Triciribine | Dna alkylator

Acta Biomaterialia

journal homepage: www.elsevier.com/locate/actbio
Acta Biomaterialia xxx (xxxx) xxx

Full length article
Biomimetic matrix for the study of neuroblastoma cells: A promising combination of stiffness and retinoic acid
Beatrice Labata,∗, Nimrod Buchbinder b, Sandrine Morin-Grognet a, Guy Ladama,
Hassan Atmania, Jean-Pierre Vannierc
a Normandie Univ, UNIROUEN, INSA Rouen, CNRS, PBS UMR 6270, 55 rue Saint-Germain, 27000 Évreux, France
b Pediatric Oncology, CHU Rouen University Hospital, Rouen, France
c Normandie Univ, UNIROUEN, PANTHER – INSERM 1234 – UFR de Médecine et de Pharmacie de Rouen 22, boulevard Gambetta 76000 Rouen, France

a r t i c l e i n f o a b s t r a c t

Article history:
Received 3 May 2021
Revised 11 August 2021
Accepted 12 August 2021

Keywords: Tumor-like ECM Rigidity Retinoic acid Neuritogenesis Genipin

Neuroblastoma is the third most common pediatric cancer composed of malignant immature cells that are usually treated pharmacologically by all trans-retinoic acid (ATRA) but sometimes, they can sponta- neously differentiate into benign forms. In that context, biomimetic cell culture models are warranted tools as they can recapitulate many of the biochemical and biophysical cues of normal or pathological microenvironments. Inspired by that challenge, we developed a neuroblastoma culture system based on biomimetic LbL ﬁlms of physiological biochemical composition and mechanical properties. For that, we used chondroitin sulfate A (CSA) and poly-L-lysine (PLL) that were assembled and mechanically tuned by crosslinking with genipin (GnP), a natural biocompatible crosslinker, in a relevant range of stiffness (30– 160 kPa). We then assessed the adhesion, survival, motility, and differentiation of LAN-1 neuroblastoma cells. Remarkably, increasing the stiffness of the LbL ﬁlms induced neuritogenesis that was strengthened by the combination with ATRA. These results highlight the crucial role of the mechanical cues of the neuroblastoma microenvironment since it can dramatically modulate the effect of pharmacologic drugs. In conclusion, our biomimetic platform offers a promising tool to help fundamental understanding and pharmacological screening of neuroblastoma differentiation and may assist the design of translational biomaterials to support neuronal regeneration.

Statement of signiﬁcance

Neuroblastoma is one of the most common pediatric tumor commonly treated by the administration of all-trans-retinoic acid (ATRA). Unfortunately, advanced neuroblastoma often develop ATRA resistance. Ac- cordingly, in the ﬁeld of pharmacological investigations on neuroblastoma, there is a tremendous need of physiologically relevant cell culture systems that can mimic normal or pathological extracellular matrices. In that context, we developed a promising matrix-like cell culture model that provides new insights on the crucial role of mechanical properties of the microenvironment upon the success of ATRA treatment on the neuroblastoma maturation. We were able to control adhesion, survival, motility, and differenti- ation of neuroblastoma cells. More broadly, we believe that our system will help the design of in vitro pharmacological screening strategy.
© 2021 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction

Neuroblastoma (NB) is the most common extra-cranial pediatric cancer, arising from undifferentiated primitive cells that derived

∗ Corresponding author at: 55 rue St Germain, CS 40486, 27004 Evreux cedex, France.
E-mail address: [email protected] (B. Labat).
from the neural crest. About 40% of patients belong to the high- risk groups. One of the strongest prognostic factors is the ampli- ﬁcation of the MYCN oncogene in malignant cells, which is asso- ciated with unfavorable outcome [1]. The N-Myc oncoprotein was reported to be responsible for impaired sympathetic nervous sys- tem maturation [2]. Neuroblastoma cell lines in culture are highly heterogeneous with three distinct phenotypic groups: (i) tumori- genic N-type cells (Neuroblastic) that can differentiate, (ii) non- tumorigenic S-type cells (Substrate-adherent), and (iii) I-type cells

https://doi.org/10.1016/j.actbio.2021.08.017

1742-7061/© 2021 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Please cite this article as: B. Labat, N. Buchbinder, S. Morin-Grognet et al., Biomimetic matrix for the study of neuroblastoma cells: A promising combination of stiffness and retinoic acid, Acta Biomaterialia, https://doi.org/10.1016/j.actbio.2021.08.017

B. Labat, N. Buchbinder, S. Morin-Grognet et al. Acta Biomaterialia xxx (xxxx) xxx

(Intermediate) that can be induced to convert into either N- or S- type cells, with both tumorigenic and adhesive properties, respec- tively [3,4]. I-type cells organize in spheroids in vitro with tumori- genic phenotype [5] but they can spontaneously differentiate into post-mitotic mature neurons presenting neurite extensions or even axons while regressing to a benign state [6] as they lose their pro- liferative and stemness phenotypes [7].
Therefore, inducing NB cells to differentiate into mature cells is a key therapeutic objective that is currently pharmacologically con- ducted by retinoid treatments that reduce the expression of proto- oncogenes such as MYCN [8] and induce in vitro differentiation of various NB cells. Accordingly, clinical therapeutics like all-trans- retinoic acid (ATRA) and 13-cis-retinoic acid and 9-cis-retinoic acid are the most commonly administered retinoids in pediatric oncol- ogy as a therapy for high-risk NB [9]. For example, overall sur- vival was signiﬁcantly improved with 13-cis-RA after myeloablative therapy [10,11]. Unlike the other RA isomers, 13-cis RA is the less abundant RA isoform, weakly associated with the retinoic acid re- ceptor RAR and requires higher doses to eﬃciently induce NB cell differentiation [12]. In vitro, ATRA plays a prominent role in the in- duction of proliferating primitive cells to differentiate into mature cells that develop neurites [13]. However, the eﬃcacy of the in vivo ATRA treatment is limited in time for advanced NBs that often de- velop resistance to RA such that the outcome for high-risk patients unfortunately remains fatal [14].

Besides, it is widely acknowledged that the cell biophysical mi- croenvironment, deﬁned by its roughness, topography, and rigid- ity for instance, plays a pivotal role in the regulation of the tumor growth and the development of metastasis [15–17] . The stiffness of the extracellular matrix (ECM) has been focused for the recent years since it is well admitted that cells can sense mechanical cues of their microenvironment by exerting traction forces mainly via their integrins that convey the mechano-signal outside-in through the actin cytoskeleton, for a ﬁnal transduction into biochemical sig- nals that direct the cell fate [18]. More speciﬁcally, the mechanical properties of the ECM have been evidenced to inﬂuence the ad- hesion, proliferation and differentiation of NB cells [19]. Undiffer- entiated NB cells are rather rounded but they develop neurite ex- tensions as they differentiate [20] and the promotion of neurites was demonstrated to be related to an increase of their substrate stiffness [21].
For all these reasons, there is a growing need to develop cell culture systems able to recapitulate as many ECM parameters as possible, in a biomimicry approach. In vitro ECM-like matrices containing glycosaminoglycans (GAGs) are of particular interest as GAGs can act as bioactive reservoir-like molecules, imitating their in vivo biomolecules storage role [22], and modulate the NB dif- ferentiation [23]. Particularly, chondroitin sulfates (CSs), sulfated GAGs that contain repeating disaccharides units of D-glucuronic acid (GlcA) and N-acetylgalactosamine (GalNAc) with different sul- fation patterns, can exert positive or negative regulatory effects on neuritogenesis, depending on their structure. They have been re- ported to be closely associated with the unfavorable outcome of NB. For example, Neurocan (NCAN), a chondroitin sulfate proteo- glycan, has been reported to induce an undifferentiated phenotype, and to promote the malignancy of NB cells [24]. Whereas highly sulfated CSD [25] and CSE [26] subtypes, exhibit substantial neurite outgrowth, as well as for fully sulfated mimetics composed of cel- lulose sulfate [27], CSA exerts a poor stimulation or even inhibits neurite extension [28].

In that context, we developed a tunable biomimetic ECM-like substrate, containing CSA and poly-L-lysine (PLL). At physiological pH, these two biopolymers behave as polyelectrolytes such that they can be assembled to form “Layer-by-Layer” (LbL) ﬁlms (also called polyelectrolyte multilayers, PEM), whose stiffness can be ad- justed by crosslinking with genipin (GnP), a biocompatible natu-
ral crosslinker [29,30]. GnP is becoming a reference as natural- crosslinking agent used for stiffening both 2D and 3D scaffolds. For example, Silva et al. used GnP to crosslink layer-by-layer nanoﬁlms composed of chitosan and alginate that were tuned to control their adhesive properties toward endothelial cells [31]. Another example in the case of natural-based 3D scaffolds is the study of the ad- hesion and proliferation of neuroblastoma that Chiono et al. devel- oped with blends of gelatin and chitosan, crosslinked with gradual concentrations of GnP [32].
We investigated the potential beneﬁts of these mechanically controlled LbL ﬁlms in combination with ATRA treatments towards NB cells differentiation. After characterizing the topography and the stiffness of our LbL ﬁlms, we examined the behavior of LAN- 1 cells in terms of survival, adhesion, proliferation, migration and ultimately, differentiation. Remarkably, we demonstrated a promis- ing synergistic combination of stiffened LbL ﬁlms with ATRA treat- ment, since LAN-1 cells were able to develop neurites extension, favorable to the maturation of NB along the neuronal lineage.
To our best knowledge, our results validate for the ﬁrst time a biomimetic ECM-like matrix culture system of controlled stiff- ness able to ﬁne-tune the differentiation/maturation of LAN-1 cells
when combined to ATRA treatment, with a striking discrimina- tion in actin/β3-tubulin synthesis during the neurites formation and extension. We clearly demonstrated that although ATRA is a
prerequisite for β3-tubulin synthesis, LAN-1 cells needed to grow onto appropriate mechanical microenvironment to synergistically coordinate β3-tubulin with actin and consequently, to initiate neu- rites outgrowth.

2. Materials and methods

2.1. Cell culture

MYCN-ampliﬁed human neuroblastoma LAN-1 cells [33] were kindly provided by Dr J. Benard, (CNRS UMR 8126, Institut Gus- tave Roussy, Villejuif, France) and were beforehand characterized by short tandem repeat analysis (STR) using the Promega pow- erplex 21 PCR kit (Euroﬁns). They were cultured in low glucose (1 g L−1) RPMI, supplemented with 10% FBS, 2 mM L-glutamine, 1% HEPES, 5000 UI L−1 penicillin and 50 mg L−1 streptomycin and maintained at 37 °C in a 5% CO2 humidiﬁed atmosphere. In some cases, cells were treated with 5 μM all-trans retinoic acid (ATRA). ATRA was prepared in absolute ethanol at 1 mM and stored at
−80 °C until used.

2.2. CSA/PLL assembly and post-crosslinking with genipin (GnP)

Solutions of chondroitin sulfate A (CSA, 20–30 kDa, Sigma) and poly-L-lysine (PLL, 40–60 kDa, Sigma) were prepared at 1 mg mL−1, and solutions of poly(ethyleneimine) (PEI, 60 kDa, Aldrich) were prepared at 5 mg mL−1 in Tris/NaCl buffer (10 mM Tris, 150 mM NaCl, physiological pH 7.4).
Based on the Layer-by-Layer (LbL) method developed by Decher [34], we built-up (CSA/PLL)6 LbL ﬁlms (where 6 corresponds to the number of (CSA/PLL)pairs of layers), either directly on top of the plastic wells bottoms, or atop glass slides, as previously de- scribed [29]. Brieﬂy, after an initial PEI deposition for 30 min, CSA and PLL were alternatively adsorbed onto the substrates by means of 10-min immersions, each followed by three rinses with Tris/NaCl buffer, until we obtained native PEI/(CSA/PLL)6 or PEI/(CSA/PLL)6/CSA LbL architectures, denoted in the following by PLL or CSA, respectively.
Some LbL ﬁlms were kept in their native state while some oth- ers were post-treated for 16 h at room temperature in the dark with various solutions of genipin (GnP, Wako Chemicals): 0%, 0.10%, 0.25%, 0.50%, 0.75%, and 1% (w/v) in Dulbecco’s PBS, denoted as

B. Labat, N. Buchbinder, S. Morin-Grognet et al. Acta Biomaterialia xxx (xxxx) xxx
GnP0, GnP10, GnP25, GnP50, GnP75, and GnP100, respectively. Fi- nally, all LbL ﬁlms were rinsed in PBS and UV-sterilized for 30 min before cell seeding.

2.3. Roughness and stiffness by AFM

For AFM analysis, we built-up native and crosslinked LbL ﬁlms on top of glass slides. Roughness and stiffness (global elastic mod- ulus) measurements were performed with a Pico SPM AFM setup from Molecular Imaging (Scientec) in wet conditions (Tris/NaCl buffer for the native ﬁlms; PBS buffer for the GnP-treated ﬁlms) to maintain the hydrated structure of the ﬁlms. For imaging and quantifying the LbL roughness (average roughness, Ra), we used pyramidal Si3N4 tips mounted on gold-coated cantilevers with a nominal spring constant of 0.3 N m−1 at 1 Hz scan rate and op- erated in contact mode. For the stiffness measurements, we used conical tips mounted on non-covered cantilevers with a nominal spring constant of 3.5 N m−1. This allowed to probe the global elastic modulus of the system as described in our previous work [30]. Elastic moduli were derived from the force-displacement curves by using the Sneddon model.

2.4. Cell density and viability

First, to determine the optimal cell density for the further ex- periments, cells were seeded at different densities (25 × 103, 50 × 103, 100 × 103 and 150 × 103 cells cm−2) atop native LbL ﬁlms that were directly assembled in 24-well plates. After 2 days, the culture medium was gently removed and replaced with a 0.4% trypan blue solution for 2 min. Then, 5 random ﬁelds were pho- tographed using a Sony camera mounted on an Axiovert 135 light microscope (Zeiss, Le Pecq, France). Total and dead (blue) cells were counted using ImageJ software (NIH, Bethesda, MD, USA).
Then, to test the effect of GnP treatment of the ﬁlms upon cell survival, we seeded the cells at the lowest cell density (25 000 cell cm−2) atop native or crosslinked ﬁlms with increasing concentra- tions of GnP and cultured them for 4 days. We stained cells with trypan blue as previously described.

2.5. Cell adhesion assay

LbL ﬁlms were directly assembled into 24-well plates and cells were seeded at a low cell density of 25 × 103 cells cm−2 atop native and crosslinked ﬁlms, to assess the strength of cell/ﬁlm interactions rather than cell/cell ones. After 5 h, the culture medium was gently removed and cells were put in contact with trypsin/EDTA solution for 7 min until they detached from the LbL ﬁlms. Cell concentrations were then determined using a Coulter Counter (Beckman Coulter, Brea, California).

2.6. Cell cycle assay

Cells were cultured for 2 and 5 days (intermediate cell den- sity of 50 × 103 cells cm−2) atop native and crosslinked ﬁlms that were built up into 24-well plates, with or without ATRA (5 μM) in the culture medium. At each time point, the culture medium was removed and cells were ﬁxed in cold absolute ethanol (VWR, Fontenay-sous-Bois, France), washed in PBS containing 0.5% FBS, and then incubated with a 50 μg mL−1 propidium iodide solution (Sigma–Aldrich, St Quentin Fallavier, France) containing RNAse A (100 μg mL−1). After 20 min into dark, cells were analyzed with a FACScalibur ﬂow cytometry system (BDBioscience).

2.7. Cell proliferation assay

Native or crosslinked LbL ﬁlms were built-up into 24-well plates and cells were seeded at a density of 50 × 103 cells cm−2 to
provide enough space for the cells to adapt and proliferate on the surface of the different substrates. After 5 days, the cul- ture medium was removed and cells were enzymatically detached (0,01X trypsin/EDTA solution) and counted using a Coulter Counter (Beckman Coulter, Brea, California).

2.8. Immunoﬂuorescence

First, native or crosslinked LbL ﬁlms were built up in 8-well Millicell EZ slides (Millipore, Ireland). Then, cells were seeded at 50 × 103 cell cm−2 for 24 h before ATRA (5 μM) treatment. This cell density would provide appropriate place for cell to elongate. Four days later, cells were washed with PBS, ﬁxed with 4% p- formaldehyde, permeabilized with 0.1% Triton X-100 and incubated with blocking solution containing 1% BSA in 0.1% Triton X-100. Cells were then incubated for 1 h with anti-β3-tubulin primary mouse antibody diluted at 1:100. After 3 PBS washes, FITC-coupled anti-mouse secondary antibody was used at 1:100 and incubated for 1 h. Actin cytoskeleton was labeled using rhodamin-coupled phalloidin at 1:450, and nuclei were stained using DAPI at 1 μg mL−1. Pictures were obtained using a ﬂuorescence microscope (Ax- iovert 200 M, Zeiss).

2.9. Measurement of neurite length, percentage of cell generating neurites and number of neurites per cell

Cells were seeded at the same density as for immunoﬂuores- cence (50 × 103 cells cm−2) into 12-well plates containing native or crosslinked LbL ﬁlms and let to adhere for 24 h. Then, cells were incubated with ATRA (5 μM) for 96 h. For each LbL ﬁlm condition, at least 8 randomly ﬁelds were photographed using a Sony cam- era mounted on an Axiovert 135 light microscope (Zeiss, Le Pecq, France). The total length of neurites was measured using Saisam software (Microvision Instruments, Evry, France), and the average neurite length was determined by dividing the total length of neu- rites by the number of counted cells. Only neurites that exhibited processes >10 μm in length were considered and the length was measured from the tip of a neurite to the cell body. We also de- termined the percentage of cell that were able to generate neu- rites and the number of neurites per cell from the same 8 random ﬁelds.

2.10. Time-lapse videomicroscopy

Cells were seeded at 25 × 103 cell cm−2 atop native or crosslinked LbL ﬁlms that were built-up in 24-wells plates. That low cell density offers enough surface for cells to move on. Plates were placed inside the microscope integrated incubator (37 °C and 5% CO2) and images were recorded with a temporal resolution of t = 10 min for 24 h such that the videos contained 145 im- ages (1032 × 1300 pixels). Manual cell tracking was performed for cells that moved within a ROI (region of interest) during the entire recording time by using ImageJ equipped with the MTrackJ plugin [35].

2.11. Western blot

Cells were seeded (50 × 103 cells cm−2) atop native or crosslinked LbL ﬁlms directly built up into 6-well plates. Twenty- four hours after seeding, cells were treated with ATRA (5 μM) for 4 days. Then, cells were washed in PBS, harvested, and solu- bilized in RIPA lysis buffer containing freshly added protease in- hibitors (1% protease inhibitor cocktail and 1 mM phenyl methyl sulﬁde ﬂuoride) and 0.1% phosphatase inhibitor (Pierce Biotech- nology, Perbio Science) and kept in ice for 20 min. Lysates wereclariﬁed by centrifugation at 10 000 g for 15 min at 4 °C. West- ern blots were performed using the Xcell II system (Invitrogen, Cergy Pontoise, France) according to the manufacturer’s protocol. Proteins were assayed by using Bradford assay and 30 μg of pro- teins were diluted in loading buffer, heated for 5 min at 70 °C, size-separated in a 8% polyacrylamide gel, and transferred to a polyvinylidene diﬂuoride membranes (Amersham, GE Healthcare). Next, membranes were blocked with 5% non-fat milk in 0.1 M Tris-base, 0.2 M NaCl and 0.1% Tween solution, washed, and in- cubated with anti-β3 tubulin antibodies according to the manu- facturer’s protocol (Cell Signaling, Denver, MA, USA). Horseradish peroxidase-coupled anti mouse antibodies, and anti-GAPDH were used at 1:5000 dilution. Antibody binding was revealed using the enhanced chemiluminescence system (Amersham, GE Healthcare). Membranes were stripped in a 62.5 mM Tris buffer containing 2% SDS and 100 mM β-mercaptoethanol (pH 6.7).

2.12. Statistical analysis

All the experiments were performed at least in triplicates and repeated at least twice. Quantitative data were reported either as mean ± SEM when 3 biological replicates were performed (n = 3) or by means of scatter plots when n = 2. Data were also statistically analyzed by using the nonparametric Mann-Whitney- Wilcoxon test. Statistical differences were assigned with P-values deﬁned as ∗ < 0.05.

3. Results

3.1. Genipin stiffens (CSA/PLL)6 LbL ﬁlms

We ﬁrst inspected the surfaces of the LbL ﬁlms by using AFM imaging (Fig. 1A). For all the conditions, we obtained similar to- pographical features, exhibiting islets-like structures with a rough- ness Ra ∼ 30 nm. We also used AFM nano-indentation for the measurement of the surface mechanical properties of the ﬁlms that are probed by the cells. By increasing the concentration of GnP, the
LbL ﬁlms gradually stiffened and exhibited a global Young modulus ranging from ∼30 kPa for the softest ﬁlm to 160 kPa for the stiffest one (Fig. 1B).

3.2. Stiffened ﬁlms promote survival and enhance adherence of LAN-1 cells

Atop native CSA-ending ﬁlms, LAN-1 cells had troubles sustain- ing their survival (Fig. 2A-bottom). Increasing cell density from 25 × 103 to 150 × 103 cells cm−2 slightly improved the cell sur- vival but only 40% of living cells were obtained for the highest cell density (150.000 cells cm-2) after 2 days of culture. Conversely, on top of native PLL-ending ﬁlms, cells exhibited a better survival ranging from 40% to 75% of living cells, from the lowest density to the density of 100 × 103 cells cm−2, respectively.
We next tested the impact of GnP treatment of the ﬁlms upon
the viability of cells for a 4 days culture period with the lowest cell seeding density of 25 × 103 cells cm−2. Whereas some cells were able to survive 2 days atop native ﬁlms (Fig. 2A), they could not subsist any longer (Fig. 2B). In contrast, GnP-crosslinked ﬁlms strikingly improved cell survival regardless of the ending layer of the ﬁlms and whatever the concentration of GnP.
Finally, we assessed cell adherence atop native or crosslinked ﬁlms 5 h after cell seeding at the same density (25 × 103 cells cm−2), followed by a trypsin treatment (Fig. 2C). Whereas on top of native ﬁlms, numerous cells were quite easily detached (54% for
CSA-ending ﬁlms and 21% for PLL-ending ﬁlms), crosslinked LbL ﬁlms strengthened the cell adherence, irrespective of the ending layer (only between 4% and 6% cells were detached for PLL- and CSA-ending ﬁlms, respectively).

3.3. Stiffened ﬁlms reduce apoptosis and stimulate proliferation of LAN-1 cells

Flow cytometry analysis was performed at D2 and D5 after cul- turing cells atop native or crosslinked ﬁlms to assess their cell cy- cle (Fig. 3A). At D2, atop native ﬁlms, we noted a signiﬁcant per- Genipin crosslinking stiffens (CSA/PLL)6 LbL ﬁlms. (A) Representative 3D AFM images (20 μm x 20 μm) exhibiting surface topographies of native (CSA/PLL)6 LbL ﬁlms (GnP0) and crosslinked ﬁlms with increasing concentrations of genipin (GnP10, GnP25, GnP50, GnP75, GnP100). (B) Corresponding Global Young’s modulus (kPa) obtained from AFM nano-indentation. Each value was expressed as mean ± SEM (n = 3), ∗ corresponding to p<0.05.

Stiffened ﬁlms promote survival and enhance adherence of LAN-1 cells. (A) Inﬂuence of cell density on cell survival 2 days after plating on top of native (CSA/PLL)6 (top) or (CSA/PLL)6 /CSA (bottom) LbL ﬁlms. Trypan Blue staining and percentage of live (white) vs. dead (blue) cells. Scale bar: 40 μm. Each value correspond to 15 random ﬁelds and was expressed as mean ± SEM (n = 2), (B) Percentage of live cells cultured on top of native or crosslinked (CSA/PLL)6 ± CSA LbL ﬁlms with GnP50 and GnP100, 4 days after plating cells at a density of 25 × 103 cells cm−2 . Data are presented with scatterplots and mean values; (n = 2), (C) Cell adherence on top of same native or crosslinked ﬁlms as in (B), 5 h after plating. Each value was expressed as mean ± SD (n = 3), ∗ corresponding to p < 0.05for CSA-ending ﬁlms. Most cells were in G0/G1 phase (∼80%). In contrast, at both D2 and D5, cells cultured atop crosslinked ﬁlms exhibited a 2-fold increase of S + G2/M phase compared to those cultured atop native ﬁlms.
This was conﬁrmed by the cell proliferation assessment also performed at D5 (Fig. 3B). Contrary to native ﬁlms, crosslinked ﬁlms were favorable to cell proliferation, even for the weakest con- centration of genipin (GnP10), with similar rates to the other GnP concentrations and whatever the ending layer of the ﬁlms.

3.4. Stiffened ﬁlms + ATRA strongly reduce apoptosis and proliferation but rather increases G0/G1 phase of LAN-1 cells

Stiffened ﬁlms reduce apoptosis and stimulate proliferation of LAN-1 cells.
(A) Representative cell-cycle distribution of cells cultured on top of native or GnP50 crosslinked LbL ﬁlms, for 2 and 5 days - SubG0, G0/G1 and S + G2/M phases.
(B) Corresponding cell proliferation at D5 atop LbL ﬁlms, in their native state or crosslinked with GnP10, GnP50, and GnP100. Data are presented with scatterplots and mean values; (n = 2), ∗ corresponding to p < 0.05centage of hypodiploid LAN-1 cells in SubG0 phase, with a higher rate for PLL-ending ﬁlms (50%) than for CSA-ending ones (34%). Conversely, atop crosslinked ﬁlms, a very few cells were in SubG0 phase (4% and 1%, for PLL GnP50 and CSA-GnP50, respectively) and nearly half of the cells were in S + G2/M phase. From D2 to D5 atop native ﬁlms, the LAN-1 cells population in SubG0 dramatically decreased from 50% to 2% for PLL-ending ﬁlms and from 34% to 5%

We next supplemented the culture medium with exogenous ATRA (5 μM) and we performed a new ﬂow cytometry analysis of LAN-1 cells after 2 days of treatment (Fig. 4).On top of PLL-ending LbL ﬁlms, the ATRA treatment induced a dramatic decrease of SubG0 cells, for every stiffness condition with a much more marked effect in the case of native ﬁlms (from 25% down to 3%) (Fig. 4A). Conversely, on top of CSA-ending ﬁlms, the treatment of ATRA did not change the proportion of cells in SubG0, whatever the stiffness of the ﬁlms.
The total cell cycle distribution represented in Fig. 4B shows that in the absence of ATRA, the stiffening of LbL ﬁlms provoked an increase of S phase. In contrast, ATRA exposure induced a re- duction of S phase and an increase of the G0/G1 one, for both PLL- and CSA-ending LbL ﬁlms.

3.5. Stiffened ﬁlms + ATRA synergistically increases LAN-1 motility onto CSA-ending ﬁlms

Migration is a key property of metastatic cancer cells to in- vade tissues. LAN-1 cells motility and migration speed were as- Stiffened ﬁlms + ATRA strongly reduce apoptosis and proliferation but rather increases G0/G1 phase of LAN-1 cells. Analysis of cell cycle of LAN-1 cells, treated or not with 5 μM ATRA, cultured atop native or crosslinked ﬁlms (GnP25, GnP50, GnP75, and GnP100) for 2 days. (A)% DNA fragmentation represented by the SubG0 sub- population of cells. Data are presented with scatterplots and mean values; (n = 2), ∗ corresponding to p < 0.05, (B) Representative total percentage of DNA content (SubG0, G0/G1, S, and G2/M phases).

Stiffened ﬁlms + ATRA synergistically increases LAN-1 motility onto CSA-ending ﬁlms. Representative example of individual cell tracking for 24 h atop GnP50 crosslinked (CSA/PLL)6 ± CSA ﬁlms treated or not with ATRA (5 μM). (A) Trajectories of cells represented by colored lines (n cells = 7 to 9). (B) Corresponding average migration speed of cells cultured on top of the same GnP50 LbL ﬁlms and atop native LbL ﬁlms (GnP0). Data are presented with scatterplots and mean values; (n = 2), ∗ corresponding to p < 0.05sessed on top of native and crosslinked LbL ﬁlms, with or with- out ATRA treatment. On top of PLL-ending ﬁlms, migrating cells adopted a ﬂattened proﬁle (Fig. 5A-left) and had a similar migra- tion speed on top of both native and crosslinked LbL ﬁlms (Fig. 5B). Conversely, CSA-ending ﬁlms forced all migrating cells to remain rounded, developing very small membrane extensions necessary to migrate (Fig. 5A-right) and cells speeded-up their migration when cultured onto stiffer LbL ﬁlms (Fig. 5B), particularly when treated with ATRA. Increasing the concentration of genipin up to GnP100 did not induce signiﬁcant difference in the cell motility. The no- ticeable effect was obtained from GnP50.

3.6. Stiffened ﬁlms + ATRA markedly promote neuritogenesis of LAN-1 cells

To test our hypothesis that LAN-1 cells cultured on top of crosslinked ﬁlms and treated with 5 μM ATRA were likely dif- ferentiating, we investigated their propensity to adopt a mature neuron-like phenotype. For that, we examined their morphology and particularly, their actin cytoskeleton associated to microtubule synthesis, hallmark of neurite extension. We ﬁrst used ﬂuores- cent labeling of actin cytoskeleton, β3-tubulin and nuclei. As expected, on top of native ﬁlms (GnP0), all cells remained rounded, unable to elongate, whatever the ending-layer (Fig. 6A,B), although some of them (between 20 and 60% - Fig. 6D) generated a few number (between 2 and 3 - Fig. 6E) of small neurites exten- sions. However, ATRA treatment obviously induced the synthe- sis of both actin and β3-tubulin, but in a meshwork-like struc- ture rather than bundles, accordingly to β3-tubulin bands ob- served in the western blots (Fig. 6C). Conversely, atop crosslinked ﬁlms, almost all cells were able to spread and develop elon- gating neurites (Fig. 6A,D), in a greater number (Fig. 6E) for both GnP50 and GnP100, still more extended after ATRA treat- ment (Fig. 6B) but without signiﬁcant difference between GnP50 and GnP100, and expressing β3-tubulin (Fig. 6C). Overall, the longest and the more numerous neurites were obtained for cells cultured onto crosslinked PLL-ending ﬁlms and treated with ATRA.

4. Discussion

It is generally admitted that ECM stiffening is a common hall- mark of cancerous tissues [36,37]. Based on our previous work on GnP-crosslinked hyaluronan hydrogels for the study of cancerous Stiffened ﬁlms + ATRA markedly promote neuritogenesis of LAN-1 cells. LAN-1 cells differentiation assessed by neurites elongation, and actin and β3-tubulin expression, when cultured on top of native or crosslinked ﬁlms, and treated or not with ATRA (5 μM). (A) Representative immunoﬂuorescence micrographs of cells observed 4 days after ATRA treatment. Nuclei were labeled with DAPI (blue), actin cytoskeleton with phalloidin-TRITC (red), and β3-tubulin with anti-β3-tubulin primary mouse antibody, itself labeled with FITC-coupled anti-mouse secondary antibody (green). Scale bar corresponds to 20 μm. (B) Average neurite length measured from light microscopy images of 8 random ﬁelds. Each value was expressed as mean ± SD (n = 3), (C) Representative example of western blot analysis showing β3-tubulin and GAPDH (as loading control) expressions, 4 days after ATRA treatment. (D) Percentage of cell generating neurites and (E) Number of neurites per cell, both obtained from images of 8 random ﬁelds. Each value was expressed as mean ± SD (n = 3). ∗ corresponding to p < 0.05 with the corresponding native ﬁlm condition.glioblastoma cells [38], we assessed the behavior of neuroblastoma cells (LAN-1) on top of biomimetic mechanically tunable LbL ﬁlms composed of a glycosaminoglycan (CSA) and a polypeptide (PLL) that were gradually stiffened by using genipin.

The inspection of all LbL ﬁlms surfaces revealed comparable to- pographies, exhibiting homogeneously distributed islets-like struc- tures of similar roughness, whatever the concentration of GnP. This is in line with a previous work where we described that GnP treatments of the same biomimetic CSA-containing LbL ﬁlms gen- erated semi-interpenetrated polymer network (semi-IPN) architec- tures, composed of a GnP-crosslinked PLL network in which CSA molecules remained free to diffuse, displaying islets features [30]. We demonstrated that this topography was not a determinant pa- rameter that could alter pre-osteoblast cells behavior. Likewise, considering the comparable roughness values of the present LbL ﬁlms, we assume that their surface features would not modulate the NB cells behavior.
The stiffness of the cell microenvironment remains a pivotal biophysical cue controlling cell processes. In vivo tissues exhibit a large spectrum of
rigidities, ranging from 0.1 kPa to 10 GPa, from brain to bone, respectively [18,39]. In the case of in vitro studies of neural tissue, it has been demonstrated that the very soft sub- strates (0.1–10 kPa) promoted neurogenesis and gliogenesis [40], while the stiffest ones (∼ 500 kPa) were favorable to the activation of synaptic connectivity and transmission [41,42]. Here, by varying the concentration of GnP, we were able to modulate the surface mechanical properties of the LbL ﬁlms, in a physiologically rele- vant range (∼30 kPa to 160 kPa). Noteworthy, the extent of the data dispersion reported on box plots are likely to be related to the use of sharp tips (conically-shaped) for AFM nano-indentation experiments that exerted a very local contact onto the surface of the LbL ﬁlms either on top of or between the islet-like structures (as shown on Fig. 1A). Subsequently, the AFM tips probed more or less dense areas.

Cell survival atop native CSA-ending LbL ﬁlms was very poor, even for the highest cell density. Because cells are negatively charged, the polyanionic nature of CSA could explain such a cell- repellence. Over a prolonged cell culture period, we observed the formation of cell aggregates (data not shown), likely because cells developed cell-cell interactions rather than cell-substrate interac- tions for their survival. Of note, LAN-1 cells belong to N-type cells, identiﬁed as small and loosely adherent that tend to ag- gregate in spheroid-like structures depending on their microen- vironment [43,44]. They required suitable interactions with their substrates to circumvent anoikis. Moreover, MYCN-ampliﬁed LAN-
1 cells were reported to downregulate CD44 adhesion receptors [45,46] involved in GAG/cells interactions. In contrast, native PLL- ending ﬁlms were much more favorable to cell survival. As a cationic polypeptide, PLL is often used for the coating of cell cul- ture plates to favor cell adhesion through electrostatic interactions. We thus believe that the positive charges of PLL supported LAN-1 cells adhesion and subsequently, facilitated their survival.

Markedly, the treatment of the LbL ﬁlms with GnP improved the cell viability, whatever the ending layer, indicating that the stiffness of the ﬁlms was likely a key parameter that governed cell survival. This is in line with frequently noticed cancer cell sur- vival to be inﬂuenced by ECM stiffness and whose malignancy has been shown to be enhanced through integrin-dependent mechano- transduction [47].
We next conﬁrmed that stiffened ﬁlms dramatically improved cell adhesion, in accordance with another work wherein NB cells cultured onto collagen I-coated polyacrylamide substrates of vari- ous stiffnesses exhibited a better adhesion onto the stiffest ones
[21] and with another work using blends of gelatin and chitosan also crosslinked with GnP that demonstrated a better adhesion of neuroblastoma S5Y5 cells on the stiffest substrates [32]. Cell adhe- sion is mainly controlled by the anchoring of transmembrane inte- grins to ECM molecules (collagen, laminin, ﬁbronectin), while therole of CS is not elucidated yet. It is generally admitted that CS- containing substrates display anti-adhesive properties for several different types of cells, including neuron cells [48,49].

The cell cycle analysis clearly conﬁrmed that LAN-1 cells cul- tured atop the native ﬁlms during the ﬁrst 2 days were mostly apoptotic whereas the GnP treatment of the ﬁlms obviously im- proved the cell survival. Over time and atop native LbL ﬁlms, only a few cells remained in Sub/G0. This could probably be related to the renewal of culture medium between D2 and D5, such that apop- totic, low-adhering and dead cells were eliminated. Consequently, only strongly adhering cells remained anchored and likely entered into a quiescent or maybe in a senescent state. This is in accor- dance with a recent work reporting that NB cells cultured within 3D printed hydrogels presented a greater number of anti-apoptosis markers than apoptosis ones as the hydrogel stiffness increased[50]. Conversely, stiffened LbL ﬁlms clearly promoted mitotic activ- ity as indicated by the increase of cells being in S + G2/M phase and by the results of cell proliferation. This is in line with many other studies that reported cancer cell growth increased when ECM stiffened [51–54]. However, our results differ from another study that demonstrated an inverse correlation between ECM stiffness and cell proliferation for SK-N-DZ neuroblastoma cells cultured on top of polyacrylamide gels with rigidity varying from 0.1 kPa to 10 MPa [21]. Although the range of stiffness overlapped therange tested here, the coating of the gels was exclusively protein in nature (type I collagen), whereas we developed a more tumor- like ECM, including a GAG component. Indeed, CS has been re- ported to be up-regulated in many tumor microenvironments [55], promoting proliferation and a recent study found that Neurocan (NCAN), a secreted CS proteoglycan, stimulated malignant pheno- type in NB cancers, by inducing the formation of tumor spheres [24].

Because ATRA is a pharmacological molecule commonly used in the treatment of NB cancers, we implemented it in our study to test its potentiality upon LAN-1 cells cultures when combined to our mechanically-controlled LbL ﬁlms. As previously discussed, the stiffening of the ﬁlms induced both an anti-apoptotic and pro- proliferative effects upon LAN-1 cells, whatever the ending-layer. But, when the cells were treated with ATRA, we observed an anti- proliferative effect, in line with many others works [57,58]. ATRA is known to stop NB cells proliferation in favor of their differen- tiation [56,59,60]. Strikingly, the ending layer had a crucial dis- criminating effect. Indeed, while on top of CSA-ending LbL ﬁlms, ATRA had no effect on the SubG0 phase, atop PLL-ending ﬁlms, the proportion of apoptotic cells was markedly reduced, particu- larly for native PLL-ending ﬁlms. The biological activity of ATRA is potentiated after binding to retinoic acid receptors (RARs). ATRA contains a carboxylate group at its polar cluster that electrostati- cally interacts with RARs through the positively charged arginine residues [61]. In the case of PLL-ending ﬁlms, positive charges are exposed at the upper surface of the ﬁlms. Thus, by adding ATRA in the culture medium, there is a possibility that it partially bound to PLL through electrostatic interactions, possibly being entrapped within the LbL ﬁlms such that it could increase its biodisponibil- ity to cells. Besides, GnP has been reported to be a good candidate as anti-cancer drug by inhibiting the UCPs family (Uncoupling pro- teins), mitochondrial proteins involved in ATP production [62,63]. UCP2 was shown to be overexpressed in many cancer cells lines and is known to promote tumorigenesis both in vitro and in vivo [64]. Thus, one cannot rule out a possible additional biological ef- fect of GnP upon cell survival in our study.

Cell migration is a crucial process in cancer invasion. N- type LAN-1 cells are known to disseminate and produce micro- metastases in the alveolar interstitium [65]. We evidenced that the biochemical nature of the surface of the LbL ﬁlms also affected cell motility in terms of cell shape and speed. Irrespective to stiffness,cells cultured on top of PLL-ending ﬁlms adopted a ﬁbroblast-like shape while atop CSA-ending ﬁlms they remained rather rounded, consistently with the poor cell adhesion observed in Fig 2C. It is generally admitted that soft matrices are not mechano-compatible to traction forces exerted by many cells to spread or migrate [66]. Interestingly, the LbL ﬁlms stiffening induced the modulation of cell migration speed only atop CSA-ending LbL ﬁlms. Cells doubled their motility, exhibiting a more migratory phenotype. Remarkably, ATRA exerted a converse effect upon cell migration depending on the ending layer of the ﬁlms. While it slightly slowed down the motility of cells when cultured on top of PLL-ending ﬁlms, it accel- erated their migration speed onto CSA-ending ones. This indicates that the biochemical nature of the outmost surface of a matrix can subtlety modulate the cell motility.
Finally, the degree of differentiation of NB cells could be mea- sured by the percentage of cell generating neurites, the num- ber of neurites per cell and the length of neurites extension, driven by ECM cues, particularly the biochemical composition, and stiffness. Undifferentiated NB cells are non-polarized and rather rounded cells. During neurite extension, microtubules and actin ﬁbers cooperate to spatially organize the cytoskeleton. More pre- cisely, microtubules align and structure in bundles, while actin ﬁl- aments rearrange to initiate growth cones. Actin acts as a guide to the stabilization of microtubule network [67]. Thus, without a strong adhesion to substrates, NB cells cannot exert traction forces through mechanosensitive integrins to generate neuronal growth cones [68] that are known to be highly sensitive to substrate stiff- ness [69] and necessary to initiate and extend neurites.

On top of the native LbL ﬁlms, even though some cells could produce some very small extensions sensing their substrates, cells were unable to generate actin stress ﬁbers, and subsequently, un- able to develop neurites, likely because of a lack of focal adhe- sions. Strikingly, the ATRA exposure effectively induced the syn-thesis of both actin and β3-tubulin but this was not suﬃcient
to initiate neurite sprouting and even less their elongation. This result demonstrates that ATRA is certainly a prerequisite for β3- tubulin production but, as such, it is not suﬃcient to coordinatewith actin and initiate neurites outgrowth, in line with the study of Lam et al. [21] showing that single RA exposure was not self- suﬃcient to achieve optimal differentiation. In contrast, the com- bination of ATRA with a mechanically suitable ﬁlm, favored the de- velopment of ﬁlopodia prior to neurite extensions, consistent with a recent work demonstrating a synergistic effect of RA and loss modulus upon the neuronal differentiation of SHSY-5Y neuroblas- toma cells [70]. Of note, as for cell survival, one cannot exclude an intrinsic role of GnP upon neurite outgrowth as demonstrated by Yamazaki et al. with Neuro2 cells through an NO–cGMP–PKG signaling pathway [71]. Accordingly, we tried to test GnP in solu- tion (as control), but as it is largely admitted in the literature, GnP bonds every amino-containing component [72,73], which was the case for serum proteins contained in the culture medium, such that the latter turned to blue after GnP dissolved.

5. Conclusion

This study describes a promising cell culture platform based on versatile biomimetic LbL ﬁlms suitable for the study of neurob- lastoma cells. These ﬁlms provide an extracellular matrix-like mi- croenvironment that can be biochemically and mechanically tuned. By playing with the outmost layer (chondroitin sulfate A or poly- L-lysine) and the stiffness of the ﬁlms, we could modulate the cell survival, cell cycle, proliferation, motility, and the neuritogenesis of LAN-1 cells. Importantly, our work demonstrated that by combin- ing an appropriate substrate stiffness with a conventional retinoic acid treatment, undifferentiated NB cells were able to spread and extend neurites, similarly to mature neurons.

While the most promising direction of RA therapy is the pos- sible synergy of multidrug treatments, targeting different signal- ing pathways, these results highlight the crucial role of the bio- chemical composition and the mechanical cues of the neuroblas- toma microenvironment since they can dramatically modulate the effect of pharmacologic drugs by switching their activity from pro- to anti-, and conversely. In conclusion, our biomimetic platform of- fers a promising tool to help fundamental understanding in neu- roblastoma differentiation and may help the design of transla- tional biomaterials to support neuronal regeneration. We believe that this new ﬁndings would also be of particular interest in the screening of pharmacological drugs upon pathological microenvi- ronments such as NB and likely from other pathological tissues.

Declaration of Competing Interest

The authors declare that they have no known competing ﬁnan- cial interests or personal relationships that could have appeared to inﬂuence the work reported in this paper.

Acknowledgments

This work was partially supported by EPN (Evreux Portes de Normandie) and Région Normandie. The authors also want to thank Lucia Bertolini-Forno, Elisabeth Legrand and Catherine Bu- quet for their technical assistance.

References
[1] S.L. Cohn, A.D.J. Pearson, W.B. London, T. Monclair, P.F. Ambros, G.M. Brodeur,
A. Faldum, B. Hero, T. Iehara, D. Machin, V. Mosseri, T. Simon, A. Garaventa,
V. Castel, K.K. Matthay, INRG Task Force, The International Neuroblastoma Risk Group (INRG) classiﬁcation system: an INRG Task Force report, J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol. 27 (2009) 289–297, doi:10.1200/JCO.2008.16.6785.
[2] G.M. Marshall, D.R. Carter, B.B. Cheung, T. Liu, M.K. Mateos, J.G. Meyerowitz,
W.A. Weiss, The prenatal origins of cancer, Nat. Rev. Cancer. 14 (2014) 277– 289, doi:10.1038/nrc3679.
[3] V. Ciccarone, B.A. Spengler, M.B. Meyers, J.L. Biedler, R.A. Ross, Phenotypic di- versiﬁcation in human neuroblastoma cells: expression of distinct neural crest lineages, Cancer Res. 49 (1989) 219 http://cancerres.aacrjournals.org/content/ 49/1/219.abstract.
[4] G.M. Brodeur, Neuroblastoma: biological insights into a clinical enigma, Nat. Rev. Cancer. 3 (2003) 203–216, doi:10.1038/nrc1014.
[5] J.B. Jensen, M. Parmar, Strengths and limitations of the neurosphere culture system, Mol. Neurobiol. 34 (2006) 153–161 https://doi.org/10.1385/MN:34:3: 153.
[6] J.M. Maris, M.D. Hogarty, R. Bagatell, S.L. Cohn, Neuroblastoma, Lancet Lond. Engl. 369 (2007) 2106–2120, doi:10.1016/S0140-6736(07)60983- 0.
[7] J.A. Tomolonis, S. Agarwal, J.M. Shohet, Neuroblastoma pathogenesis: deregula- tion of embryonic neural crest development, Cell Tissue Res. 372 (2018) 245– 262, doi:10.1007/s00441-017-2747- 0.
[8] T.T. Amatruda, N. Sidell, J. Ranyard, H.P. Koeﬄer, Retinoic acid treatment of human neuroblastoma cells is associated with decreased N-myc expres- sion, Biochem. Biophys. Res. Commun. 126 (1985) 1189–1195, doi:10.1016/ 0006-291x(85)90311- 0.
[9] N. Bayeva, E. Coll, O. Piskareva, Differentiating neuroblastoma: a systematic re- view of the retinoic acid, its derivatives, and synergistic interactions, J. Pers. Med. 11 (2021) 211, doi:10.3390/jpm11030211.
[10] C.P. Reynolds, K.K. Matthay, J.G. Villablanca, B.J. Maurer, Retinoid therapy of high-risk neuroblastoma, Cancer Lett. 197 (2003) 185–192, doi:10.1016/ s0304-3835(03)00108- 3.
[11] K.K. Matthay, C.P. Reynolds, R.C. Seeger, H. Shimada, E.S. Adkins, D. Haas- Kogan, R.B. Gerbing, W.B. London, J.G. Villablanca, Long-Term Results for chil- dren with high-risk neuroblastoma treated on a randomized trial of myeloab- lative therapy followed by 13-cis-Retinoic acid: a Children’s Oncology Group Study, J. Clin. Oncol. 27 (2009) 1007–1013, doi:10.1200/JCO.2007.13.8925.
[12] G.J. Veal, M. Cole, J. Errington, A.D.J. Pearson, A.B.M. Foot, G. Whyman,
A.V. Boddy, Pharmacokinetics and metabolism of 13- cis -retinoic acid (isotretinoin) in children with high-risk neuroblastoma – a study of the United Kingdom Children’s Cancer Study Group, Br. J. Cancer. 96 (2007) 424–431, doi:10.1038/sj.bjc.6603554.
[13] S. Påhlman, A.-.I. Ruusala, L. Abrahamsson, M.E.K. Mattsson, T. Esscher, Retinoic acid-induced differentiation of cultured human neuroblastoma cells: a compar- ison with phorbolester-induced differentiation, Cell Differ. 14 (1984) 135–144, doi:10.1016/0045-6039(84)90038-1.
[14] F. Peinemann, E.C. van Dalen, H. Enk, F. Berthold, Retinoic acid postcon- solidation therapy for high-risk neuroblastoma patients treated with autolo- gous haematopoietic stem cell transplantation, Cochrane Database Syst. Rev. 8 (2017) CD010685, doi:10.1002/14651858.CD010685.pub3.

[15] S.N. Hanumantharao, C.A. Que, B.J. Vogl, S. Rao, Engineered three-dimensional scaffolds modulating fate of breast cancer cells using stiffness and morphology related cell adhesion, IEEE Open J. Eng. Med. Biol. 1 (2020) 41–48, doi:10.1109/ OJEMB.2020.2965084.
[16] C. Rianna, M. Radmacher, Inﬂuence of microenvironment topography and stiffness on the mechanics and motility of normal and cancer renal cells, Nanoscale 9 (2017) 11222–11230, doi:10.1039/C7NR02940C.
[17] K.K.B. Tan, C.S.Y. Giam, M.Y. Leow, C.W. Chan, E.K.F. Yim, Differential cell adhe- sion of breast cancer stem cells on biomaterial substrate with nanotopograph- ical cues, J. Funct. Biomater. 6 (2015) 241–258, doi:10.3390/jfb6020241.
[18] A.J. Engler, S. Sen, H.L. Sweeney, D.E. Discher, Matrix elasticity directs stem cell lineage speciﬁcation, Cell 126 (2006) 677–689, doi:10.1016/j.cell.2006.06. 044.
[19] N.D. Leipzig, M.S. Shoichet, The effect of substrate stiffness on adult neu- ral stem cell behavior, Biomaterials 30 (2009) 6867–6878, doi:10.1016/j. biomaterials.2009.09.002.
[20] Z. Zhang, A. Lei, L. Xu, L. Chen, Y. Chen, X. Zhang, Y. Gao, X. Yang, M. Zhang,
Y. Cao, Similarity in gene-regulatory networks suggests that cancer cells share characteristics of embryonic neural cells, J. Biol. Chem. 292 (2017) 12842– 12859, doi:10.1074/jbc.M117.785865.
[21] W.A. Lam, L. Cao, V. Umesh, A.J. Keung, S. Sen, S. Kumar, Extracellular matrix rigidity modulates neuroblastoma cell differentiation and N-myc expression, Mol. Cancer 9 (2010) 35, doi:10.1186/1476-4598- 9- 35.
[22] A.D. Theocharis, D. Manou, N.K. Karamanos, The extracellular matrix as a mul- titasking player in disease, FEBS J. 286 (2019) 2830–2869, doi:10.1111/febs. 14818.
[23] H. Matsushima, E. Bogenmann, Modulation of neuroblastoma cell differentia- tion by the extracellular matrix, Int. J. Cancer 51 (1992) 727–732, doi:10.1002/ ijc.2910510511.
[24] Z. Su, S. Kishida, S. Tsubota, K. Sakamoto, D. Cao, S. Kiyonari, M. Ohira,
T. Kamijo, A. Narita, Y. Xu, Y. Takahashi, K. Kadomatsu, Neurocan, an extra- cellular chondroitin sulfate proteoglycan, stimulates neuroblastoma cells to promote malignant phenotypes, Oncotarget 8 (2017) 106296–106310, doi:10. 18632/oncotarget.22435.
[25] M. Shida, T. Mikami, J.-.I. Tamura, H. Kitagawa, Chondroitin sulfate-D promotes neurite outgrowth by acting as an extracellular ligand for neuronal integrin
αVβ3, Biochim. Biophys. Acta Gen. Subj. 1863 (2019) 1319–1331, doi:10.1016/j.
bbagen.2019.06.004.
[26] A.M. Clement, K. Sugahara, A. Faissner, Chondroitin sulfate E promotes neurite outgrowth of rat embryonic day 18 hippocampal neurons, Neurosci. Lett. 269 (1999) 125–128, doi:10.1016/s0304-3940(99)00432-2.
[27] R. Menezes, S. Hashemi, R. Vincent, G. Collins, J. Meyer, M. Foston, T.L. Arinzeh, Investigation of glycosaminoglycan mimetic scaffolds for neurite growth, Acta Biomater. 90 (2019) 169–178, doi:10.1016/j.actbio.2019.03.024.
[28] H. Wang, Y. Katagiri, T.E. McCann, E. Unsworth, P. Goldsmith, Z.-.X. Yu,
F. Tan, L. Santiago, E.M. Mills, Y. Wang, A.J. Symes, H.M. Geller, Chondroitin- 4-sulfation negatively regulates axonal guidance and growth, J. Cell Sci. 121 (2008) 3083, doi:10.1242/jcs.032649.
[29] F. Gaudière, I. Masson, S. Morin-Grognet, O. Thoumire, J.-.P. Vannier, H. At- mani, G. Ladam, B. Labat, Mechano-chemical control of cell behaviour by elas- tomer templates coated with biomimetic Layer-by-Layer nanoﬁlms, Soft Matter 8 (2012) 8327–8337, doi:10.1039/C2SM25614B.
[30] F. Gaudière, S. Morin-Grognet, L. Bidault, P. Lembré, E. Pauthe, J. Vannier, H. At- mani, G. Ladam, B. Labat, Genipin-cross-linked layer-by-layer assemblies: bio- compatible microenvironments to direct bone cell fate, Biomacromolecules 15 (2014) 1602–1611, doi:10.1021/bm401866w.
[31] J.M. Silva, J.R. García, R.L. Reis, A.J. García, J.F. Mano, Tuning cell adhesive prop- erties via layer-by-layer assembly of chitosan and alginate, Acta Biomater. 51 (2017) 279–293, doi:10.1016/j.actbio.2017.01.058.
[32] V. Chiono, E. Pulieri, G. Vozzi, G. Ciardelli, A. Ahluwalia, P. Giusti, Genipin- crosslinked chitosan/gelatin blends for biomedical applications, J. Mater. Sci. Mater. Med. 19 (2008) 889–898, doi:10.1007/s10856-007-3212- 5.
[33] G.M. Brodeur, R.C. Seeger, M. Schwab, H.E. Varmus, J.M. Bishop, Ampliﬁcation of N-myc in untreated human neuroblastomas correlates with advanced dis- ease stage, Science 224 (1984) 1121–1124, doi:10.1126/science.6719137.
[34] G. Decher, Fuzzy nanoassemblies: toward layered polymeric multicomposites, Science 277 (1997) 1232–1237, doi:10.1126/science.277.5330.1232.
[35] E. Meijering, O. Dzyubachyk, I. Smal, Chapter nine - methods for cell and par- ticle tracking, in: P.M. conn (Ed.), Methods Enzymol, Academic Press, 2012, pp. 183–200, doi:10.1016/B978- 0- 12-391857- 4.00009-4.
[36] J.L. Leight, A.P. Drain, V.M. Weaver, Extracellular matrix remodeling and stiff- ening modulate tumor phenotype and treatment response, Annu. Rev. Cancer Biol. 1 (2017) 313–334, doi:10.1146/annurev- cancerbio- 050216-034431.
[37] R. Burgos-Panadero, F. Lucantoni, E. Gamero-Sandemetrio, L. de la Cruz- Merino, T. Álvaro, R. Noguera, The tumour microenvironment as an integrated framework to understand cancer biology, Cancer Lett. 461 (2019) 112–122, doi:10.1016/j.canlet.2019.07.010.
[38] S. Bonnesœur, S. Morin-Grognet, O. Thoumire, D.Le Cerf, O. Boyer, J.-.P. Van- nier, B. Labat, Hyaluronan-based hydrogels as versatile tumor-like models: tun- able ECM and stiffness with genipin-crosslinking, J. Biomed. Mater. Res. A. 108 (2020) 1256–1268, doi:10.1002/jbm.a.36899.
[39] M. Akhmanova, E. Osidak, S. Domogatsky, S. Rodin, A. Domogatskaya, Physical, spatial, and molecular aspects of extracellular matrix of in vivo niches and ar- tiﬁcial scaffolds relevant to stem cells research, Stem Cells Int. (2015) e167025 2015, doi:10.1155/2015/167025.
[40] K. Saha, A.J. Keung, E.F. Irwin, Y. Li, L. Little, D.V. Schaffer, K.E. Healy, Substrate Modulus Directs Neural Stem Cell Behavior, Biophys. J. 95 (2008) 4426–4438, doi:10.1529/biophysj.108.132217.
[41] Q.-.Y. Zhang, Y.-Y. Zhang, J. Xie, C.-X. Li, W.-Y. Chen, B.-L. Liu, X. Wu, S.-N. Li,
B. Huo, L.-H. Jiang, H.-C. Zhao, Stiff substrates enhance cultured neuronal net- work activity, Sci. Rep. 4 (2014) 6215, doi:10.1038/srep06215.
[42] S. Yao, X. Liu, X. Wang, A. Merolli, X. Chen, F. Cui, Directing neural stem cell fate with biomaterial parameters for injured brain regeneration, Prog. Nat. Sci. Mater. Int. 23 (2013) 103–112, doi:10.1016/j.pnsc.2013.02.009.
[43] V. Ciccarone, B.A. Spengler, M.B. Meyers, J.L. Biedler, R.A. Ross, Phenotypic di- versiﬁcation in human neuroblastoma cells: expression of distinct neural crest lineages, Cancer Res. 49 (1989) 219–225.
[44] S. Acosta, C. Lavarino, R. Paris, I. Garcia, C. de Torres, E. Rodríguez, H. Beleta,
J. Mora, Comprehensive characterization of neuroblastoma cell line subtypes reveals bilineage potential similar to neural crest stem cells, BMC Dev. Biol. 9 (2009) 12, doi:10.1186/1471-213X-9-12.
[45] N. Schwankhaus, C. Gathmann, D. Wicklein, K. Riecken, U. Schumacher,
U. Valentiner, Cell adhesion molecules in metastatic neuroblastoma models, Clin. Exp. Metastasis. 31 (2014) 483–496, doi:10.1007/s10585-014-9643- 8.
[46] N. Gross, K. Balmas Bourloud, C.B. Brognara, MYCN-related suppression of functional CD44 expression enhances tumorigenic properties of human neu- roblastoma cells, Exp. Cell Res. 260 (2000) 396–403, doi:10.1006/excr.2000. 5007.
[47] K.R. Levental, H. Yu, L. Kass, J.N. Lakins, M. Egeblad, J.T. Erler, S.F.T. Fong,
K. Csiszar, A. Giaccia, W. Weninger, M. Yamauchi, D.L. Gasser, V.M. Weaver, Matrix crosslinking forces tumor progression by enhancing integrin signaling, Cell 139 (2009) 891–906, doi:10.1016/j.cell.2009.10.027.
[48] D.R. Friedlander, P. Milev, L. Karthikeyan, R.K. Margolis, R.U. Margolis,
M. Grumet, The neuronal chondroitin sulfate proteoglycan neurocan binds to the neural cell adhesion molecules Ng-CAM/L1/NILE and N-CAM, and inhibits neuronal adhesion and neurite outgrowth, J. Cell Biol. 125 (1994) 669–680, doi:10.1083/jcb.125.3.669.
[49] J. Jin, S. Tilve, Z. Huang, L. Zhou, H.M. Geller, P. Yu, Effect of chondroitin sul- fate proteoglycans on neuronal cell adhesion, spreading and neurite growth in culture, Neural. Regen. Res. 13 (2018) 289–297, doi:10.4103/1673-5374.226398.
[50] E. Monferrer, S. Martín-Vañó, A. Carretero, A. García-Lizarribar, R. Burgos- Panadero, S. Navarro, J. Samitier, R. Noguera, A three-dimensional bioprinted model to evaluate the effect of stiffness on neuroblastoma cell cluster dynam- ics and behavior, Sci. Rep. 10 (2020) 6370, doi:10.1038/s41598- 020- 62986- w.
[51] M. Alonso-Nocelo, T.M. Raimondo, K.H. Vining, R. López-López, M. de la Fuente,
D.J. Mooney, Matrix stiffness and tumor-associated macrophages modulate ep- ithelial to mesenchymal transition of human adenocarcinoma cells, Biofabrica- tion 10 (2018) 035004, doi:10.1088/1758-5090/aaafbc.
[52] A. López-Carrasco, S. Martín-Vañó, R. Burgos-Panadero, E. Monferrer,
A.P. Berbegall, B. Fernández-Blanco, S. Navarro, R. Noguera, Impact of extracellular matrix stiffness on genomic heterogeneity in MYCN- ampliﬁed neuroblastoma cell line, J. Exp. Clin. Cancer Res. 39 (2020) 226, doi:10.1186/s13046- 020-01729- 1.
[53] I. Tadeo, A.P. Berbegall, S. Navarro, V. Castel, R. Noguera, A stiff extracellular matrix is associated with malignancy in peripheral neuroblastic tumors, Pedi- atr. Blood Cancer 64 (2017) e26449, doi:10.1002/pbc.26449.
[54] V. Gkretsi, T. Stylianopoulos, Cell adhesion and matrix stiffness: coordinating cancer cell invasion and metastasis, Front. Oncol. 8 (2018) 145, doi:10.3389/ fonc.2018.00145.
[55] Y. Shi, J. Wei, Z. Chen, Y. Yuan, X. Li, Y. Zhang, Y. Meng, Y. Hu, H. Du, Integra- tive analysis reveals comprehensive altered metabolic genes linking with tu- mor epigenetics modiﬁcation in pan-cancer, BioMed. Res. Int. (2019) e6706354 2019, doi:10.1155/2019/6706354.
[56] M.A. Marzinke, M. Clagett-Dame, The all-trans retinoic acid (atRA)-regulated gene Calmin (Clmn) regulates cell cycle exit and neurite outgrowth in murine neuroblastoma (Neuro2a) cells, Exp. Cell Res 318 (2012) 85–93, doi:10.1016/j. yexcr.2011.10.002.
[57] N. Sidell, Retinoic acid-induced growth inhibition and morphologic differen- tiation of human neuroblastoma cells in vitro, J. Natl. Cancer Inst. 68 (1982) 589–596.
[58] A. Voigt, P. Hartmann, F. Zintl, Differentiation, proliferation and adhesion of human neuroblastoma cells after treatment with retinoic acid, Cell Adhes. Commun. 7 (2000) 423–440, doi:10.3109/15419060009109023.
[59] C.J. Thiele, C.P. Reynolds, M.A. Israel, Decreased expression of N-myc precedes retinoic acid-induced morphological differentiation of human neuroblastoma, Nature 313 (1985) 404–406, doi:10.1038/313404a0.
[60] M. Clagett-Dame, E.M. McNeill, P.D. Muley, Role of all-trans retinoic acid in neurite outgrowth and axonal elongation, J. Neurobiol. 66 (2006) 739–756, doi:10.1002/neu.20241.
[61] J.P. Renaud, N. Rochel, M. Ruff, V. Vivat, P. Chambon, H. Gronemeyer, D. Moras, Crystal structure of the RAR-gamma ligand-binding domain bound to all-trans retinoic acid, Nature 378 (1995) 681–689, doi:10.1038/378681a0.
[62] J. Kreiter, A. Rupprecht, L. Zimmermann, M. Moschinger, T.I. Rokitskaya,
Y.N. Antonenko, L. Gille, M. Fedorova, E.E. Pohl, Molecular mechanisms respon- sible for pharmacological effects of genipin on mitochondrial proteins, Biophys. J. 117 (2019) 1845–1857, doi:10.1016/j.bpj.2019.10.021.
[63] M.K. Shanmugam, H. Shen, F.R. Tang, F. Arfuso, M. Rajesh, L. Wang, A.P. Ku- mar, J. Bian, B.C. Goh, A. Bishayee, G. Sethi, Potential role of genipin in cancer therapy, Pharmacol. Res. 133 (2018) 195–200, doi:10.1016/j.phrs.2018.05.007.
[64] V. Ayyasamy, K.M. Owens, M.M. Desouki, P. Liang, A. Bakin, K. Thangaraj,
D.J. Buchsbaum, A.F. LoBuglio, K.K. Singh, Cellular model of Warburg effect identiﬁes tumor promoting function of UCP2 in breast cancer and its suppres- sion by genipin, PLoS ONE 6 (2011) e24792, doi:10.1371/journal.pone.0024792.
[65] U. Valentiner, F.-.U. Valentiner, U. Schumacher, Expression of CD44 is associ- ated with a metastatic pattern of human neuroblastoma cells in a SCID mouse xenograft model, Tumour Biol. J. Int. Soc. Oncodevelopmental Biol. Med. 29 (2008) 152–160, doi:10.1159/000143401.
[66] A. Pathak, S. Kumar, Independent regulation of tumor cell migration by ma- trix stiffness and conﬁnement, Proc. Natl. Acad. Sci. 109 (2012) 10334–10339, doi:10.1073/pnas.1118073109.
[67] L. Dehmelt, S. Halpain, Actin and microtubules in neurite initiation: are MAPs the missing link? J. Neurobiol. 58 (2004) 18–33, doi:10.1002/neu.10284.
[68] P. Lamoureux, R.E. Buxbaum, S.R. Heidemann, Direct evidence that growth cones pull, Nature 340 (1989) 159–162, doi:10.1038/340159a0.
[69] D. Koch, W.J. Rosoff, J. Jiang, H.M. Geller, J.S. Urbach, Strength in the periphery: growth cone biomechanics and substrate rigidity response in peripheral and central nervous system neurons, Biophys. J. 102 (2012) 452–460, doi:10.1016/j. bpj.2011.12.025.
[70] N. Mundhara, S. Yadav, P.U. Shirke, D. Panda, A. Majumder, Substrate loss mod- ulus promotes the differentiation of SHSY-5Y neuroblastoma cells, Materialia 15 (2021) 100968, doi:10.1016/j.mtla.2020.100968.
[71] M. Yamazaki, K. Chiba, Genipin exhibits neurotrophic effects through a com- mon signaling pathway in nitric oxide synthase-expressing cells, Eur. J. Phar- macol. 581 (2008) 255–261, doi:10.1016/j.ejphar.2007.12.001.
[72] A. Nishiguchi, T. Taguchi, A pH-driven genipin gelator to engineer decel- lularized extracellular matrix-based tissue Triciribine adhesives, Acta Biomater. (2021), doi:10.1016/j.actbio.2021.06.033.
[73] M.I.L. Neves, M.M. Strieder, E.K. Silva, M.A.A. Meireles, Manufacturing natu- ral blue colorant from genipin-crosslinked milk proteins: does the heat treat- ment applied to raw milk inﬂuence the production of blue compounds? Future Foods 4 (2021) 100059, doi:10.1016/j.fufo.2021.100059.