MH-S, IC-21, and RAW 264.7 murine cell lines were purchased from American Type Culture Collection (ATCC, Manassas, Virginia, United States). All the cells were cultured between passages 2–10 in complete media, RPMI 1640 Medium (ThermoFisher Scientific) supplemented with 10% heat-inactivated fetal bovine serum (certified Gibco heat inactivated, USA origin) and 1% penicillin/streptomycin. Primary BMMs and PMs were isolated from healthy BALB/c and C57BL/6 mice (6–12 weeks old, Jackson Laboratories, Bar Harbor, Maine, United States), respectively, following the approved Institutional Animal Care and Use Committee (IACUC) protocol at the University of Delaware. Bone marrow was extracted from femurs and tibias, and cells were cultured in BMM differentiation media (complete media supplemented with 30% L929 conditioned media) until confluent in a T-75 flask before using for the experiments. PMs were isolated using a peritoneal lavage. Cells were isolated and cultured per standard techniques (Gonçalves and Mosser, 2015). The cells were centrifuged (500 g for 5 min) and resuspended in complete media and allowed to attach overnight before using for the experiments.

Cells were seeded, allowed to adhere for 24 h, then stimulated with M1 or M2 stimuli for 24 h. Immortalized cell lines (seeded at 10e5 cells/well in a 6-well plate) were polarized with M1 and M2 stimuli for 24 h to compare the polarization profile of the cell line. The polarization profile of primary cells was determined by treating the cells (seeded at 50e5 cells/well in a 6-well plate) with M1 or M2 stimuli for 24 h. Cell lines were seeded at a lower density to account for their higher proliferation rates, assuming the cells lines would double approximately twice between seeding and flow cytometry analysis while primary cells would exhibit limited proliferation and potentially lower viability within the same timeframe.

Lipopolysaccharides (LPS) (20 ng/mL, Escherichia coli O111:B4 [Millipore Sigma, Rockville, MD, United States]) and interferon (IFN)-γ (20 ng/mL, PeproTech^®) were used as M1 stimuli. Interleukin (IL)4/IL13 (10 ng/mL each, PeproTech^®) was used as M2 stimuli. IL4/IL13 (10 ng/mL each) supplemented with 50% tumor-conditioned media (TCM) from A549 cells (IL4/IL13/TCM) was used as a tumor-conditioned stimuli. The cytokine concentration was decided based on values found in previous literature, where concentrations between 10 ng/mL to 100 ng/mL are typical for polarization stimuli (Andreakos et al., 2004; Edwards et al., 2006; Genin et al., 2015; Bomb et al., 2022).

To compare the plasticity of different cell lines, pre-polarized macrophages (M1 or M2) receiving 24 h of dose 1 were repolarized with either the same stimuli or a different stimulus for an additional 24 h with dose 2 (Supplementary Figure S1). The polarization profile was quantified using flow cytometry.

The plasticity index (PI) was defined to assess the relative polarization of macrophages to alternative stimuli compared to the polarization with the same stimuli. PI was quantified separately for each marker using median fluorescence intensity (MFI) using Eqs 1, 2 (for M1 markers) and Eqs 3, 4 (for M2 markers), where subscripts indicate either dose 1 (initial polarization) or dose 2 (plastic change). A (+) sign convention was adopted for when the repolarization would be expected to align with upregulation of the marker type (i.e., + for M1 markers of CD86 and MHCII when M2→M1; (−) for M1 markers when M1→M1). Absolute values >1 indicate that cells were plastic and experienced repolarization upon secondary stimuli based on the respective marker.

M1 Markers (CD86, MHCII) P I M 1   M 1 → M 2 = − I F N γ 1 → I F N γ 2 I F N γ 1 → I L 4 / 13 2 (1) P I M 1 M 2 → M 1 = +   I L 4 / 13 1 → I F N γ 2 I L 4 / 13 1 → I L 4 / 13 2 (2)

M2 Markers (CD206, EGR2) P I M 2   M 1 → M 2 = +   I F N γ 1 → I L 4 / 13 2 I F N γ 1 → I F N γ 2 (3) P I M 2 M 2 → M 1 = −   I L 4 / 13 1 → I L 4 / 13 2 I L 4 / 13 1 → I F N γ 2 (4)

The phagocytic ability of cell lines was determined by quantifying the uptake of fluorescent (FITC) phosphatidylcholine (PC) coated lipid microparticles (3 μm diameter, Echelon Biosciences, Salt Lake City, Utah, United States). Uniform PC particle suspension (achieved by sonication and vortexing) was seeded (∼2 particles per cell) on polarized cell lines (20 ng/mL IFNγ or 10 ng/mL each IL4/13). The particle uptake was determined after 6 h by quantifying particle positive cell population [% PC positive (+ve)] and MFI of the internalized particles using flow cytometry on FITC channel.

Cells on TCP were detached using Accutase (Innovative Cell Technologies, Inc.) for 30 min and washed with phosphate buffered saline (PBS), followed by staining with Zombie Yellow (10 min). Cells were blocked with anti-CD16/32 (Fc block) and immunostained with anti-CD86 and anti-MHCII (30 min). Cells were then fixed with 4% paraformaldehyde (PFA) in PBS (15 min), permeabilized using Intracellular Staining Permeabilization Wash Buffer (BioLegend), and immunostained with anti-CD206 and EGR2 (30 min). Further information on antibodies can be found in Supplementary Table S1. BMMs and PMs were gated on CD11b and F4/80 double positive. For internalization studies, cells were washed twice with FACS buffer (2% FBS in PBS) to remove unbound particles and quenched with 0.4% Trypan Blue dye to quantify internalized particles. Samples were analyzed using ACEA NovoCyte Flow Cytometer. The gates were set using untreated samples and MFI was plotted for each marker (Supplementary Figures S2–S4).

Cell lines (seeded at 10e4 cells/well) and primary cells (seeded at 50e4 cells/well) were washed with PBS and fixed with 4% PFA (15 min). Note, higher cell seeding density was selected for terminally differentiated primary cells given their lack of proliferation over time. Cells were permeabilized by washing with 0.2% Triton-X (5 min) and stained with 2:250 Phalloidin-Rhodamine (ActinRed™ 555 ReadyProbes™ Reagent, Invitrogen™) for 20 min. Cells were washed with PBS and incubated with 70 nM 4',6-diamidino-2-phenylindole (DAPI) (Invitrogen™) for 10 min to stain nuclei. Cells were imaged on BioTek Cytation 5 Multimode Imager. Cell area and circularity were analyzed using Fiji imaging package for the ImageJ2 1.53q software. For morphological analysis, the area selection tool was used to trace the macrophage area and determine the cell spread manually. Circularity for each macrophage was then quantified using the built-in circularity analysis tool.

Bioinks and activators (Cat. no. Px02.29PH, 1.1 kPa with RGD, GFOGER, YIGSR, and HA) were generously provided by Inventia Life Science (Sydney, Australia). The 3D cultures were printed using the RASTRUM™ bioprinter (Inventia Life Science) following protocols previously described for breast cancer cell lines (Engel et al., 2022). Briefly, the printing protocol was created via RASTRUM™ Cloud (Inventia Life Science). Cells were seeded in the F177 activator at a density of 20e6 cells/mL (BMMs) and 5e6 cells/mL (RAW 264.7) and printed within Large Plug model in a 96-well plate (Corning—#CLS3904). The Large Plug model (0.5 mm in height) fills the whole surface of a well in a standard 96 well plate (surface area ∼0.32 cm²) as shown in previous literature (Gomila Pelegri et al., 2023). Seeding densities were varied to account for differences in proliferation between the cell types, as shown in Supplementary Figure S6. A lower seeding density of RAW 264.7 cells allowed for proliferation over the experiment time course with the aim of having the number of RAW 264.7 cells be similar to the number of BMMs after 7 days; the initial cell densities also provided a mix of cell-cell and cell-matrix interactions with single cells/small clusters of encapsulated cells throughout the 3D culture upon printing. RAW 264.7 cells were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) complete media (10% heat-inactivated fetal bovine serum and 1% penicillin/streptomycin) and BMMs were maintained in BMM culture media (DMEM with 10% heat-inactivated fetal bovine serum and 1% penicillin/streptomycin supplemented with 30% L929 conditioned media).

The viability of macrophages in 3D culture was assessed by a LIVE/DEAD Viability/Cytotoxicity Kit (ThermoFisher Scientific). The LIVE/DEAD kit contained calcein-AM to indicate viable cells (live cells fluoresce green; excitation [ex.] 494 nm, emission [em.] 517 nm) and ethidium homodimer-1 to label dead cells (dead cells fluoresce red; ex. 528 nm, em. 617 nm). Hydrogels were washed 3x with PBS + 3 min PBS incubation, then incubated (37°C, 5% CO₂) for 10 min in a solution of calcein-AM (2 µM) and ethidium homodimer-1 (4 µM). After incubation, hydrogels were washed 3x with PBS + 3 min PBS incubation. Hydrogels were imaged with a confocal microscope (LSM 800, Zeiss; ×10 objective and frame size of 1,024 × 1,024, 180 µm z-stack, 8.38 µm/stack, three images per sample). Cell viability was quantified with Imaris 9.9.1 imaging software (Oxford Instruments) using Find Surfaces for AF488 (green) channel and Rhoda (red) channel functions. The percentage of viable cells was calculated by the number of green cells/total number of cells × 100%.

For 3D Immunostaining, cells were washed with PBS and fixed with 4% PFA (15 min). Cells were permeabilized by washing with 0.2% Triton-X (5 min) and stained with Phalloidin-Rhodamine (2:50, 20 min). Cells were washed with PBS and incubated with Hoechst 33342 solution (ThermoFisher Scientific) (1:200) to stain nuclei. Hydrogels were imaged with a confocal microscope (LSM 800, Zeiss; ×20 objective and frame size of 1,024 × 1,024, 200 µm z-stack, 2 µm/slice, three images per sample). Cell area and circularity were analyzed using Fiji imaging package for the ImageJ2 1.53q software. For morphological analysis, the area selection tool was used to trace the macrophage area and determine the cell spread manually. Circularity for each macrophage was then quantified using the built-in circularity analysis tool.

All centrifugation steps and washes were complete at 500 g for 5 min. Cells were retrieved from hydrogel matrix to complete downstream flow cytometry analysis. Here, 75 µL of cell retrieval solution (F235, Inventia Life Science) was added to each well for 45 min to enzymatically degrade the hydrogel matrix followed by addition of 100 µL of PBS. To remove contents from the well, 3x PBS washes were performed, passed through a 70 μm cell strainer, and centrifuged and washed 2x with 2% FBS in PBS buffer. Cells were used for flow cytometry analysis as described in Section 2.5.

GraphPad Prism 9 was used to perform statistical analyses. All quantitative data are represented as mean ± standard error of the mean. Tukey’s multiple-comparison test was used to generate p-values in ANOVA multiple comparisons unless stated otherwise. Three or more biological replicates were used for statistical analysis.

We first investigated the baseline expression of M1 and M2 markers and morphology for both untreated cell lines and primary cells to understand basal profiles. RAW 264.7 (BALB/c origin, monocyte-derived), IC-21 (C57BL/6 origin, peritoneum), and MH-S (BALB/c origin, lungs) cells were selected as immortalized cell lines. BMMs (BALB/c and C57BL/6 origin) and PMs (C57BL/6 origin) were selected as primary cells. Cells were seeded for 24 h before analyzing M1 (CD86 and MHCII) and M2 (CD206 and EGR2) markers using flow cytometry. M1 and M2 markers were selected based on our previous work (Jarai and Fromen, 2022), where we showed the specific marker expression strongly correlated with polarization stimuli.

Statistically significant differences were observed for all the markers between different cell lines (Figure 2A). RAW 264.7 cells showed the lowest marker expression for both M1 and M2 markers compared to MH-S and IC-21 cells. We hypothesize that the low marker expression for RAW 264.7 cells is due to the monocyte-derived origin of the cell line, while both MH-S and IC-21 are sourced from tissue-resident macrophage cell lines, where the baseline expression differs depending on the tissue type. This is also reported in previous literature where BMMs showed differences in marker expression compared to tissue-resident macrophage counterparts both in the lung and the peritoneal cavity (Misharin et al., 2017; Zhao et al., 2017). IC-21 cells showed the highest basal M2 marker expression of both CD206 and EGR2. This basal phenotypical skewing is somewhat counterintuitive to the genetic background of the strain used to isolate the cells; RAW 264.7 and MH-S cells are derived from BALB/c mice, which is characterized as a more Th2 strain, whereas IC-21 cells are derived from C57BL/6 mice, which is characterized as a more Th1 strain (Santos et al., 2006; Jones et al., 2013). However, we speculate these basal M1 and M2 dispositions are attributed to the extent of macrophage differentiation. This is also highlighted in previous studies, which showed a higher M2 marker expression on cells correlated with macrophage differentiation marker expression (F4/80) irrespective of the genetic background of the mice (Chamberlain et al., 2009; Chamberlain et al., 2015). Furthermore, MH-S cells had the lowest surface area and a statistically higher circular morphology compared to both IC-21 to RAW 264.7 cells. This may be due to the increased surface tension observed for AMs resulting in a more circular morphology. Also, F-actin staining of MH-S cells showed that actin structure was diffuse in the cytoplasm, whereas for both IC-21 and RAW 264.7 cells, increased formation of cytoplastic protrusions was observed.

Next, we looked at the marker expression and morphology for untreated ex vivo primary cells and again observed statistically significant differences between PMs and BMMs of two strains (Figure 2B). Like the cell lines, PMs showed statistically higher MHCII expression than both BMMs (C57BL/6 and BALB/c); however, contrary to cell lines, PMs showed statistically lower CD206 expression compared to BMMs. BMMs from C57BL/6 origin (Th1 skewed) showed higher CD206 and EGR2 expression compared to those from BALB/c origin (Th2 skewed), which again is counterintuitive. However, we hypothesize the hyper-physiological conditions of TCP may contribute to this basal expression and, as shown in our later evaluations, basal expression levels do not directly correlate to polarization potential. Unlike the cell lines, PMs and BMMs (C57BL/6 and BALB/c) showed similar elongated cell morphology. C57BL/6 BMMs had lowest mean cell area (218.7 µm²) compared to BALB/c BMMs (357 μm²) and PMs (362 μm²). These findings highlight the differences in marker expressions and cell morphology between the cell lines and the primary cells from similar tissue origin.

Next, we studied the response of M1 and M2 stimuli on macrophage cell lines. Cells were polarized for 24 h with M1 (LPS or IFNγ) or M2 (IL4/13 or IL4/3/13/TCM) stimuli for 24 h before analyzing the phenotypical profile using flow cytometry and normalizing MFI to the untreated (UT) condition of each cell type, where upregulation of a particular marker is indicated by a normalized MFI greater than 1. For M1 markers (CD86 and MHCII, Figures 3A, B, respectively), IFNγ promoted a robust marker expression across all the cell lines; however, LPS only resulted in increased M1 marker expression for RAW 264.7 cells. MHCII, an antigen presentation marker, was highly upregulated in M1-stimulated RAW 264.7 cells, followed by MH-S and IC-21 cells (Figure 3B). RAW 264.7 cells are of monocyte-derived origin that has been shown to closely resemble primary BMMs (Berghaus et al., 2010). Furthermore, Zajd et al. reported that compared to PMs, primary BMMs (C57BL/6) are skewed more toward a M1 phenotype when stimulated, which may explain both increased MHCII and CD86 marker expression in RAW 264.7 cells (Zajd et al., 2020). Reduced M1 marker expression upon M1 stimulation in IC-21 cells may be attributed to the origin of tissue-resident macrophages. This was also reported by Stevens et al., who showed that PMs show a lower M1 response to mycobacterial infection compared to AMs (Stevens et al., 2021). Interestingly, for IC-21 cells, M2 stimuli downregulated the expression for M1 markers (statistically significant for CD86), whereas, for other cell lines, no such difference in the expression levels was observed.

For M2 markers, we observed interesting trends. CD206 expression was upregulated by both M2 stimuli in IC-21 and RAW 264.7 cells, while no changes were observed in MH-S cells (Figure 3C). Surprisingly, for EGR2, both M2 stimuli failed to upregulate EGR2 expression (Figure 3D). These data indicate that different cell lines upregulate M2 marker expressions to different levels. A higher dose of M2 stimuli or a repeated dose of M2 stimuli might be required to effectively polarize cells and generate persistent M2 marker expression. Furthermore, M1 stimuli of LPS and IFNγ upregulated the marker expression of both CD206 and EGR2. While LPS and IFNγ are thought to be only M1 stimuli, macrophages exist on a spectrum and therefore can express both inflammatory and anti-inflammatory markers at the same time (Mosser and Edwards, 2008; Smith et al., 2016). Further, macrophages generally progress from inflammatory to anti-inflammatory over time (Risser et al., 2023). Previous works have shown that LPS led to increased expressed of M2-like markers and cytokines, (such as IL-10, Arg1, CD206) (Ci et al., 2010; Zhang et al., 2019; Zhen et al., 2023). Further, immortalized cells lines have chromosomal abnormalities or mutations that allow their propagation, which may also affect polarization markers; therefore, the choice to use a cell line should be carefully considered.

For M1 stimuli, compared to IFNγ stimulation, LPS-stimulated cells showed reduced M1 marker expression across the cell lines. IFNγ stimulation is modulated by JAK-STAT pathway, which promotes the upregulation of pro-inflammatory genes leading to immediate response to IFNγ (Sikorski et al., 2012). In addition to pro-inflammatory genes, interferon regulatory factors are also secreted, which helps sustain IFNγ induced response. On the other hand, LPS stimulation is modulated by Toll Like Receptor 4 (TLR4) signaling to promote upregulation of pro-inflammatory genes and induce an M1 response. However, due to a lack of sustained pro-inflammatory response, LPS stimulation might generate a weaker M1 stimulation compared to IFNγ at equivalent mass dosages. Furthermore, IFNγ stimulation also regulates the transcription of interferon-stimulated gene, which contributes to the upregulation of MHC molecules (Ivashkiv, 2018). Our results follow this trend where MHCII expression is upregulated more by IFNγ compared to LPS across all the cell lines.

For M2 stimuli, IL4 and IL13 were used together for M2 stimulation because both IL4 and IL13 have been shown to involve phosphorylation of STAT6, which promotes STAT6 translocation to the nucleus to regulate the transcription of M2 genes and promote M2 polarization (Gordon and Martinez, 2010). Further, adding tumor-conditioned stimuli with IL4 and IL13 yielded a similar phenotypical profile as IL4/13 stimulation. This may be due to the heterogeneous source of tumor-conditioned stimuli, which in our study was obtained from A549 cells (human-derived lung cancer cells) for use with murine macrophages. Previous studies have successfully shown differences in the phenotypical profile of macrophages using tumor-conditioned stimuli from homogenous sources for both mice and human models (Benner et al., 2019; Little et al., 2019). Based on these observations, we selected IFNγ and IL4/13 for further evaluation.

Overall, our results indicate that the RAW 264.7 cell line is the most responsive to both M1 and M2 stimuli compared to both MH-S and IC-21 cells. This robust response, as well as quick proliferation, may be a reason why these cells are extensively used within in vitro characterizations in the literature.

Next, we assessed the effect of these stimuli on the cell morphology of different cell lines by quantifying cell area and cell circularity (Figure 4A). In general, MH-S cells exhibited a trend of lower cell spread and higher circular morphology compared to both IC-21 and RAW 264.7 cells (Figures 4B, C). This difference could arise from the cell origins, where AMs reside at the air-liquid interface in the lung and face increased surface tension resulting in a more circular morphology with compact actin fiber assembly. The area of both MH-S and IC-21 cells was independent of M1 and M2 stimuli, while LPS significantly increased the area of RAW 264.7 cells, highlighting differences in stimuli-dependent morphology changes amongst different cell lines (Figure 4B). Next, we quantified the circularity of the cells to determine cell elongation upon stimuli treatment (Figure 4C), where McWhorter et. al have reported a higher circularity upon M1 stimuli treatment and lower circularity or increased cell elongation upon M2 stimuli treatment (McWhorter et al., 2013). We observed that cell morphology, like cell spreading, exhibited trends of dependence on tissue of origin. For example, MH-S cells showed a more circular morphology independent of the stimuli, while IC-21 cells showed a higher circularity with M1 stimuli and a more elongated morphology upon M2 stimuli treatment, as previously reported in the literature. For IC-21 cells, morphological changes correlated well with phenotypical changes, where M1 marker expression (CD86) was increased with M1 stimulation and decreased with M2 stimulation (Figure 3A). Interestingly, for RAW 264.7 cells, IL4/13 treatment significantly increased cell circularity.

Overall, we report cell line-specific differences in morphology in response to different stimuli. Except for RAW 264.7 cells, we did not observe statistically significant changes in morphology upon stimuli treatment. We hypothesize this may be due to the stiff TCP substrate used in the experiments. The stiffness of many body tissues is in the kilopascal (kPa) range, while the stiffness of these culture dishes is in the gigapascal range, which may negate the effect of soluble stimuli on cell morphology. We, along with others, have shown an increased morphological response to stimuli treatment when cultured on a more physiologically relevant substrate (Sridharan et al., 2019; Bomb et al., 2022).

Phagocytosis is a critical macrophage function responsible for the uptake and removal of foreign pathogens to maintain homeostasis and is often impaired in diseased conditions (Allard et al., 2018). We investigated the effect of polarization stimuli on phagocytosis in different cell lines by using fluorescent lipid microparticles coated with PC as phagocytic mimics and allowed the cells to internalize the particles for 6 h. Uptake of the PC microparticles was confirmed in all the cell lines with the presence of a second peak (Figure 5A, gray), which is quantified in Figure 5B. For the untreated condition, RAW 264.7 cells showed the highest particle uptake, followed by IC-21 cells and MH-S cells. This trend was observed for both the M1 and M2 polarization conditions as well. Wang et al. reported similar observation in primary cells, where BMMs showed increased uptake compared to PMs, highlighting phagocytic similarities between primary cells and immortalized cell lines (Wang et al., 2013). Quantitative analysis of % PC + ve cells further confirmed the uptake profile (Figure 5B). For all the experimental conditions, MH-S cells (∼30% UT, ∼34% IFNγ, ∼34% IL4/13) showed the lowest uptake, followed by IC-21 (∼37% UT, ∼36% IFNγ, ∼41% IL4/13) and RAW 264.7 cells (∼45% UT, ∼45% IFNγ, ∼45% IL4/13). Furthermore, stimuli-related effects on phagocytosis were cell-line dependent. While RAW 264.7 cells did not show changes in phagocytosis on polarization, phagocytosis by both MH-S and IC-21 cells was sensitive to stimuli treatment in different ways. For MH-S cells, the polarization with both M1 and M2 stimuli increased the uptake compared to untreated cells; however, for IC-21 cells, only M2 stimuli increased the uptake, again highlighting the impact of tissue origin differences.

We also evaluated uptake via the normalized MFI (Figure 5C), where a higher normalized MFI value indicates a higher number of average internalized particles. For MH-S cells, we observed a significantly higher normalized MFI with IFNγ stimulation, while the normalized MFI was significantly reduced with IL4/13 stimulation compared to the untreated condition. A similar trend was observed for RAW 264.7 cells but was non-significant. This suggests that, for different cell lines, the polarization of macrophages not only impacts how many cells uptake particles but also influences how active individual cells are in their ability to internalize multiple particles.

Overall, we highlight that different cell lines have different propensities to internalize particles, which are further influenced by the polarization stimuli. Both the proportion of phagocytic cells and normalized MFI yield differential insights and are important considerations in investigating the phagocytic capacity of the macrophages.

We next investigated the relative repolarization capacity of different cell lines. Pre-polarized cells were treated with either IFNγ or IL4/13 (dose 2), and the phenotype of the repolarized macrophages after 24 h with dose 2 was quantified using flow cytometry. Repolarization of macrophages was determined by quantifying the PI for each marker, which normalizes the change in expression of final phenotype to the initial phenotype (Figure 6). A negative PI value indicates: 1) a decrease in M1 marker expression for (M1 to M2) repolarized macrophages compared to M1 macrophages and 2) a decrease in M2 marker expression for (M2 to M1) repolarized macrophages compared to M2 macrophages. In contrast, a positive PI value indicates: 1) an increase in the M1 marker expression for the (M2 to M1) repolarized macrophages compared to the M2 macrophages and 2) an increase in the M2 marker expression for (M1 to M2) repolarized macrophages compared to M1 macrophages. We compared PI for (M1 to M2) polarization (black bars) with (M2 to M1) polarization (pink bars). If the magnitude of the PI for (M1 to M2) polarization is higher than (M2 to M1) polarization, then it is easier to polarize cells to M2 phenotype, and vice versa. Similar values of the PI highlight that the cells can be repolarized to either stimulus effectively.

For example, for RAW 264.7 cells, PI for MHCII (M1 marker) was ∼ (−) 12.07 for (M1 to M2) repolarized macrophages and ∼ (+) 2.74 for (M2 to M1) repolarized macrophages; similarly, for IC-21 cells, PI for EGR2 (M2 marker) was ∼ (+)17.04 for M1 to M2 repolarized macrophages and ∼ (−)4.09 for M2 to M1 repolarized macrophages, highlighting changes in the individual marker expression upon repolarization (Figure 6). Overall, for MH-S cells, (M1 to M2) polarized cells had a higher PI for EGR2; (M2 to M1) polarized cells had higher PI for CD86. At the same time, both MHCII and CD206 showed similar PI values, indicating a similar level of plasticity. For IC-21 cells, (M1 to M2) polarized cells had a higher PI for MHCII and EGR2, while (M2 to M1) polarized cells had a higher PI for CD86 and CD206. For RAW 264.7 cells, (M1 to M2) polarized cells had a higher PI for CD86, MHCII, and EGR2. In contrast, CD206 had a higher PI for (M2 to M1) polarized cells.

These results suggest that not all macrophage cell lines respond similarly to repolarization. For example, for RAW 264.7 cells, it is easier to repolarize M1 polarized cells to M2, whereas the M2 polarized cells dampen the repolarization response to M1 stimuli. This observation is in-line with their strain background of Th2, indicating a preference towards M2-phenotypes. On the other hand, MH-S cells were more plastic and responded similarly to M1 and M2 repolarization. IC-21 cells showed a relatively higher EGR2 PI than other markers, suggesting an increased propensity to repolarize towards the M2 phenotype. Previous literature, to our knowledge, has investigated either M1 or M2 repolarization of macrophages, which limits our understanding of macrophage plasticity to either phenotype. For example, PMs have been shown to repolarize to either M1 or M2 macrophage phenotype in the presence of different therapeutic modalities (Farajzadeh et al., 2018; Moradi-Chaleshtori et al., 2021). However, to determine plasticity, the repolarization of macrophages should be investigated using both M1 and M2 stimuli simultaneously. This was previously demonstrated by Smith et al. where BMMs increasingly repolarize from the M1 to M2 phenotype; however, the M2 phenotype inhibits repolarization to the M1 phenotype (Smith et al., 2016). These observations, along with our results, are important for immune-engineering applications reliant on repolarization paradigms (Jarai et al., 2023).

Analogous to Figures 2, 3, we next assessed the phenotypical profile and morphology of primary macrophages extracted from mice of relevant genetic background and cultured ex vivo. As mentioned, macrophage phenotypes vary among strains; therefore, we assessed BMMs from both BALB/c and C57BL/6 mice. For both types of BMMs, M1 stimuli of IFNγ significantly upregulated both CD86 (Figure 7A) and MHCII (Figure 7B) marker expression. Notably, C57BL/6 BMMs had higher upregulation of both M1 markers (MHCII and CD86), which is consistent with previous literature that shows macrophages from C57BL/6 mice are more responsive to inflammatory stimuli (Santos et al., 2006). While M2 stimuli IL4/13 significantly upregulated CD206 (Figure 7C) for both strains, IL4/13 upregulated EGR2 for only BMMs from BALB/c mice, not C57BL/6 mice (Figure 7D), again consistent with the genetic background. Further, CD206 was more strongly increased in BALB/c BMMs compared to C57BL/6 BMMs. These data further align with BALB/c being a predominantly Th2 strain (Watanabe et al., 2004; Santos et al., 2006; Jones et al., 2013). This point is further illustrated by calculating the ratio of M1 and M2 markers (Supplementary Figure S5). C57BL/6 BMMs exhibited higher M1/M2 when stimulated with IFNγ. BALB/c BMMs exhibited higher M2/M1 when stimulated with IL4/IL13. For PMs, IFNγ significantly upregulated MHCII (Figure 7B) and CD206 (Figure 7C), while IL4/13 significantly upregulated MHCII (Figure 7B) and EGR2 (Figure 7D). Unlike BMMs from the same strain (C57BL/6), IL4/13 strongly upregulated EGR2 in PMs but not CD206. Upregulation of MHCII holds across all PMs and BMMs tested in response to IFNγ, while only PMs did show not an increase in CD86 when treated with either M1 or M2 stimuli.

Comparing cell lines to their primary cell counterparts (Figures 3, 7), some differences in the marker expression in the primary cells were observed: these include 1) IL4/13 stimulation upregulated EGR2 marker expression in BMMs (BALB/c) and PMs, whereas no change in the marker expression was observed for the cell line counterparts (RAW 264.7 and IC-21); 2) for IFNγ stimulation, the MHCII expression was similar for both RAW 264.7 and BMMs (BALB/c), whereas PMs showed a much higher increase in the fold change (∼44.1X) compared to IC-21 cells (∼1.2X). We hypothesize this is due to the transcriptional and genetic changes (Ben-David et al., 2018) introduced by the immortalization of primary cells, which has previously been reported for multiple cell types including macrophages (Levenson et al., 2018) and endothelial cells (Deng et al., 2020).

Next, we assessed the morphological changes of primary cells under both M1 and M2 stimulation. BMMs from both strains and PMs showed similar cell area (Figure 7E), and circularity (Figure 7F), and elongated morphology (Figure 7G), irrespective of the stimuli. Only BALB/c BMMs showed differences in cell area due to stimuli; IFNγ treated cells had lower cell area compared to the other conditions. Similar morphology was observed for both RAW 264.7 and IC-21 cell lines compared to BMMs (BALB/c) and PMs (C57BL/6), respectively, suggesting both primary cells and cell lines show similar morphology under different activation conditions. Non-significant differences in the morphology are again attributed to the hyper-physiological stiffness of the TCP culture substrate, which promotes a more elongated morphology. Overall, our results indicate that notable phenotypical differences exist in the primary macrophages obtained from different tissues and between the cell lines and the primary cells from the same origin.

Although 2D cell culture on TCP is the standard of practice, cell function and fate are dictated by a range of biochemical and biophysical cues in their multi-dimensional microenvironments. Bioinspired 3D culture models provide opportunities for understanding immune responses in more physiologically relevant environments that are specific to the tissue of interest. Among the cell types analyzed above in 2D culture, the most widely used immortalized (RAW 264.7) and primary (C57BL/6) murine bone-marrow macrophages were chosen for culture in three dimensions for probing similarities and differences in responses of immortalized and primary cells in 3D culture. Here, we used the RASTRUMTM bioprinter to create well-defined 3D cultures in a multi-well plate format (Figure 8A), providing rigor, reproducibility, and accessibility and enabling facile imaging and cell harvesting for downstream assays. We selected a bioprinted synthetic ECM composition (1.1-kPa cell-degradable PEG-peptide hydrogel with RGD, GFOGER, YIGSR, and HA) with properties inspired by tissues from which macrophages were derived (e.g., stiffness and ligands from ECM proteins [fibronectin, collagen, laminin] relevant to lung, peritoneal cavity) (Yen et al., 1997; Bomb et al., 2022; Wodzanowski et al., 2022).

Cell viability up to 7 days was confirmed using Live/Dead™ staining. The results suggest that cells maintain viability >70% after 7 days (Figure 8), showing that RASTRUM™ printed hydrogels are suitable for 3D culture of both immortalized and primary macrophages. As expected, RAW 264.7 cells exhibited increased cell number and metabolic activity that are indicative of proliferation (Supplementary Figure S6). As seen in Figure 8B, RAW 264.7 cells also formed large clusters over time, consistent with reports of their behavior in other 3D applications (Kim et al., 2019; Fang et al., 2020). However, cell clustering and proliferation may make RAW 264.7 cells undesirable for future 3D culture applications, limiting single cell analysis. Notably, unlike RAW 264.7 cells, BMMs remained mainly single cells (Figure 8C), with greater uniformity of cells throughout the hydrogel. Lack of cell clustering with BMM cultures is likely attributed to their limited proliferation, as expected and supported by the total cell counts and metabolic data over the 7-day period (Supplementary Figure S6). To the best of our knowledge, these results represent the first use of the RASTRUM™ bioprinter for 3D culture of macrophages, presenting future opportunities for tuning of cell-matrix and cell-cell interactions using the features of this platform (Du et al., 2022; Engel et al., 2022; Jung et al., 2022; Sullivan et al., 2023).

To probe any differences between immortalized and primary macrophage responses now in 3D culture, we examined the polarization of RAW 264.7 cells and BMMs encapsulated within the synthetic ECM upon the application of soluble stimuli. Analogous to Figure 7, M1 (CD86/MHCII) and M2 markers (CD206/EGR2) were quantified using flow cytometry after 24-h polarization with M1 stimuli (IFNγ) or M2 stimuli (IL4/IL13). IFNγ significantly upregulated M1 markers (CD86/MHCII) for RAW 264.7 cells (Figures 9A, B). Although not statistically significant, BMMs treated with IFNγ also showed slight upregulation of CD86 and MHCII compared to IL4/13 treated cells (Figures 9A, B). Neither cell line showed significant upregulation of M2 markers (Figures 9C, D).

Analogous to analyses on TCP, we assessed morphological differences (area and cell circularity) based on stimuli (IFNγ and IL4/13) (Figure 9E). Only BMMs were quantitatively analyzed for morphology; single cell statistics cannot be determined accurately with large cell clusters, and therefore, RAW 264.7 cells were not analyzed for volume and sphericity. We observed similar mean cell areas (350–400 μm²) and circularities (∼0.8) of BMMs across conditions (Figure 9F, G). Similarities across treatment groups are expected based on flow cytometry results, which suggest that polarization across conditions was insignificant. We observed the largest ranges of circularity and area of cells in the untreated condition, with lower circularity and larger cell areas in the untreated condition compared to the other conditions. Unlike on TCP, we observed mainly rounded morphologies, which is expected for macrophages encapsulated within soft materials (Cha et al., 2017; Chen et al., 2020; Bomb et al., 2022; Wodzanowski et al., 2022).

Compared to TCP, macrophages cultured in hydrogels were less responsive to external stimuli (IFNγ and IL14/IL13) when compared to the untreated condition for the respective cell type. This observation aligns with previous literature. For example, Lee et al. found that RAW 264.7 cells cultured in 3D hydrogels were less responsive to stimuli (LPS) (Lee and Ki, 2020). The biophysical properties of the microenvironment (e.g., matrix stiffness, density, degradability, structure) as well as biochemical cues (e.g., ligand type, density) all influence cellular responses (Bomb et al., 2022). We hypothesize that the lack of significant upregulation of M1 or M2 markers relative to TCP is due to culture in compliant materials relative to culture on rigid TCP. Soft materials have been shown to activate inflammatory responses in macrophages (Cha et al., 2017; Chen et al., 2020), where Cha et al. showed higher levels of CD86 expression in PEG diacrylate (PEGDA) hydrogels compared to gelatin methacrylate hydrogels (Cha et al., 2017). Kim et al. also noted that the 3D environment itself can induce inflammation compared to TCP (Kim et al., 2019). Further, Lee et al. found that RAW 264.7 cell response to an immunomodulatory polysaccharide was less in 3D culture in PEG-based hydrogels relative to 2D culture on TCP, potentially owing to the low availability of receptors in 3D culture like in tissues (Lee and Ki, 2020).

Hydrogel compliance also plays an important role in polarization; however, exactly how the modulus correlates to polarization remains unclear based on the current literature. Some researchers have noted that cells seeded on top of hydrogels display a more M2-like phenotype with stiffer gels (Chen et al., 2020; Bomb et al., 2022) while others have found that increased stiffness promotes M1-like phenotype (Sridharan et al., 2019; Bu et al., 2022). This trend can change based on the range of stiffnesses tested and the chemistry of the hydrogel. Further, 3D cultures and 2D hydrogel culture trends do not always align. Several researchers noted that larger pores size and lower modulus promote more M2-like phenotypes (He et al., 2018; Bu et al., 2022; Cicuéndez Maroto et al., 2024). Others also have reported that RAW 264.7 cells in PEG-based hydrogels increase baseline expression of both M1 and M2 markers in 3D culture in both compliant and stiff cell-degradable PEG-peptide hydrogels relative to 2D culture on TCP (Kim et al., 2019). Future work is needed to address how stiffness impacts the phenotype in such well-defined 3D cultures alongside synergistic effects of stiffness and external stimuli.

Based on previous literature as noted above (Cha et al., 2017; Kim et al., 2019; Lee and Ki, 2020), we speculate that higher base level expression of inflammatory markers prior to stimulation may hinder subsequent polarization with external stimuli when compared to the untreated control. There may also be less availability of receptors in 3D culture for binding to applied ligands or nonspecific interactions between the hydrogel components and cytokines that impact ligand-receptor binding. Note, the diffusion of cytokines through the full hydrogel thickness is expected to occur within an hour, based on the literature and assuming one-dimensional Fickian diffusion (t_d = L²/D, where the length (L) is 0.055 cm and the diffusion coefficient (D) of the proteins through a PEG-hydrogel is estimated to be ∼1E-6 cm²/s) (Weber et al., 2009). In this context, we do not expect cytokine diffusion to be limiting the macrophage response in these studies. Macrophages may need higher concentrations of cytokines and longer incubation times, or repeated exposures to see pronounced responses to stimuli. Opportunities for future studies include probing tissue-specific responses within such culture systems for their validation and application. Overall, our results establish the relevance of a bioprinted synthetic ECM for well-defined 3D culture of both immortalized and primary macrophages and provide insights into differences in their function in 3D culture toward informing cell type selection in both fundamental and applied studies, from biological mechanisms to biomaterials designs.