Characterization and printability of Sodium alginate -Gelatin hydrogel for bioprinting NSCLC co-culture

3D bioprinting improves orientation of in vitro tumor models by offering layer by layer positioning of cancer cells and cancer associated fibroblasts (CAFs) which can replicate tumor microenvironment. Aim of this study was to develop a sodium alginate -gelatin (SA-GL) hydrogel by optimizing rheological parameters to print non-small cell lung cancer (NSCLC) patient derived xenograft (PDX) cells and lung CAFs co-cultures. SA-GL hydrogels were prepared, and rheological properties were evaluated. Both the cells were mixed with the hydrogel and printed using INKREDIBLE bioprinter. Hydrogels prepared with 3.25% and 3.5% (w/v) SA and 4% (w/v) GL showed higher printability and cell viability. A significant decline in viscosity with shear rate was observed in these hydrogels suggesting the shear thinning property of hydrogels. Spheroid size distribution after 15 days was in the diameter range of 50–1100 µm. Up-regulation of vimentin, α-SMA and loss of E-cadherin in co-culture spheroids confirmed cellular crosstalk. This study demonstrates that rheological optimization of SA-GL hydrogel enhances printability and viability of NSCLC PDX and CAF co-culture which allows 3D co-culture spheroid formation within the printed scaffold. Therefore, this model can be used for studying high throughput drug screening and other pre-clinical applications.

SA is the common natural polymer for hydrogel preparation due to its potential biomedical applications and possibility of formulating wide range of hydrogel viscoelasticity by ionic crosslinking with CaCl 2 . Further, CaCl 2 crosslinked SA was reported as biocompatible and does not affect cell viability during the gelling process with a concentration range up to 100 mM 19 .
GL has also been utilized in combination with SA as a hydrogel for printing tumor cells. GL at 37 °C is a solution but can form gel at <29 °C due to its conformational change from coil to helix. This property of GL is used to control the viscosity of alginate solution and thereby printable at different temperature range 20 . Pan et al. (2016), Li.Z et al. (2018), Chung et al. (2013), and Jiang et al. (2017) have already demonstrated the printability and biocompatibility of SA/GL hydrogel for printing of various living cells 8,[21][22][23] . However, characterization of rheological properties e.g. viscosity, viscoelasticity, shear thinning, or shear thickening behaviors of hydrogel for bio-printing of patient specific cancer cells has not been well described. The presence of CaCl 2 crosslinked SA matrixes and the coacervation between SA and GL, increases the thermal stability of GL and its delay in GL exudation from the bio fabricated scaffolds at 37 °C [24][25][26] .
Among advanced non-small-cell lung cancer (NSCLC) patients, epidermal growth factor receptor (EGFR) T790M mutation is the most common mechanism which results in approximately 60% of acquired resistance to first generation tyrosine kinase inhibitors (TKIs). Several preclinical studies with various treatment strategies have been reported but disappointingly acquired resistance to EGFR-TKIs develops within 12 months of treatment 27,28 . One of the reasons for the higher rate of clinical application failures in NSCLC drug discovery is currently available 2D models cannot replicate the three-dimensionality and heterogeneity of the tumor microenvironment. 3D cell culture can represent a stringent model for in vitro drug screening due to possessing several in vivo features such as cell-cell interactions, drug penetration, drug response, drug resistance and production of extracellular matrix 29,30 . Moreover, Patient derived tumor xenograft cells (PDX) in 3D culture recapitulate patient specific tumor microenvironment in vitro and can generate promising outcomes in preclinical drug evaluation. A major portion of non-small cell lung cancer (NSCLC) tumor stroma consist of Cancer Associated Fibroblasts (CAFs) which are associated with NSCLC progression, angiogenesis, epithelial to mesenchymal transition and metastasis. CAFs were also responsible for both intrinsic and acquired resistance to Tyrosine Kinase Inhibitor (TKIs) of NSCLC harboring T790M second mutation in EGFR 31 . Therefore, distinctive micro-environmental influences on NSCLC in vitro tumor model with co-cultured CAFs and patient derived tumor cells (EGFR T790M mutants) will be helpful for better understanding of stromal contribution to NSCLC. However, choice of materials and bioprinting of NSCLC PDX and CAFs co-culture with consistent viability and functionality currently are not well established.
The use of majority of reported natural or synthetic hydrogels as bioprinting material is limited due to biocompatibility and printability issues. Different hydrogels interact differently with different cell types; therefore, development of a hydrogel for printing cancer cells is very important. The primary requirements for a bioprinting material are a) Cell viability, b) printability (shear-thinning, compatible printing pressure, continuous extrusion without air bubbles), c) structure stability (stability of printed structure before crosslinking and after crosslinking), and d) non-toxic crosslinking. In this context we attempted to address all the important factors to develop a NSCLC co-culture platform using 3D bioprinting. This study for the first time aims to investigate a) Co-culture spheroid growth using patient derived NSCLC cell line and lung cancer associated fibroblasts (CAFs) in bioprinted scaffold. b) Printability of SA-GL hydrogels by tuning SA and GL concentration to improve cell viability and spheroid formation after bioprinting. c) a correlation between hydrogel rheology and patient derived cell survival and co-culture spheroid formation and d) scaffolds stiffness and stability during incubation period.

Results
Printability of SA and GL hydrogels. SA was used as a main component for bioprinting. SA solution at low concentration (<4%w/v) showed very low viscosity and ill-defined structural infusion between the spaces which made it difficult to print. In contrast, fully interconnected, well defined structures of scaffolds were printed after addition of GL to the SA solution. In the preliminary study, various hydrogels with 1-6% (w/v) SA and 3-8% GL were prepared and printed at room temperature using the INKREDIBLE bioprinter (Cellink, Sweden). Scaffolds with 3% SA with 4% GL (S1) showed deformation of the printed structure but 3.25-4% SA with 4% GL (S2) showed high structure fidelity. The extrusion pressures for printing 3, 3.25, 3.5, 3.75, 4% SA with 4% GL (named as S1, S2, S3, S4 and S5) at room temperature were approximately 20 kPa, 35 kPa, 60 kPa, 90 kPa and 120 kPa respectively. The extrusion pressure for printing hydrogels with 7:3, 6:4 and 4:8 (SA and GL ratios) was approximately 115 kPa, 145 kPa and 250 kPa respectively. Viscoelasticity of hydrogels. The rheological properties of five hydrogel samples with different concentrations of SA and fixed GL concentration were assessed. The viscoelasticity data of all sample hydrogels before and after crosslinking with CaCl 2 showed higher storage modulus (G′) than loss modulus (G′′) at room temperature (25 °C). Also, with the increase in SA concentration, storage modulus was increased which suggested that the mechanical properties of the hydrogel increased. Figure 1(A) shows that at 0.1 Hz for S1, S2, S3, S4 and S5, the storage modulus was approximately 59 Fig. 1(B) Tan δ vs frequency of the hydrogel mixtures. Tan δ values were found <1 for all the hydrogel mixtures. The strength of SA/GL hydrogel was increased as SA concentration increased in the hydrogel. Figure 1(C) demonstrates that increasing the concentration of SA in the system increases the elastic modulus of the system. Figure 1(D) also demonstrates the increase of storage modulus with increased concentration of both SA and GL.
Bioprinted scaffold stiffness and stability during incubation period. The stiffness of printed cell laden scaffolds at day 1,3,7,10 and 12 is depicted in Fig. 3(A,B) was evaluated by oscillation-frequency sweep and oscillation-time sweep measurement mode respectively. In all measurement, the storage modulus was higher than loss modulus and parallel in all frequencies (Fig. 3A). At day 1, scaffolds had higher stiffness which gradually decreased during the incubation period. The stiffness of all cell laden scaffolds was in the range of 1 kPa-8 kPa for 12 days. Figure 3(B) shows the stiffness of cell laden scaffolds which were evaluated by oscillation time sweep mode.
Development of Co-culture spheroids within printed scaffold. PDX and CAFs were observed to form small spheroids within the cell laden construct after 4 days of bioprinting. The spheroid size was increased over time. After 15 days, the spheroid size distribution pattern was characterized. Figure 5A shows spheroid size distribution in the printed scaffold and the sizes were in diameter range of 50 µm-1100 µm. At day 15, 25% of the spheroids had a spheroid size diameter range of 600 µm-1000 µm. NucBlue-ActinGreen staining in Fig. 5B shows the cytoskeleton of the printed co-culture spheroids on days 15 and 25. Figure 6 shows the loss of the cell adhesion molecule E-cadherin in the printed co-culture cells. The higher expressions of mesenchymal marker vimentin was also observed. CAFs in the printed cells showed higher expression of α-SMA levels. This phenomenon was observed in various cells and spheroids.

Discussion
Despite significant improvement in both technology and scientific approaches, NSCLC remains a leading cause of cancer associated deaths worldwide. Development of drug resistance and mutations in particular genes has decreased the treatment outcome as well as survival of the patient 32,33 . NSCLC PDX cells retain most of the mutations, pathologic and molecular characteristics of the original tumor. Moreover, PDX models are considered valuable in preclinical trials and the results will be used as a therapeutic guideline for future patients 34 . CAFs are important in the tumor microenvironment as they intensify tumor proliferation, angiogenesis, invasion and metastasis. CAFs help in TME remodeling and facilitate CAF activation and maintenance. Activation of CAFs results in cancer progression. CAFs are also known as one of the main regulators of tumor architecture and ECM. The synthesis and degradation of tumor architecture and ECM involves collagen, fibronectin secretion, and matrix metalloproteinase (MMP). CAFs are also considered as one of the most promising targets because genetically CAFs are normal, do not vary among patients and most importantly are less prone to evolve resistance [35][36][37] . Addition of CAFs with PDX cells into the 3D tumor model improves predictability of that model as well as enables better understanding and predictability of patients TME 38,39 . Although many co-culture models with 3D structures were reported with CAFs and tumor cells, most of them fail to provide a better understanding of the organization of different cells in the tumor microenvironment. Our hypothesis for this experiment was to develop an in vitro CAFs and tumor cells containing co-culture model which will enable studies to understand CAF-tumor cell interaction as in cancer patients. Further, this is the first report where we have used a SA-GL www.nature.com/scientificreports www.nature.com/scientificreports/ scaffold, which has been optimized using rheological parameters, to print tumor cells and CAFs using a 3D bioprinter which has shown to form viable spheroids.
Bioprinting technology not only allows/augments the cells to form spheroids but also facilitates the spheroids to form in different layers of the scaffold to mimic in vivo tumor microstructure. Therefore, this study is important to precisely develop multicellular spheroids using 3D bioprinting technology. This study will also allow researchers to study the complexity of the disease and to advance preclinical drug screening in the future. Three fundamental properties for cell bioprinting, (i) formation of consistent and cylindrical fibers (ii) stack printed layers into a coherent structure without spreading and (iii) lower extrusion pressure to achieve minimum stress, were used during preliminary screening of various SA/GL mixtures for 3D bioprinting. From these observations we chose five SA-GL ratios for potential hydrogel material for 3D bioprinting. All the five hydrogels were printed with the same protocol to evaluate changes in shape and size of printed structure after crosslinking.
Geometrics have been considered as a principal element in the role of the 3D printed cell laden scaffolds biofabrication and the right lattice design will help to mimic cellular microenvironment in addition to the bio-ink physicochemical properties 40,41 . Complexity of the biological tissues can be fabricated using multiple biomaterials and print heads. In bio-printing, printed scaffolds have an interconnected pore structure which ensures nutrient diffusion within the construct. Moreover, this interconnected porous structure also allows waste products diffusion from the scaffold. Scaffold size and the pore size are important as the cells primarily interact within the scaffold. Using that information, scaffolds can be designed to study drug delivery mechanisms and high throughput drug screening in pharmaceutical development.
In extrusion bioprinting, mechanical properties of the hydrogel plays an important role to determine the printed structure accuracy and cell viability 15 . In this study, all the rheology data was measured and stored within the rheometer control software with relevant rheological parameters (i.e., shear rate, shear stress, torque, normal forces, and angular frequencies). Further, this study does not include any conductivity test since our main focus was on the rheological properties of the hydrogels. Hydrogels which have higher viscosity, experience high shear forces to extrude through the nozzle during printing. Shear thinning behavior enables highly viscous hydrogels printable with structure accuracy 15,42,43 . High extrusion pressures are required in highly viscous hydrogels due to nozzle jamming. On the flip side, spreading of hydrogels is also a significant issue to print low viscous hydrogels 44,45 . During extrusion, hydrogels provide mechanical protection to the cells which can maximize the  Histogram shows the viability of PDX cells before printing, immediately after printing, 4 and 15 days after bioprinting with S1, S2, S3, S4 and S5 hydrogel. The viability of the cells was higher before and immediately after the bioprinting in all the hydrogels. After 4 days, the cell viability was higher in S1, S2 and S3 hydrogel than S4 and S5 hydrogel. After 15 days, more than 50% cells were found dead in both S4 and S5 hydrogel but S2and S3 showed highest cell viability. (B) Live-DEAD staining of printed PDX cells 1 day after bioprinting. Representative images show live cells (green) and dead (red) cells. www.nature.com/scientificreports www.nature.com/scientificreports/ modulus with increasing concentration of its constituents suggests the strength of the hydrogel. In only alginate solution, G′′ is dominant over G′ and behaves as a viscous fluid in all frequencies. Dominance of G′′ indicates the fluid-like property of the material. Due to insufficient energy or storage modulus, SA solution could not hold the printed structure 49,50 . Chung et al. (2013) reported that above 25 °C GL hydrogel alone and GL with 2% SA behaves like fluid (G′′ > G′) and after cooling G′ increased rapidly and crossed over G′′ and showed gel like characteristics. This report also suggested that SA-GL hydrogel should be printed at lower temperature where the hydrogel exhibits gel-like behavior 51 . In our experiment, frequency oscillation data showed that increase in storage (G′) and loss (G′′) modulus for 3-4% SA-4% GL hydrogel mixtures. The Tan δ value below 1 suggested that all the hydrogel mixtures are more gel like mixture than liquid like. The viscosity curves of all the hydrogel mixtures showed a similar pattern which suggests that all the hydrogels had shear thinning behaviors and the viscosity of these hydrogels were dependent on the dry content of the hydrogel and temperature besides to shear force and shear rate.
It has also been observed that 4% GL contributes a profound viscoelastic property when mixed with 3-4% SA which resulted in consistent extrusion through the nozzle. Non crosslinked 3-4% (w/v) SA solution alone cannot form a stable structure upon printing. Addition of GL imparts structural integrity to the printed structure as GL physically crosslinks at the printing temperature. Previous reports have shown that GL provides elastic characteristics of the hydrogel and improves cell adhesion within the hydrogel and SA contributes to the viscosity of the hydrogel and scaffolds stiffness 22,47,52 . Also, with the increase in SA concentration, pore diameter increases, and filament width decreases due to increase in the viscosity and storage modulus of the hydrogel 23 . Here, we fixed GL concentration and varied printability of the hydrogel by changing SA concentration. With increasing SA concentration, the modulus values increased which indicates the gelling behavior of the hydrogel. Moreover, printed scaffold resolution is dependent on the material rheological property and bioprinter condition which includes nozzle diameter, print-bed temperature and printing speed. The scaffold width depends on the inner diameter of the printing nozzle and the viscoelastic property of the hydrogel. The small nozzle diameter with high extrusion pressure results higher resolution but it increases cellular damage. Previous studies reported that higher nozzle diameter with low pressure printing results a structure with low shape fidelity or limited resolution 53 . Here, SA-GL hydrogel with multiple process parameters was printed with 410 µm (inner diameter) nozzles and was found favorable for cell viability after printing.
Earlier studies conducted by other investigators established the printability and biocompatibility of SA-GL hydrogel with SA concentration 3-7 w/v% and GL concentration 6-8 w/v% 8,21,23,54 . In our experiments, all the hydrogels with 3.25-4% SA and 4% GL allowed construction of 3D scaffolds with good processability and shapes at room temperature. SA lower than 3.25% (w/v) and with 4% (w/v) GL could not hold the printed structure due to lack of required storage modulus. Giuseppe et al. (2018) reported that increasing the concentration of SA led to more accurate printing and higher concentration of GL (10%) reduced the printing accuracy 48 . In our experiments, 3.25% and 3.5% SA with 4% GL showed suitable flow properties through the nozzle tip at room temperature with superior shape fidelity and high cell viability immediately after the extrusion and 15 days after printing. www.nature.com/scientificreports www.nature.com/scientificreports/ Our results are supported by a study done by Katja et al. (2016) who showed that bioinks with viscosity range of 30-6 × 10 7 mPa can be printable by extrusion type printer 8 .
Scaffold's stiffness is also a vital parameter to maintain the structural integrity as well as tumor cell growth. Haobo et al. (2018) demonstrated that manipulation of mechanical stiffness of bio-fabricated 3D scaffolds helps to ensure customized physiological stiffness and mechanical support for high cell attachments. Further, stiffness of the hydrogel influences the cell proliferation and tumor spheroid growth 24 . Wang et al. (2014) and Yunfeng et al. (2018), reported that stiff hydrogel decreases cell proliferation by delaying/arresting cell cycle progression. Stiffer hydrogels create compact micro-environments which decrease cell growth and proliferation and our study suggests that the stiffness of hydrogels can be controlled by polymer concentration 55,56 . Cavo et al. (2016) reported that soft hydrogel allows cells to form bigger spheroids with time whereas the cluster size did not change 57 . In this study, we observed that the stiffness of printed cell laden scaffolds were in the range of physiological lung tumor stiffness which is reported to be in the range of 1 kPa to 50 kPa 58,59 . The stiffness of our bio-printed cell laden scaffolds was 1.00-8.00 kPa which is in the range of lung tumors stiffness.
In patient tumors, cancer cells may activate the stromal fibroblast into CAFs through stimulation of paracrine growth factor 60 . Transforming growth factor-β (TGFβ) helps to activate fibroblasts and platelet-derived growth factor (PDGF) from cancer cells play important role in fibroblast activation and proliferation 61,62 . Therefore, a mutual interaction between cancer cell induced fibroblast activation and fibroblast induced cancer proliferation and metastasis makes CAFs tumor supporter. In this study our main goal was to develop a method for co-culturing spheroids using bioprinting technology. We tried different ratios of PDX and CAFs for bioprinting (data not shown) but observed that at lower density, enough fibroblasts did not survive in the co-culture. As a preliminary study, we used a 50-50 cell density to develop the model. Further studies are required to optimize the hydrogel recipe (Bioink) which can enhance fibroblast growth. PDX and CAFs started forming spheroids within the printed heterogeneous cell laden scaffold 4 days after printing. Increasing spheroid size over time up to 15 days indicates proliferation of individual cells. Jiang et. al (2017) showed that bio-printed MDA-MB-231 and IMR-90 cells self-aggregated to form multicellular spheroids and the size of the spheroids increased over time 20 . In our study we found a similar trend of increasing spheroid size over time.
Current in vitro 3D cell culture models consist of mainly cancer cells whereas co-culture of cancer cell with CAFs is always required as CAFs are involved in the initiation and progression of the tumor.  ) showed that cross-talk between lung cancer PC9 and H358 cells and CAFs is responsible for EGFR-TKI Erlotinib resistance in NSCLC 69 . Although several studies mentioned a clear link between TKI resistance and EMT but there are very few effective NSCLC-CAFs co-culture models which have been developed for high throughput drug screening. This is the first report to develop a Lung T790M PDX cells co-cultured with NSCLC CAFs using 3D bioprinting technology. Here E-cadherin, was chosen as a marker for adherent cells, vimentin indicated the shift of epithelial cells into mesenchymal phenotype and α-SMA indicated mesenchymal phenotype of CAFs 68 . Moreover, expression of α-SMA typically identifies the presence of CAFs in the co-culture. In this study, down-regulation in E-cadherin, concomitant up-regulation of vimentin and higher expression of α-SMA in the printed co-culture spheroids suggested the crosstalk between CAFs and PDX cells which promotes EMT. Forming spheroids using patient derived cells with controlled and uniform size is always a challenge in cancer research. However, Cavo et. al. 2016 showed that the stiffness of the hydrogel determines the spheroid size 57 . Using rheological principles, other polymers can be used which may reflect similar cellular behavior and spheroid growth. Moreover, custom built bioprinting system may enable to generate a controlled and uniform spheroid by eliminating manual cell-seeding and non-uniform cell distribution in the hydrogel.

conclusion
We demonstrate that 3D bioprinting technology can be used to develop in vitro tumor co-culture spheroids by using optimized SA and GL hydrogel. This model also allows investigating tumor stroma interactions to mimic in vivo tumor microenvironments. However more detailed studies are required to overcome some other concerns, but this study preliminarily showed that increased cell viability and tumor-stroma interaction can be achieved in NSCLC PDX cells and CAFs co-culture through 3D bioprinting.

Materials and Methods
Sodium alginate (MW 50000 Da) and gelatin type B from bovine skin (MW ~50,000-100,000 Da), calcium chloride analytical grade, DMEM/F12, RPMI, bFGF and EGF were purchased from Sigma-Aldrich (USA). B27 supplement, NucBlue, Actin-Green488 reagents were purchased from Life technologies (USA). Live-Dead assay reagent was purchased from Biotium Inc. (USA). Vimentin, E-cadherin and α-SMA were purchased from Cell Signaling Technology (USA). Anti-rabbit and anti-mouse FITC and Rh conjugated secondary antibodies were purchased from Santa Cruz biotechnology (USA).
Preparation of hydrogels. GL type B was dissolved in 0.45%w/v saline solution at 40 °C for 2 h using overhead stirrer (LIGHTNIN, USA) and magnetic stirrer. SA powder was measured and mixed into the GL solution for 3 hours. The hydrogels were placed under UV light for 30 minutes and stored at 4 °C until use. Hydrogels were kept at room temperature for 3 h before printing. (2019) 9:19914 | https://doi.org/10.1038/s41598-019-55034-9 www.nature.com/scientificreports www.nature.com/scientificreports/ Rheological characterization of hydrogels. The mechanical properties of the hydrogels were studied using a combination of rotational and oscillatory rheology experiments on the AR 1500Ex Rheometer (TA instrument, USA). To load hydrogel sample, 2 ml hydrogel was placed on the lower plate of the rheometer and then melted to 34 °C for 3 minutes. The gap was set to 1 mm (parallel plate geometry, 20 mm diameter), and then cooled to 4 °C at a rate of 1 °C/min. Excess sample around trim gap was removed with spatula. For each measurement fresh samples were used with independent three replicates. The elastic and viscous moduli of hydrogels were evaluated at 25 °C using oscillatory rheology measurements (oscillation time sweep geometry) at a strain of 1% and frequency 10 rads −1 . Next, frequency sweep measurements were carried out in the range 0.1-100 rad/s at a constant strain of 0.1% to show samples physical (viscoelastic and gel like) behaviors. Viscosities of all the samples were evaluated at 25 °C and at shear rate 1 to 100 s −1 .
Scaffold fabrication and hydrogel printability. Inkredible 3D bioprinter (CELLINK Inc, Sweden) was used to print all the hydrogels. The hydrogels were loaded into 3 ml printing cartridges (CELLINK Inc, Sweden) and connected to a controllable pressure regulator (0-1000 kPa). The hydrogel printing flow rate was optimized by adjusting the extrusion pressure (25-125 kPa) and Z-axis of 0.1 mm from nozzle tip to print bed. Before printing, hydrogels were taken out from the refrigerator and kept at room temperature for 3 h. Three layered scaffolds (10 × 10 × 1 mm) were printed on a petridish using 22 G nozzle (410 µm diameter). The dimensions of 3D printed scaffolds were measured after CaCl 2 cross-linking. To measure the line width, scaffolds were printed (n = 5) with all the hydrogel mixtures. The line width of all the scaffolds was measured using CellSens software (Olympus, Japan).
Bioprinted scaffold stiffness and stability during incubation period. The stiffness of scaffolds was evaluated with AR 1500Ex Rheometer (TA instrument, USA) with 1 mm measurement gap (parallel plate geometry, 20 mm plate diameter). To measure the stiffness of cell laden scaffolds, designed scaffolds with a 20 mm diameter,1 mm height and 65% infill density were printed and cross-linked with 100 mM CaCl 2 for 3 minutes. Excess cross-linkers were removed and washed with HBSS. The scaffolds were incubated in 6 well plates with DMEM high glucose medium, at 37 °C with 5% CO2. The cultured scaffold was taken out from the well plates and viscoelasticity was ascertained at rheometer temperature of 37 °C and with plate to plate gap of 1 mm. Oscillation frequency sweep was run from 0.1 rad/s to 100 rad/s with 1% strain to check the frequency dependency stiffness of the scaffolds. Additionally, oscillation time sweep with a strain of 1% and angular frequency of 10 rad/s was used for stiffness measurement. The stiffness of the scaffold was measured at day 1, 3,7,10 and 12 at 37 °C. All scaffolds were constructed from 3.25% SA and 4% GL and each cell-laden scaffold had the initial estimated cells of 2 × 10 5 cells per scaffold.

Cell printing. NSCLC PDX (EGFR T790M) cell line was obtained from Dr. Rishi's Laboratory at Wayne
State University and Lung CAFs (AA0022) was obtained from Dr. Noyes Laboratory, Moffit Cancer Center. A total of 10 × 10 6 cells were mixed homogeneously with the hydrogel using a spatula in 10:1 ratio and printed with INKREDIBLE bioprinter (Cellink, Sweden). For co-culture, 5 × 10 6 cells of each PDX and CAFs were homogeneously mixed with the hydrogel and printed with print head 1 in INKREDIBLE bioprinter. 22 G needle (CELLINK, Sweden) was used for bio-printing throughout the experiment. After printing, scaffolds were crosslinked with 100mMCaCl 2 crosslinker solution and transferred to 6 well plates. The printed cell laden scaffolds were cultured in DMEM/F12 and RPMI media supplemented with B27 supplement, recombinant human epidermal growth factor (EGF) and recombinant human basic fibroblast growth factor (bFGF).
Cell viability. Printed cell viability was studied by using Live-Dead Assay Kit (Biotium Inc.) Briefly, printed cells were stained with 2uM calcein AM/4 uM EthD-III for 1 h at 37 °C, 5% CO 2 . Cell laden scaffolds were washed twice with serum free media and photographs were taken by using Olympus IX70 microscope. Percentage of live and dead cells were determined by counting the cells using Image-J (NIH, USA) software and a histogram was made to compare the cell viability in different printed groups.
Microscopy for spheroid formation. Nucleus and Actin staining was performed to identify co-culture spheroid growth inside the printed scaffold. Briefly, cell laden scaffolds were fixed using 4% formaldehyde for 1 h at room temperature and washed twice with HBSS for 15 minutes each. The scaffolds were then permeabilized with 0.1% Triton X-100 in HBSS. NucBlue and ActinGreen488 (Life Technologies) staining was performed according to manufacturer protocol. Images were taken by using confocal microscope (Nikon).
Immuno-fluorescence staining for cellular crosstalk. Immuno-fluorescence staining was performed to determine the cross-talk between printed PDX and CAF in co-culture spheroid. After 15 days, the co-culture spheroids were removed from the printed scaffold by dissolving the scaffold in Algimatrix dissolving buffer (Gibco). Co-culture spheroids were fixed with 4% formaldehyde for 1 h at room temperature and then treated with 0.1% triton X-100 for another 1 h at 37 °C. The co-culture spheroids were then washed in PBS twice and blocked overnight using 5% BSA in PBST. E-cadherin, vimentin and alpha-SMA antibodies with 1:100 dilutions in 3%BSA were incubated overnight. FITC and Texas Red conjugated secondary antibodies were used for 2 h and finally NucBlue staining was performed. Images were taken by using a fluorescence microscope (Olympus).

Statistical analysis.
The experimental results were expressed as means ± standard deviation. Statistical analyses were performed by Students T Test. The p value < 0.05 were considered as statistically significant.