoxic effects of acrylates and methacrylates: Relationships of monomer structures and cytotoxicity, J. Biomed. Mater. Res., 1997, 37(4), 517–524 CrossRef CAS PubMed .
17. M. J. Männel, C. Fischer and J. Thiele, A Non-Cytotoxic Resin for Micro-Stereolithography for Cell Cultures of HUVECs, Micromachines, 2020, 11(3), 246 CrossRef PubMed .
18. E. G. Garza, A. Wadajkar, C. Ahn, Q. Zhu, L. A. Opperman and L. L. Bellinger, et al., Cytotoxicity evaluation of methacrylate-based resins for clinical endodontics in vitro, J. Oral Sci., 2012, 54(3), 213–217 CrossRef CAS PubMed .
19. D. A. Azhar, S. Syed, M. Luqman and A. A. Ali, Evaluation of methyl methacrylate monomer cytotoxicity in dental lab technicians using buccal micronucleus cytome assay, Dent. Mater. J., 2013, 32(3), 519–521 CrossRef CAS PubMed .
20. B. Venzac, S. Deng, Z. Mahmoud, A. Lenferink, A. Costa and F. Bray, et al., PDMS Curing Inhibition on 3D-Printed Molds: Why? Also, How to Avoid It?, Anal. Chem., 2021, 93(19), 7180–7187 CrossRef CAS PubMed .
21. K. Piironen, M. Haapala, V. Talman, P. Järvinen and T. Sikanen, Cell adhesion and proliferation on common 3D printing materials used in stereolithography of microfluidic devices, Lab Chip, 2020, 20(13), 2372–2382 RSC .
22. S. Kreß, R. Schaller-Ammann, J. Feiel, J. Priedl, C. Kasper and D. Egger, 3D Printing of Cell Culture Devices: Assessment and Prevention of the Cytotoxicity of Photopolymers for Stereolithography, Materials, 2020, 13(13), 3011 CrossRef PubMed .
23. X.-S. Chai, F. J. Schork and E. M. Oliver, ATR-UV monitoring of methyl methacrylate miniemulsion polymerization for determination of monomer conversion, J. Appl. Polym. Sci., 2006, 99(4), 1471–1475 CrossRef CAS .
24. H. Kaczmarek, Effect of Irgacure 651 Initiator on Poly(Methyl Methacrylate) Photostability Studied by UV-Vis Spectroscopy, Open Process Chem. J., 2013, 1(1), 8–11 CrossRef .

This journal is © The Royal Society of Chemistry 2024