I. INTRODUCTION

Section:

ChooseTop of pageABSTRACTI. INTRODUCTION <<II. MATERIALS AND METHODSIII. RESULTS AND DISCUSSI...IV. CLOSING REMARKSREFERENCESCancer progression can be characterized by the aggressive proliferation and chaotic activities of cells that disrupt the tissue architecture, thereby leading to improper organ function.1,21. D. Hanahan and R. A. Weinberg, Cell 144, 646 (2011). https://doi.org/10.1016/j.cell.2011.02.0132. D. Hanahan, Cancer Discov. 12, 31 (2022). https://doi.org/10.1158/2159-8290.CD-21-1059 Fundamentally, cancer invasion is a three-dimensional (3D) process. In some cancers, the process stimulates the initiation of metastasis to distant organs. It is widely suggested that tumor invasion is triggered by multifactorial determinants including accumulated mutations,33. I. Bozic et al., Proc. Natl. Acad. Sci. U.S.A. 107, 18545 (2010). https://doi.org/10.1073/pnas.1010978107 growth-induced competition for space,44. J. M. Tse et al., Proc. Natl. Acad. Sci. U.S.A. 109, 911 (2012). https://doi.org/10.1073/pnas.1118910109 deprivation of nutrients and oxygen,55. J. H. Garcia, S. Jain, and M. K. Aghi, Front. Cell Dev. Biol. 9, 683276 (2021). https://doi.org/10.3389/fcell.2021.683276 reduction of intercellular adhesion,66. M. Janiszewska, M. C. Primi, and T. Izard, J. Biol. Chem. 295, 2495 (2020). https://doi.org/10.1074/jbc.REV119.007759 and increasing interactions with the surrounding extracellular matrix (ECM);77. V. Gkretsi and T. Stylianopoulos, Front. Oncol. 8, 145 (2018). https://doi.org/10.3389/fonc.2018.00145 however, detailed mechanism behind each determinant has remained elusive. Efforts in computational modeling suggest that invasion is an emergent property of the collective cell response to environmental cues.88. A. R. A. Anderson, A. M. Weaver, P. T. Cummings, and V. Quaranta, Cell 127, 905 (2006). https://doi.org/10.1016/j.cell.2006.09.042 Additionally, the reduction in intercellular adhesion, caused by genomic instability and various stress factors, may further promote cancer cells to increasingly interact with their physical surroundings. Such interactions reciprocally stimulate cell-ECM engagement and cell migration out of the tumor mass. Once initiated, the invasive cancer cells, individually or collectively, penetrate the ECM and spread across the tumor microenvironment (TME).Cancer cell migration is thought to be regulated not only by cell-generated contractility and proteolytic activity but also by extrinsic factors such as the presence of adhesion molecules,66. M. Janiszewska, M. C. Primi, and T. Izard, J. Biol. Chem. 295, 2495 (2020). https://doi.org/10.1074/jbc.REV119.007759 soluble factors (e.g., chemotaxis), and topographical and mechanical cues (e.g, durotaxis).7,9–117. V. Gkretsi and T. Stylianopoulos, Front. Oncol. 8, 145 (2018). https://doi.org/10.3389/fonc.2018.001459. R. J. Pelham and Y. L. Wang, Proc. Natl. Acad. Sci. U.S.A. 94, 13661 (1997). https://doi.org/10.1073/pnas.94.25.1366110. C. Rianna, P. Kumar, and M. Radmacher, Semin. Cell Dev. Biol. 73, 107 (2018). https://doi.org/10.1016/j.semcdb.2017.07.02211. P. P. Provenzano, K. W. Eliceiri, J. M. Campbell, D. R. Inman, J. G. White, and P. J. Keely, BMC Med. 4, 1 (2006). https://doi.org/10.1186/1741-7015-4-38 Due to the dense network of fibrous ECM proteins, early pioneering cancer cells must degrade the ECM via membrane-bound and secreted enzymes (e.g., Matrix metalloproteinases—MMPs) to create space for growth and migration.12–1412. S. Kumar, A. Das, A. Barai, and S. Sen, Biophys. J. 114, 650 (2018). https://doi.org/10.1016/j.bpj.2017.11.377713. A. Das, M. Monteiro, A. Barai, S. Kumar, and S. Sen, Sci. Rep. 7, 1 (2017). https://doi.org/10.1038/s41598-017-14340-w14. K. Wolf et al., J. Cell Biol. 201, 1069 (2013). https://doi.org/10.1083/jcb.201210152 Existing 3D tumor invasion models employ enzymatically degradable hydrogel platforms (e.g., Matrigel™ or collagen) to emulate the migratory processes. Furthermore, contact guidance, imposed by the topographical features of a substrate, has been well-known to direct cell migration.11,1511. P. P. Provenzano, K. W. Eliceiri, J. M. Campbell, D. R. Inman, J. G. White, and P. J. Keely, BMC Med. 4, 1 (2006). https://doi.org/10.1186/1741-7015-4-3815. A. Ray et al., Nat. Commun. 8, 14923 (2017). https://doi.org/10.1038/ncomms14923 For instance, cancer cells, both individually and collectively, have been reported to align their migration with collagen fibers to facilitate local invasion.1616. W. Han et al., Proc. Natl. Acad. Sci. U.S.A. 113, 11208 (2016). https://doi.org/10.1073/pnas.1610347113 Nevertheless, cancer cell migration is quite dynamic and constantly evolving. While cancer invasion is critically dependent on proteolytic enzymes, an increasing number of studies have reported the dynamic shift to nonproteolytic modes of invasion.17–2117. K. M. Yamada and M. Sixt, Nat. Rev. Mol. Cell Biol. 20, 738 (2019). https://doi.org/10.1038/s41580-019-0172-918. L. Wullkopf et al., Mol. Biol. Cell 29, 2378 (2018). https://doi.org/10.1091/mbc.E18-05-031919. L. C. Kelley et al., Dev. Cell 48, 313 (2019). https://doi.org/10.1016/j.devcel.2018.12.01820. P. Strzyz, Nat. Rev. Mol. Cell Biol. 20, 136 (2019). https://doi.org/10.1038/s41580-019-0104-821. C. D. Paul, P. Mistriotis, and K. Konstantopoulos, Nat. Rev. Cancer 17, 131 (2017). https://doi.org/10.1038/nrc.2016.123 Cancer cells can squeeze through accessible pores of collagen networks or confined microchannels independent of proteolytic activities by adopting amoeboid migration.21–2521. C. D. Paul, P. Mistriotis, and K. Konstantopoulos, Nat. Rev. Cancer 17, 131 (2017). https://doi.org/10.1038/nrc.2016.12322. F. Sabeh, R. Shimizu-Hirota, and S. J. Weiss, J. Cell Biol. 185, 11 (2009). https://doi.org/10.1083/jcb.20080719523. A. W. Holle et al., Nano Lett. 19, 2280 (2019). https://doi.org/10.1021/acs.nanolett.8b0472024. D. Fanfone et al., Elife 11, e73150 (2022). https://doi.org/10.7554/eLife.7315025. Y. J. Liu et al., Cell 160, 659 (2015). https://doi.org/10.1016/j.cell.2015.01.007 Thus, cancer cells can dynamically adapt various modes of migration in response to heterogeneous environmental cues,17,1817. K. M. Yamada and M. Sixt, Nat. Rev. Mol. Cell Biol. 20, 738 (2019). https://doi.org/10.1038/s41580-019-0172-918. L. Wullkopf et al., Mol. Biol. Cell 29, 2378 (2018). https://doi.org/10.1091/mbc.E18-05-0319 yet the underlying mechanism is not clearly demonstrated. The continued development of new biomaterials and biointerfaces for studying tumor invasion provides additional insight into these mechanisms from different perspectives.Many aspects of tumor invasion are modulated through interactions with the ECM, which represents a complex 3D-compliant network of biological interfaces (e.g., collagens, fibronectin, elastin, and proteoglycans).26–2826. C. Frantz, K. M. Stewart, and V. M. Weaver, J. Cell Sci. 123, 4195 (2010). https://doi.org/10.1242/jcs.02382027. J. Kim et al., “Collective ECM remodeling organizes 3D collective cancer invasion,” pp. 1–6 (2019). http://arxiv.org/abs/1903.03290.28. F. A. Venning, L. Wullkopf, and J. T. Erler, Front. Oncol. 5, 224 (2015). https://doi.org/10.3389/fonc.2015.00224 The ECM is an aqueous, tissue-specific network formed from proteins and polysaccharides that is strikingly heterogeneous. Proteoglycans and fibrous proteins are the two main classes of biomacromolecules of the ECM.2626. C. Frantz, K. M. Stewart, and V. M. Weaver, J. Cell Sci. 123, 4195 (2010). https://doi.org/10.1242/jcs.023820 Recent advancements in biotechnologies and material science have made progress in developing materials that are recapitulative of in vivo ECM to study cancer progression. Multifunctional hydrogels and microgels have emerged as the main classes of materials employed for the development of novel three-dimensional preclinical models.29–3129. K. Wagner et al., Soft Matter 15, 9776 (2019). https://doi.org/10.1039/C9SM01226E30. C. Jensen and Y. Teng, Front. Mol. Biosci. 7, 1 (2020). https://doi.org/10.3389/fmolb.2020.0003331. S. R. Caliari and J. A. Burdick, Nat. Methods 13, 405 (2016). https://doi.org/10.1038/nmeth.3839 As compared to conventional cell monolayer models, these 3D in vitro models of invasion facilitate more realistic tumor architectures and microenvironments that faithfully recapitulate in vivo conditions.32–3432. J. Drost and H. Clevers, Nat. Rev. Cancer 18, 407 (2018). https://doi.org/10.1038/s41568-018-0007-633. D. Anton, H. Burckel, E. Josset, and G. Noel, Int. J. Mol. Sci. 16, 5517 (2015). https://doi.org/10.3390/ijms1603551734. D. T. Nguyen et al., Cells 11, 1 (2022). 10th Anniversary of Cells—Advances in Cell Microenvironment. In previous studies, we demonstrated a liquid-like solid (LLS) platform for the 3D culture of cells and microtissues.34,3534. D. T. Nguyen et al., Cells 11, 1 (2022). 10th Anniversary of Cells—Advances in Cell Microenvironment.35. M. A. Schaller et al., JCI Insight 6, 1 (2021). https://doi.org/10.1172/jci.insight.148003 The LLS is made from an ensemble of soft aqueous microgels3636. T. Bhattacharjee et al., ACS Biomater. Sci. Eng. 2, 1787 (2016). https://doi.org/10.1021/acsbiomaterials.6b00218 that provides an accessible platform for in situ observations of drug screening and disease pathogenesis. The LLS can be made of high-water-content hydrogel materials37–4037. C. D. Morley et al., Nat. Commun. 10, 3029 (2019). https://doi.org/10.1038/s41467-019-10919-138. C. S. O’Bryan et al., MRS Bull. 42(8), 1 (2017). https://doi.org/10.1557/mrs.2017.16739. T. Bhattacharjee et al., Sci. Adv. 1, e1500655 (2015). https://doi.org/10.1126/sciadv.150065540. K. J. Leblanc et al., ACS Biomater. Sci. Eng. 2, 1796 (2016). https://doi.org/10.1021/acsbiomaterials.6b00184 including polyethylene glycol (PEG), polyacrylamide, and others. The sizes of these microgels are designed with characteristic interstitial spaces through which cells can crawl and liquid media can be perfused.Glioblastoma is a particularly aggressive and fast-growing form of brain cancer with an abysmal median survival of under 2 years.5,415. J. H. Garcia, S. Jain, and M. K. Aghi, Front. Cell Dev. Biol. 9, 683276 (2021). https://doi.org/10.3389/fcell.2021.68327641. S. J. Price and J. H. Gillard, Br. J. Radiol. 84, S159 (2011). https://doi.org/10.1259/bjr/26838774 A key characteristic of glioblastoma is its invasion and general lack of a defined tumor margin. Postmortem studies have shown that 20%–27% of glioblastomas invade as much as 10 mm,42,4342. H. J. Scherer, Brain 63, 1 (1940). https://doi.org/10.1093/brain/63.1.143. P. C. Burger, E. R. Heinz, T. Shibata, and P. Kleihues, J. Neurosurg. 68, 698 (1988). https://doi.org/10.3171/jns.1988.68.5.0698 while approximately another 20% show extensive invasion greater than 30 mm,4343. P. C. Burger, E. R. Heinz, T. Shibata, and P. Kleihues, J. Neurosurg. 68, 698 (1988). https://doi.org/10.3171/jns.1988.68.5.0698 and a further 8% show grossly disseminated spread.41,4441. S. J. Price and J. H. Gillard, Br. J. Radiol. 84, S159 (2011). https://doi.org/10.1259/bjr/2683877444. A. T. Parsa et al., J. Neurosurg. 102, 622 (2005). https://doi.org/10.3171/jns.2005.102.4.0622 Due to the aggressiveness of the disease, glioblastoma has been widely employed as an in vitro model for tumor invasion. In this study, we report on the development of tunable biointerfaces on a 3D LLS medium with a controlled density of ECM biomacromolecules to facilitate the studies of glioblastoma invasion. In particular, we successfully functionalized and conjugated type 1 collagen onto the surface of LLS microgel particles (COL1-LLS) to enable cell adhesion. The interstitial space between the microgels formed randomly interconnected 3D microchannel networks, allowing cancer cell invasion without the need for ECM degradation. The microgel particles, made of polyacrylamide, were sized to facilitate an interstitial space on the order of 7–10 μm to mimic the characteristics of a 3D capillary network.4545. D. F. J. Tees, P. Sundd, and D. J. Goetz, “10 - A flow chamber for capillary networks: Leukocyte adhesion in capillary-sized, ligand-coated micropipettes,” in Principles of Cellular Engineering, edited by M. R. KING (Academic Press, 2006), pp. 213–231. https://doi.org/10.1016/B978-012369392-1/50011-5 The microgel particles gravitationally settle to form a solid bed of yield-stress fluid that stably supports the tumors in 3D. The high-water-content LLS medium has a yield stress of less than 10 Pa, allowing for essentially unrestricted tumor growth and expansion. This platform enabled 3D tumor invasion via cell anchorage-dependent migration and geometrical guidance from the physical surroundings.

II. MATERIALS AND METHODS

Section:

ChooseTop of pageABSTRACTI. INTRODUCTIONII. MATERIALS AND METHODS <<III. RESULTS AND DISCUSSI...IV. CLOSING REMARKSREFERENCES

A. Cell culture

High-grade murine glioma cell line Kr158B was kindly provided by Dr. Elias Sayour, MD. The cell line was cultured in a growth medium containing DMEM without sodium pyruvate (Gibco, Billings, MT, USA, 11965-092), 10% fetal bovine serum (FBS) (Sigma-Aldrich, Sigma-Aldrich, St. Louis, MO, USA, F4135), and 1% penicillin/streptomycin (Gibco, Billings, MT, USA, 15140148) in a T75 cell culture flask at 37 °C and 5% CO2. For spheroid generation, we employed a perfusion culture method as previously published in Ref. 3434. D. T. Nguyen et al., Cells 11, 1 (2022). 10th Anniversary of Cells—Advances in Cell Microenvironment.. In short, Kr158B cells were enzymatically detached from cell culture flasks and suspended in inert LLS at a density of 106 cells/ml. The cell-LLS mixture was dispensed into each well (200 μl each) of a 24-well Darcy perfusion plate3434. D. T. Nguyen et al., Cells 11, 1 (2022). 10th Anniversary of Cells—Advances in Cell Microenvironment. and cultured at a perfusion flow rate of ∼40 μl/h/well at 37 °C, 5% CO2. Cellular spheroids formed within 72 h of culture. When the spheroid diameter reached 400–600 μm in diameter (∼day 7), the samples were harvested for invasion study.

B. Fabrication of protein-conjugated polyacrylamide LLS microgels

Acrylamide (AAm) monomer and N-acryloxysuccinimide (NAS) (at 5 wt. % total monomer concentration) were copolymerized with N,N′-methylenebisacrylamide crosslinker (BIS) (0.2 wt. %). The polymerization was initiated by ammonium persulfate and catalyzed by tetramethylethylenediamine via free radical polymerization. The polymerized P(AAm-co-NAS) hydrogel was mechanically ruptured to create microgel particles. The microgel size was further homogenized to a target mean and standard deviation with a 95% confidence interval via centrifugation. The P(AAm-co-NAS) microgels functionalized with NAS enable covalent modification via the ε-amino (NH2) group of lysine residues and the terminal NH2 groups present in target proteins. For collagen-conjugated LLS (COL1-LLS), 10 μg/ml of type I collagen (MW: 300 kDa, Nutragen, Cat. 5010) or 33.3 nM per every ml of LLS was used during the conjugation process. To prevent hydrolysis of unconjugated NAS groups to acrylic acid, ethanolamine was added postconjugation to form biologically inert hydroxylethyl acrylamide.

Alternatively, AAm monomer 4%w/v, BIS 0.3%w/v, and acrylic acid (AA) 1%w/v hydrogel were prepared via free radical polymerization in pH 5.5 MES buffer (0.7 M). The particles were then activated in a solution of EDC and NHS. The activated microgels were conjugated with collagen I in PBS for 2 h and subsequently quenched with ethanolamine.

C. 3D invasion assay

The COL1-LLS was equilibrated in cell culture media in 37 °C water bath for 30 min prior to the experiment. The solution was then centrifuged at 1000×g for 5 min to tightly consolidate the COL1-LLS. Upon centrifugation, the liquid media was removed leaving the media-infused and consolidated COL1-LLS at the bottom of the tube. Using a prewet 200 μl-wide bore pipette tip, a volume of 100 μl of COL1-LLS was deposited into the respective well of a glass-bottom 96-well plate. The tumor spheroids were then gently positioned into the COL1-LLS gel such that they were optically accessible. The 96-well plate was then centrifuged at 100×g for 5 min at a low deceleration and acceleration setting. Using a 25G syringe needle, 100 μl of media was gently added on top of the well. The plate was then secured in a custom-built stage incubator of a Nikon A1R HD25 confocal microscope for in situ time-lapse acquisition.

D. Immunofluorescence assay

The immunofluorescence (IF) staining protocol has been previously described.34,3534. D. T. Nguyen et al., Cells 11, 1 (2022). 10th Anniversary of Cells—Advances in Cell Microenvironment.35. M. A. Schaller et al., JCI Insight 6, 1 (2021). https://doi.org/10.1172/jci.insight.148003 In short, cells were fixed in 4.0% formaldehyde (Fisher Scientific, Waltham, MA, USA, BP531-500) in 1× PBS overnight at 4 °C, washed twice, and incubated in PBS for 1 h at room temperature. The samples were then permeabilized in 0.5% Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA, X100-100ML) for 2 h, washed twice, and blocked with 3% bovine serum albumin in PBS for 3 h at room temperature. After blocking, samples were washed three times with PBS and then incubated overnight with conjugated antibodies at 4 °C. The samples were then stained with Invitrogen™ Alexa Fluor™ 568 Phalloidin (InvitrogenTM, Waltham, MA, USA, A12380). After overnight incubation with the antibodies, the samples were washed three times with PBS and counterstained with Hoechst 33342 (InvitrogenTM, Waltham, MA, USA, H3570) for 10 min before imaging.

E. Microscopy

To capture time-lapse images of glioblastoma invasion, cellular spheroids were fluorescently dyed with Cell Tracker Orange CMRA (5 μM) (Thermo Fisher Scientific, Waltham, MA, USA, C34551) for 30 min in growth media at 37°C and 5% CO2. Samples were then rinsed with 1× PBS three times. All samples were imaged using a Nikon A1R HD25 confocal microscope equipped with a high-definition Galvano scanner.

III. RESULTS AND DISCUSSION

Section:

ChooseTop of pageABSTRACTI. INTRODUCTIONII. MATERIALS AND METHODSIII. RESULTS AND DISCUSSI... <<IV. CLOSING REMARKSREFERENCESIn this study, we developed a protein-conjugated liquid-like solid (LLS) from a polyacrylamide microgel ensemble. We functionalized the microgels by copolymerizing acrylamide monomers and acrylic acid N-hydroxysuccinimide esters (N-acryloxysuccinimide) with N,N′-methylenebisacrylamide crosslinker. The bioconjugated microgels were also produced via EDC/NHS coupling. In this case, we copolymerized acrylamide with acrylic acid that were then activated in situ via EDC/NHS chemistry to form the pendant activated NHS esters. The free radical polymerization was initiated by ammonium persulfate and catalyzed by a tetramethylethylenediamine (TEMED) reductant. The NHS-polyacrylamide hydrogel was mechanically disintegrated into microgel particles whose surfaces were further conjugated with an extracellular matrix protein to promote cell adhesion (Fig. 1). The amine-reactive NHS esters enabled covalent conjugations with primary amines at the N-terminus and lysine amino acid residues present on all proteins. The major advantage of the EDC/NHS method compared to the one containing N-acryloxysuccinimide is the tunability of the pH for each step, reducing the risk of hydrolysis and increasing the yield of the bioconjugation reaction. This mechanism allowed for the control of different proteins to be conjugated on LLS microgels, potentiating the recapitulation of tissue-specific ecosystems in vitro via various protein compositions. In this study, type I collagen (COL1: ∼300 kDa molecule4646. E. M. Bueno and J. W. Ruberti, J. Memb. Sci. 321, 250 (2008). https://doi.org/10.1016/j.memsci.2008.04.066 with a radius of gyration Rg ∼ 25–36 nm), an abundant ECM protein present in all tissues, was selected as a proof of concept. The surface modification of the microgels and successful conjugation of the proteins were first demonstrated by confocal microscopy (Fig. 2) and the functionality was confirmed by cell attachment and spreading (Fig. 3).The spatial distribution of the conjugated proteins within the microgel particle could be greatly controlled by the polyacrylamide gel mesh size (ξ), which is dependent on total monomer content and the ratio of monomers to crosslinkers.4747. D. I. Pedro et al., Tribol. Lett. 69, 1 (2021). https://doi.org/10.1007/s11249-020-01378-7 These factors are coupled determinants of the elastic modulus, which has been shown to regulate cell adhesion and migratory behavior.4848. J. R. Tse and A. J. Engler, “Preparation of hydrogel substrates with tunable mechanical properties,” in Current Protocols in Cell Biology (Wiley, 2010). https://doi.org/10.1002/0471143030.cb1016s47 The polymer mesh size is essentially defined as the average spacing of all polymer chains. This parameter is integral to the mechanical and transport properties of hydrogels including hydraulic permeability,4949. B. Amsden, Macromolecules 31, 8382 (1998). https://doi.org/10.1021/ma980765f osmotic pressure,5050. P. De Gennes, see https://www.eng.uc.edu/∼beaucag/Classes/Properties/Books/Pierre-giles De Gennes—Scaling concepts in polymer physics-Cornell University Press (1979).pdf for “Scaling Concepts in Polymer Physics.” (1979), p. 324. and elastic modulus.50,5150. P. De Gennes, see https://www.eng.uc.edu/∼beaucag/Classes/Properties/Books/Pierre-giles De Gennes—Scaling concepts in polymer physics-Cornell University Press (1979).pdf for “Scaling Concepts in Polymer Physics.” (1979), p. 324.51. J. M. Urueña, E. O. McGhee, T. E. Angelini, D. Dowson, W. G. Sawyer, and A. A. Pitenis, Biotribology 13, 30 (2018). https://doi.org/10.1016/j.biotri.2018.01.002 Polyacrylamide mess size, typically 1–10 nm,5252. J. M. Urueña, A. A. Pitenis, R. M. Nixon, K. D. Schulze, T. E. Angelini, and W. Gregory Sawyer, Biotribology 1–2, 24 (2015). https://doi.org/10.1016/j.biotri.2015.03.001 is commonly characterized by small-angle x-ray scattering (SAXS).5353. S. Mallam, F. Horkay, A. M. Hecht, and E. Geissler, Macromolecules 22, 3356 (1989). https://doi.org/10.1021/ma00198a029 Polyacrylamide hydrogel mesh size is modulated by the concentration of monomers and crosslinkers.47,5247. D. I. Pedro et al., Tribol. Lett. 69, 1 (2021). https://doi.org/10.1007/s11249-020-01378-752. J. M. Urueña, A. A. Pitenis, R. M. Nixon, K. D. Schulze, T. E. Angelini, and W. Gregory Sawyer, Biotribology 1–2, 24 (2015). https://doi.org/10.1016/j.biotri.2015.03.001 The region of protein conjugation on LLS microgels is highly dependent on the polymer mesh size which is directly associated with elastic modulus. For small proteins (rg ξ), the conjugation is uniform throughout the entire microgel particle. Proteins with characteristic size (e.g., the radius of gyration) larger than the polymer mesh size (rg > ξ) were conjugated at the peripheral regions of the microgels due to particle exclusion. Therefore, surface modification of the polyacrylamide microgels with conjugated proteins was achieved by modulating mesh size to avoid diffusivity of proteins to the inner region of the gel. In this study, we chose a formulation of 4.8% (w/v) Aam, 0.2% (w/v) BIS, and 0.5% (w/v) NHS esters that would yield an elastic modulus of ∼ 1–2 kPa4848. J. R. Tse and A. J. Engler, “Preparation of hydrogel substrates with tunable mechanical properties,” in Current Protocols in Cell Biology (Wiley, 2010). https://doi.org/10.1002/0471143030.cb1016s47 and a mesh size ξ of less than 10 nm after polymerization.4747. D. I. Pedro et al., Tribol. Lett. 69, 1 (2021). https://doi.org/10.1007/s11249-020-01378-7 From the evidence of surface conjugation via confocal microscopy and correlated polymer concentration, we anticipated that the PAAm characteristic mesh size in this study was less than 7 nm. As demonstrated in Fig. 2, while a green fluorescent protein (GFP) (MW ∼ 27 kDa, and Rg of 1.8 nm) diffused through a polymer network and was conjugated in the entire gel, an ITGB1 antibody, (MW ∼ 150 kDa) was conjugated exclusively at the peripheral region on the microgels. Maximum intensity projection confocal images revealed a distinct fluorescent signal at the periphery of the microgels when conjugated with ITGB1 antibody as compared to GFP, Figs. 2(c) and 2(d), respectively.Collagens are the most abundant ECM proteins present in the body and the main component of connective tissue, thus have been widely used in cell adhesion and migration studies.5454. S. Xu et al., J. Transl. Med. 17, 1 (2019). https://doi.org/10.1186/s12967-019-2058-1 In this study, type I collagen was conjugated exclusively on the periphery of the microgel particles Fig. 3(b). To visualize the COL1-LLS, we further incorporated rhodamine B acrylate (Fig. 3(a)]. This helps to confirm the microgel particles and the presence of conjugated type I collagen at the surface [Fig. 3(c)]. Since the conjugation process may induce changes in protein conformation, we cultured adherent glioblastoma cells (Kr158B) in the microgels for several days to test cell adhesion. As shown in Fig. 3(d), the cells formed network-like adhesion around microgel particles and acquired typical adherent morphology with extended actin filaments, demonstrating that the conjugated collagen proteins are functional postconjugation.In this study, we used a glioblastoma model to study invasion dynamics. As previously discussed, glioblastoma is the most aggressive cancer of the nervous system, and the tumor often grows and spreads rapidly.5555. A. Bradshaw et al., Front. Surg. 3, 1203 (2016). https://doi.org/10.3389/fsurg.2016.00048 Glioblastoma tumors are highly invasive and have been reported to adopt an aggressive migratory pattern radially away from the tumor margin.5656. A. Vollmann-Zwerenz, V. Leidgens, G. Feliciello, C. A. Klein, and P. Hau, Int. J. Mol. Sci. 21, 1932 (2020). https://doi.org/10.3390/ijms21061932 In general, cancer cells from solid tumors are adherent but can become anchorage-independent survivors in low-adhesion substrates.5757. M. C. Guadamillas, A. Cerezo, and M. A. del Pozo, J. Cell Sci. 124, 3189 (2011). https://doi.org/10.1242/jcs.072165 In such environments, these cells interact with the neighboring cells and form aggregations for optimal survival. This mechanism has been commonly employed to develop in vitro 3D tumor spheroid models by various techniques (e.g., hanging drop5858. R. Foty, J. Vis. Exp. 20(51), 4 (2011). https://doi.org/10.3791/2720 and floating spheres5959. R. L. F. Amaral, M. Miranda, P. D. Marcato, and K. Swiech, Front. Physiol. 8, 605 (2017). https://doi.org/10.3389/fphys.2017.00605). In this study, the tumor spheroids were formed using a 3D platform for cell culture in LLS by means of perfusion as previously described.3434. D. T. Nguyen et al., Cells 11, 1 (2022). 10th Anniversary of Cells—Advances in Cell Microenvironment. In short, cancer cells were cultured in inert LLS (without bioconjugation) using continuous directional perfusion. The platform ensures a constant supply of growth media and removal of metabolic waste. Due to the lack of adhesion on the surrounding substrate, cancer cells distributed within the inert LLS formed aggregates by day 2 and were grown into 400–600 μm spheroids (∼day 7) (Fig. 4, top row). To investigate the invasion of tumor spheres, we cultured glioblastoma spheroids in COL1- and inert LLS. The tumor evolution was continuously monitored over time by in situ confocal imaging. The tumor spheroids, upon being cultured in COL1-LLS, demonstrated rapid interaction with their surroundings. The cancer cells at the tumor periphery adhered and migrated radially away from the tumor mass (Fig. 4, bottom row). Once the invasion initiated, the glioblastoma cells demonstrated both individual and collective modes of migration to rapidly infiltrate the surrounding.7575. See the supplementary material at https://www.scitation.org/doi/suppl/10.1116/6.0002083 for the videos of tumor invasion in COL1-LLS.Tumor morphology has been commonly used as a diagnostic factor for progression.6060. H. Xiaopei, D. Kunfu, T. Lianyuan, L. Zhen, X. Mei, and Y. Haibo, Eur. Rev. Med. Pharmacol. Sci. 23(22), 9821 (2019). https://doi.org/10.26355/eurrev_201911_19545 Tumors with invasive phenotypes often carry significant mutational burdens, are highly proliferative with increased metabolic and oxygen consumption, demonstrate decreased cell-cell adhesion, and overproduce ECM-degrading enzymes.88. A. R. A. Anderson, A. M. Weaver, P. T. Cummings, and V. Quaranta, Cell 127, 905 (2006). https://doi.org/10.1016/j.cell.2006.09.042 These factors stimulate cancer cells to become increasingly motile and interactive with the surrounding ECM. Various water-soluble (globular) and nonsoluble (fibrous) protein compositions of the ECM present different physical barriers and intrinsic adhesion molecules, thereby promoting heterogeneous modes of tumor invasion.6161. A. G. Clark and D. M. Vignjevic, Curr. Opin. Cell Biol. 36, 13 (2015). https://doi.org/10.1016/j.ceb.2015.06.004 In this study, we examined the nonenzymatic modes of migration by reconstructing the ECM from discrete blocks of collagen-conjugated LLS microgel ensembles. This platform allows for the local discrete stochastic process while ensuring a stable yield-stress property from a global continuum. The glioblastoma cells extended their filopodia (>10 μm), probing for accessible pores, and established firm attachment to their surroundings (Fig. 4, bottom row and Fig. 5(a)]. Once properly adhered, leading cancer cells gradually pulled their body forward into the narrow space between COL1-LLS particles while maintaining appropriate distance with neighboring followers. The collective protrusions dramatically altered the tumor morphology over time, showing remarkably tortuous tumor margins [Fig. 5(b)]. We digitally analyzed tumor tortuosity by evaluating the ratio of 2D projected perimeters to the perimeter of a perfect circle of the same total internal pixel area. The characterization revealed that tumor spheres cultured in COL1-LLS medium (S1 and S2) had significantly higher tortuosity factors (>threefold), increasing monotonically over time, as compared to those cultured in inert LLS (S0) [Fig. 5(c)].In vivo observations suggested that confinement and geometrical architecture of local TME can promote cancer invasion and metastasis.21,2421. C. D. Paul, P. Mistriotis, and K. Konstantopoulos, Nat. Rev. Cancer 17, 131 (2017). https://doi.org/10.1038/nrc.2016.12324. D. Fanfone et al., Elife 11, e73150 (2022). https://doi.org/10.7554/eLife.73150 For instance, cancer cells can migrate along collagen fibers, demonstrating a sufficient mechanism for invasion.11,16,6211. P. P. Provenzano, K. W. Eliceiri, J. M. Campbell, D. R. Inman, J. G. White, and P. J. Keely, BMC Med. 4, 1 (2006). https://doi.org/10.1186/1741-7015-4-3816. W. Han et al., Proc. Natl. Acad. Sci. U.S.A. 113, 11208 (2016). https://doi.org/10.1073/pnas.161034711362. M. W. Conklin et al., Cancer Epidemiol Biomarkers Prev. 27, 138 (2018). https://doi.org/10.1158/1055-9965.EPI-17-0720 In another study, confined microchannels in microfluidic devices enable optimal cell contractility and rapid amoeboid migration.24,25,63,6424. D. Fanfone et al., Elife 11, e73150 (2022). https://doi.org/10.7554/eLife.7315025. Y. J. Liu et al., Cell 160, 659 (2015). https://doi.org/10.1016/j.cell.2015.01.00763. Y. Xia, C. R. Pfeifer, and D. E. Discher, Acta Mech. Sin. Xuebao 35, 299 (2019). https://doi.org/10.1007/s10409-018-00836-964. C. Rianna, M. Radmacher, S. Kumar, and D. Discher, “Direct evidence that tumor cells soften when navigating confined spaces,” Mol. Biol. Cell 31, 1726 (2020). https://doi.org/10.1091/mbc.E19-10-0588 Here, COL1-LLS microgels imposed a 3D channel-like and random network of interstitial space presenting potential paths for cell migration. Despite the various paths of migration, the cancer cells invaded the COL1-LLS medium along predefined directions, independently and collectively in a leader-follower mode of migration.17,6117. K. M. Yamada and M. Sixt, Nat. Rev. Mol. Cell Biol. 20, 738 (2019). https://doi.org/10.1038/s41580-019-0172-961. A. G. Clark and D. M. Vignjevic, Curr. Opin. Cell Biol. 36, 13 (2015). https://doi.org/10.1016/j.ceb.2015.06.004 As shown in Fig. 6, the invasive patterns of cancer cells did not appear to be a randomly diffusive process. Instead, tracking the evolution of invasive paths revealed a super-diffusive (directional) behavior.Since the LLS is made of polyacrylamide and is nondegradable, each microgel locally imposes a physical barrier to cell migration and forces cells to navigate around the tortuous and narrow network of interstitial space between the microgels. We hypothesized that the observed super-diffusive behavior is a result of cancer cells exploring and invading the three-dimensional space via preferential paths.65–6865. E. Toscano, L. Sepe, G. del Giudice, R. Tufano, and G. Paolella, PLoS One 17, e0272259 (2022). https://doi.org/10.1371/journal.pone.027225966. P. Dieterich, R. Klages, R. Preuss, and A. Schwab, Proc. Natl. Acad. Sci. U.S.A. 105, 459 (2008). https://doi.org/10.1073/pnas.070760310567. H. Takagi, M. J. Sato, T. Yanagida, and M. Ueda, PLoS One 3, e2648 (2008). https://doi.org/10.1371/journal.pone.000264868. I. Yurchenko, J. M. Vensi Basso, V. S. Syrotenko, and C. Staii, PLoS One 14, e0216181 (2019). https://doi.org/10.1371/journal.pone.0216181 To test this hypothesis, we performed off-lattice agent-based computer simulations of random walks in a 3D LLS network, which revealed super-diffusive behavior. We have also observed that the local invasion of different tumor models in the COL1-LLS is variable and regulated by several factors—the invasive phenotype, the gel mechanical property, and the size distribution of the LLS which modulates the interstitial space. The role of each parameter in migratory behavior has been investigated and will be reported in an independent study.6969. D. T. Nguyen et al., “Extracellular matrix stiffness and geometrical guidance regulate 3D tumorinvasion,” Manuscr. Prep. (unpublished).

IV. CLOSING REMARKS

Section:

ChooseTop of pageABSTRACTI. INTRODUCTIONII. MATERIALS AND METHODSIII. RESULTS AND DISCUSSI...IV. CLOSING REMARKS <<REFERENCES

A new form of LLS with bioconjugation of COL1 has been described and shown to provide an accessible model to investigate cancer invasion. The in vitro model provides a unique opportunity to study the role of ECM proteins and biointerfaces on cancer invasion dynamics. Glioblastoma invasion is enabled by malignant geno/phenotypes under stress and can be facilitated by adhesion-dependent opportunistic migration into accessible spaces independent of proteolytic activity. Additionally, monotonically increasing tortuosity and super-diffusive behavior of glioblastoma invasion were measured in a COL1-LLS system.

Carcinogenesis and tumor progression occupy all three spatial dimensions, but the necessary infrastructure for establishing relevant 3D in vitro models has proved a significant challenge. In this study, we reconstructed the ECM component of a TME from a bioconjugated liquid-like solid as discrete units of physical support to study tumor invasion. Investigations on mechan