Macrophage Polarization Profiling in Dynamic Culture System

Abstract

Purpose

In this study, we aimed to develop a dynamic on-chip platform to study macrophage polarization in a more physiologically relevant way by incorporating mechanical forces which have been recently shown to play important roles in macrophage biology.

Methods

We developed polymethyl methacrylate (PMMA) based platform. We examined the effects of the dynamic microenvironment on polarization states of human monocyte derived macrophages (HMDMs) towards the M1 and M2a phenotypes using lipopolysaccharide (LPS)/interferon-γ (IFN-γ) and interleukin-4 (IL-4) respectively for both static and dynamic conditions. M1 and M2 polarization levels were assessed by qPCR and flow cytometry analyses.

Results

M1 and M2 polarization was achieved successfully under dynamic and static conditions. Our platform establishes that the mechanotransductive stimulation through shear stress during polarization has direct synergistic effects with stimulants on TNF-α secretion within HMDMs. Exposure to media flow rates of 0.5, 2.5, and 5 µl/min without stimulants is insufficient to induce macrophage polarization.

Conclusion

The dynamic environment present inside our dynamic on-chip culture platform influences the human monocyte-derived macrophages (HMDMs) to become polarized into M1 phenotype at a greater level.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12195-025-00863-0.

Introduction

Microfluidic platforms are being utilized more frequently in biomedical research due to their advantages over conventional cell culture techniques [1]. These include capacity of efficiently recapitulating cellular microenvironment of native tissue through mechanical forces and high spatiotemporal control over the cellular microenvironment [2, 3]. Furthermore, employing bioengineering approaches that can faithfully mimic physiological events can reduce the number of laboratory animal use which is in line with the 3Rs (replacement, reduction and refinement) principle.

Emerging microfluidic chip systems constitute cheap, high-throughput platforms for modeling diseases and drug screening applications [4] with versatile cell types being integrated such as endothelial, epithelial [5], tumor [6] and immune cells [7]. Microenvironments generated within the microfluidic platforms can be bioengineered to provide diverse mechanical stimulation for both 2D and 3D cell culture applications. Implementation of cellular mechanotransduction in these systems is achieved through matrix stiffness, stretching or fluid shear stress [5, 8]. This makes dynamic culture systems especially attractive tools for studying various mechanosensitive cell types including numerous immune cells [9].

Macrophages are innate immune cells appearing in the earliest developmental stage. As they are localized throughout all tissues, macrophages play crucial roles in many processes from organ development to the regulation of inflammation [10, 11]. With a large receptor repertoire and high plasticity, macrophages respond to a wide variety of stimuli in the changing microenvironment [12]. Owing to their plasticity, macrophages can adopt pro-inflammatory (M1) or anti-inflammatory (M2) features depending on the signal they receive from the environment [13, 14].

While the focus has been on the biochemical aspect traditionally, the mechanical stimuli’s effect on the macrophage responses has only been visited much more recently [15]. It has been revealed that macrophages alter their function with mechanical cues such as stiffness, topography, stretch and flow [16–21]. Numerous studies have shown that there are different mechanosensitive ion channel/receptors involved in receiving and processing mechanical signals from the environment in crucial processes such as polarization, and the formation of multinucleated giant cell formation [20–24]. Determining how the mechanical factors influence macrophage polarization will provide a clearer insight into macrophage biology for various tissues in health and disease. Better in vitro disease models and clinical approaches can be developed with this improved understanding. Therefore, developing in vitro platforms that incorporate not only tissue-specific biochemical signals but also mechanical stimuli for studying macrophage polarization is crucial [25]. In the last decade, there has also been an increased interest in studying macrophage behavior in different physiological contexts via utilizing dynamic on-chip culture platforms [23, 26]. In 2018, Li et al., investigated the effects of interstitial fluid flow (IF) on macrophage polarization in the tumor microenvironment (TME) by utilization of 3D biomimetic models. They reported that the interstitial flow (~3 µm/s) drives macrophages toward an M2-like phenotype which is associated with tumor progression. Jui et al., (2024), investigated the effects of shear stress of 15 and 35 dyn/cm² on human monocyte-derived macrophages (HMDMs) with a cone-and-plate viscometer over 3, 24, and 48 h. They found upregulation of pro-inflammatory markers (TNF-α, IL-8, IL-18, fractalkine) in a time-dependent manner, with the most significant increase observed at 24 h. Son et al., (2023) investigated how shear stress (12 dyne/cm²) influences macrophage-mediated inflammation, focusing on heat shock proteins (HSPs) as key regulators by using THP-1. They reported shear stress promotes macrophage polarization toward a pro-inflammatory (M1-like) phenotype.

In this study, we have developed a dynamic on-chip culture platform to study the polarization of human primary macrophages that incorporate both biochemical signals and mechanical stimuli through shear stress caused by the constant media flow in the channels (Fig. 1). We examined the effects of the dynamic microenvironment on polarization towards the M1 and M2a phenotypes using lipopolysaccharide (LPS)/interferon-γ (IFN-γ) and interleukin-4 (IL-4) respectively.

Fig. 1.

Schematical representation of monocyte isolation from buffy coat and experimental design of static and dynamic on-chip culture conditions with stimulants

Materials and Methods

Generation of Primary Human Monocyte-Derived Macrophages

Human monocyte derived macrophages (HMDMs) were generated using healthy human buffy coat under the ethical approval of Non-invasive Research Ethical Committee of Izmir Biomedicine and Genome Center, approval number 2022-006 according to the previously published protocol [27]. Briefly, peripheral blood mononuclear cells were isolated with Ficoll-Paque PLUS (GE Healthcare, Pittsburgh, PA) gradient. Then monocytes were separated from lymphocytes using Percoll (GE Healthcare, Pittsburgh, PA) gradient. Further, monocyte enriched layer was cultured in ultra-low attachment 6 well plate using macrophage derivation media, R5, composed of RPMI 1640 (Gibco, ThermoFisher Scientific, Waltham, MA, USA), 5% heat-inactivated fetal bovine serum (FBS) (Gibco, ThermoFisher Scientific, Waltham, MA, USA), 1% penicillin/streptomycin (Gibco, ThermoFisher Scientific, Waltham, MA, USA) supplemented with 10 ng/ml human macrophage colony-stimulating factor (M-CSF) (PeproTech, Rocky Hill, NJ). Cells were cultured by addition of fresh 2 ml macrophage derivation media once every other day for a week to generate HMDMs. The macrophages were verified to be over 90% CD68⁺ by flow cytometry analysis.

Fabrication of Dynamic on-Chip Culture Platform

The platform was fabricated using polymethyl methacrylate (PMMA) sheets and double-sided adhesive film (3 M 468MP, USA) (Fig. 2A). The channels (3.0 mm in depth) were bonded to tissue culture treated polystyrene petri dishes to assist cell adhesion and topped with a PMMA cover. Channels were perfused through inlet and outlet ports with the appropriate tubing (ND-100-80, Tygon®, USA). The chips were sterilized using 70 % ethanol followed by 1 h exposure to UV light under a laminar flow hood. Static culture conditions were carried out in the custom designed PMMA wells, holding culture media volume equivalent to the amount perfused in the platform with dynamic on-chip culture conditions (Suppl. Fig. 1F).

Fig. 2.

COMSOL simulations of designed dynamic on-chip culture platform and monocyte/macrophage analysis. A PMMA layers of dynamic on-chip culture platform with dimensions and image of platform with inserted inlet and outlet tubings (bottom), Computational fluid dynamics (CFD) simulation of the dynamic platform at a flow rate of 0.5 µl/min, B Velocity magnitude (m/s) mapped across cross-sections, with velocity vectors (red arrows) indicating flow direction and magnitude. Higher velocities are observed near the inlet and along the central channel, C Velocity magnitude (mm/s) at the base of the device, D Fluid shear stress distribution (dyne/cm²) across cross-sections, showing elevated shear near the inlet and at constrictions, E Fluid shear stress (dyne/cm²) at the device base, F CD68 expressions of monocyte-enriched layer and HMDMs, G Calcein (green) and PI (red) stainings of HMDMs after 24 h of 0.5 µl/min flow rate in dynamic on-chip culture platform. Scale bar = 200 µm

Computational Fluid Dynamics Analysis

To assess the flow velocity and hydrodynamic shear stress produced in the system, the channel was designed in AutoCAD (Autodesk, UK) in 3D and then exported in .stl format. Later, the .stl file was imported into COMSOL Multiphysics (Comsol, USA). Fluid dynamics analysis of the device was modelled with the Navier-Stokes equations of continuity (eq.1) and momentum conservation (eq.2) for incompressible Newtonian fluids. Experimental flow rates of 0.5, 2.5 and 5 µl/min were imposed on the channel inlet whereas a null pressure was defined at the channel outlet. No slip wall condition was imposed on the channel walls and the 3D geometry meshed using adaptive mesh refinement.

Incompressible Navier-Stokes equation for Newtonian fluid:

$$\frac{\partial \mathrm{\rho }}{\partial t}+\mathrm{\nabla }.\left(\rho ,\mathit{u}\right)=0$$ 1

$$\mathrm{\rho }\frac{\partial \mathit{u}}{\partial t}+\mathrm{\rho }\left(\mathit{u},.,\mathrm{\nabla }\right)\mathit{u}=\mathrm{\nabla }.\left[-,p,\mathbf{I},+,\mathbf{K}\right]+F$$ 2

u: velocity vector (m/s); ρ: fluid density (kg/m³); p: fluid pressure (Pa); F: volume force vector (N/m³); K: viscous stress tensor (Pa).

The shear stress was calculated from:

$$\tau =\mu \frac{\partial \mathit{u}}{\partial y}=\mu \dot{\gamma }$$ 3

where, $\tau$ is the wall shear stress, $\dot{\gamma }$ is the shear rate, $\mu$ is the fluid viscosity, and u is the velocity.

Culture and Polarization of HMDMs in Dynamic on-Chip Culture Platform

Macrophages were seeded at a density of 7x10⁵ cell/channel in each dynamic on-chipculture platform and incubated overnight to facilitate the cell adhesion. Channels were perfused with the R5 media at 0.5 µl/min (or as otherwise noted) with a syringe pump (Harvard Apparatus PHD ULTRA, USA). For static conditions, cells received media with the same volume as the dynamic on-chip culture platform (Suppl. Fig 1H). M1 macrophage polarization was achieved by culturing the cells in the R5 media supplemented with Ultrapure LPS (100 ng/ml) (InvivoGen, San Diego, CA) and human IFN-γ (20 ng/ml) (R&D, Minneapolis, MN). M2a macrophage polarization was achieved by culturing cells with the R5 media supplemented with human IL-4 (20 ng/ml) (R&D, Minneapolis, MN). The polarization experiments lasted 6 and 24 h for mRNA and protein analysis respectively based on a previously published report [27].

Live/Dead Staining

Calcein AM and propidium iodide (PI) were used to detect live and dead cells. After washing with PBS, the dynamic on-chip culture platforms were incubated for 30 min with 1 µM Calcein AM (Biolegend, USA). Following that, the calcein solution was removed and washed again with PBS. PI (Biotium, USA) solution at a concentration of 2 µg/ml was then added and incubated for 2 min. The chips were washed with PBS before being taken to imaging.

RNA Isolation and qPCR

qPCR analysis was performed to determine the M1 and M2a gene expression levels. For M1 polarization, TNF-α, CXCL10, and IL-1β were determined. MRC1 and TGM2 were chosen as M2a markers. Total RNA was isolated using the Nucleo-Spin RNA kit (Macharey Nagel, Germany). The concentration and purity of RNA samples were measured using the NanoDrop Spectrophotometer (Thermo Scientific, Rockford, IL). cDNA conversion was performed using OneScript Plus cDNA Synthesis Kit (Applied Biological Material Inc., Canada). Quantitative PCR (qPCR) was performed with a LightCycler 480 II real time system (Roche, Switzerland) at a final volume of 10 µl containing 5 µl of 2X Mastermix, 2 µl of primers (10 µM), 2 µl of nuclease-free water, and 1 µl of cDNA (5 µg/µl) in accordance with the manufacturer’s instructions. Primer sequences for M1 and M2 macrophage polarization markers are given in Table 1. GAPDH was chosen as the reference gene. The relative quantification of gene expression was calculated by the 2^−ΔΔCT method.

Table 1.

qPCR primer sequences

Gene	Forward	Reverse
CXCL10	GTGGCATTCAAGGAGTACCTC	TGATGGCCTTCGATTCTGGATT
TGM2	CGTGACCAACTACAACTCGG	CATCCACGACTCCACCCAG
IL1b	ATGATGGCTTATTACAGTGGCAA	GTCGGAGATTCGTAGCTGGA
GAPDH	GTCTCCTCTGACTTCAACAGCG	ACCACCCTGTTGCTGTAGCCAA
TNFa	GAGGCCAAGCCCTGGTATG	CGGGCCGATTGATCTCAGC
MRC1	CTACAAGGGATCGGGTTTATGGA	TTGGCATTGCCTAGTAGCGTA

Flow Cytometry Analysis

After discarding the supernatant, cells were harvested using StemPro Accutase Cell Dissociation Reagent (Gibco, Thermo Fisher Scientific, USA). To determine the cell viability, the Zombie UV Fixable Viability Kit (BioLegend, San Diego, CA, USA) was utilized. Then, the cells were incubated with FACS buffer (1% bovine serum albumin and 0.1% sodium azide in 1X PBS) containing Human TruStain FcX antibody (Fc block) (BioLegend, San Diego, CA).

For macrophage characterization, cells were fixed and permeabilized using CytoFix/CytoPerm Buffer (BD Biosciences, USA) according to the manufacturer’s instructions, following the extracellular staining with CD20-PE/Cy7 (2H7; Tonbo Biosciences, USA), CD3-eF660 (OKT3; ThermoFisher/eBio, USA), CD14-V500 (M5E2; BD Biosciences, USA), CD16-PE/CF594 (3G8; BD Biosciences, USA) on ice for 45 min. After washing with Perm/Wash solution (BD Biosciences, USA), the cells were incubated with the human pan macrophage marker CD68 antibody (Biolegend, USA) for 45 min at room temperature.

For the M1/M2a polarization analysis, the HMDMs were incubated for 45 min on ice with the following antibodies targeting specific surface markers: CD86-BV605 (IT2.2; Biolegend, USA), HLA DR-APC/Cy7 (L243; Biolegend, USA) are associated with M1 phenotype, while CD206- AF700 (15-2; Biolegend, USA) is associated with M2a phenotype. The acquisition was performed with LSRFortessa (BD Biosciences, USA) and analyzed with FlowJo software (TreeStar, USA).

Enzyme-Linked Immunosorbent Assay (ELISA) Quantification

For the enzyme-linked immunosorbent assay of TNF-α levels for the untreated and M1 polarized HMDMs, conditioned media were collected after 24 h of incubation for static and dynamic conditions. ELISA Max Deluxe Set Human TNF-α (Biolegend, USA) was used for the detection of TNF-α levels in the media according to the manufacturer’s instructions.

Statistical Analyses

Data is expressed as mean ± standard error of the mean (SEM) of at least two different experiments performed in triplicates. The two-tailed student t-test was used for the analysis of the data. To remove potential batch effects between qPCR measurements we applied log-normalization to 2^–∆∆Ct values. Both statistical analyses and graph drawing were performed using GraphPad Prism 8 (GraphPad Software, CA, USA). p values < 0.05 were considered statistically significant and are indicated with *p < 0.05, **p ≤ 0.005, ***p ≤ 0.0001.

Results

Fabrication of Dynamic on-Chip Culture Platform and Computational Fluid Dynamics Analysis

Dynamic on-chip culture platform was fabricated with the 2 layers PMMA sheets and petri dish, which were bound with double sided adhesive tape precisely cut with a laser cutter (Fig. 2A). Top PMMA layer consists of inlet and outlet ports with 0.75 mm openings for the insertion of the tubing (Suppl. Fig. 1A). The middle part constitutes the channel with 18 × 4 × 3 mm dimensions for the media flow (Fig. 2A). Shear based cellular mechanotransduction was provided through the media flow. Simulations were performed for the determination of the flow profile inside the platform channel for the flow rate of 0.5 µl/min (Fig. 2B–E), 2.5 µl/min (Suppl. Fig. 1B–C) and 5 µl/min (Suppl. Fig. 1D–E). The simulations demonstrated that the applied media flow has uniform distribution throughout the dynamic on-chip culture platform channel even at higher flow rates (Suppl. Fig. 1B–E). Flow induced velocity magnitude at the microchannel base where cells reside ranged between 2–20 × 10⁻⁴ mm/s and Reynold’s Number Reynolds number distribution throughout the dynamic on-chip culture platform channel for 0.5 µl/min volumetric flow rate range between 8 × 10⁵–3 × 10⁸ (Suppl. Fig. 1F–G). Live/dead assay was performed to examine the viability of HMDMs cultured under continuous media flow for 24 h in the platform with a flow rate of 0.5 µl/min. Calcein AM (green)/Propidium Iodide (red) staining showed that bioengineered dynamic on-chip culture platform supports the viability of HMDMs (Fig. 2G).

Generation of Human Monocyte-Derived Macrophages

HMDMs and monocytes were characterized by flow cytometry. For monocyte-enriched layer; the monocyte population was gated according to forward- and side-scatter parameters (Suppl. Fig. 2A) [28]. Phenotypic characterization was confirmed by negativity of CD3 and CD20 (Suppl. Fig. 2B–C), and the different subpopulations of monocytes were determined with the expression level of CD14 and CD16 (Suppl. Fig. 2D). Macrophage differentiation was assessed through CD68 expression indicating a successful monocyte to macrophage differentiation (Fig. 2F).

M1 Polarization of Macrophages in Dynamic on-Chip Culture Platform

Macrophages were stimulated with LPS/IFN-γ for M1 polarization. M1 markers were assessed at mRNA level by qPCR after 6 h stimulation and at protein level by flow cytometry and ELISA after 24 h stimulation. When macrophages were stimulated with LPS/IFN-γ for 6 h, polarization was achieved under both static and dynamic conditions according to the mRNA levels of M1 marker genes relative to the GAPDH expression (Fig. 3A). When comparing M1 marker mRNA expression levels between static and dynamic conditions, normalized to the static M1 group, we did not observe a significant change between groups (Fig. 3B).

Fig. 3.

Analyses of M1 stimulated and untreated macrophages cultured under static and dynamic conditions. A Heatmap of M1 polarization marker expression (TNF, CXCL10, IL1B) across individual chips (Ch1–Ch4) under static and dynamic conditions with or without LPS + IFNγ stimulation. Color intensity represents relative expression levels. P numbers represent different human donors (biological replicates), Ch numbers represent dynamic on-chip culture platforms (technical replicates from the same donor). B Comparison of M1 marker mRNA expression between static and dynamic conditions following LPS + IFNγ stimulation, normalized to the static M1 group. C–D Flow cytometry analysis of M1 surface marker expression (HLA-DR and CD86) within macrophages under static and dynamic conditions with or without LPS + IFNγ stimulation. Representative histograms showing individual chips (Ch1–Ch3) under each condition (C), and quantification of mean fluorescence intensity (MFI) for HLA-DR and CD86 (D). E TNF-α concentrations in static and dynamically cultured untreated and LPS + IFNγ-stimulated macrophages. Each dot represents an individual chip. Data are representative of two independent experiments from different donors (C–E). Data are presented as mean ± SEM, with statistical significance indicated (*p < 0.05, **p ≤ 0.005, ***p ≤ 0.0001)

Flow cytometry analysis showed that there were no significant changes between static and dynamic conditions in terms of HLA-DR and CD86 expression levels (Fig. 3C and D). Interestingly, the production of TNF-α was significantly enhanced when macrophages were polarized under the dynamic conditions (Fig. 3E).

M2 Polarization of Macrophages in Dynamic on-Chip Culture Platform

Macrophages were stimulated with IL-4 for M2a polarization. After 6 and 24 h stimulation, qPCR and flow cytometry analyses were performed, respectively. Our results showed that our dynamic on-chip culture platform enabled M2a polarization, as indicated by the upregulation of M2a marker genes MRC1 and TGM2 (Fig. 4A, Supp Fig. 3A–D). When comparing M2a marker mRNA expression levels between static and dynamic conditions, normalized to the static M2a group, we did not observe a significant difference between groups (Fig. 4B).

Fig. 4.

Analyses of M2 stimulated and untreated macrophages cultured under static and dynamic conditions. A Heatmap of M2 polarization marker expression (MRC1 and TGM2) across individual chips (Ch1–Ch4) under static and dynamic conditions with or without IL-4 stimulation. Color intensity represents relative expression levels. P numbers represent different human donors (biological replicates), Ch numbers represent dynamic on-chip culture platforms (technical replicates from the same donor). B Comparison of M2 marker mRNA expression between static and dynamic conditions following IL-4 stimulation, normalized to the static M2 group. C–D Flow cytometry analysis of M2 surface marker expression (CD206) within macrophages under static and dynamic conditions with or without IL-4 stimulation. Representative histograms showing individual chips (Ch1–Ch3) under each condition (C), and quantification of mean fluorescence intensity (MFI) for CD206 (D). Each dot represents an individual chip. Data are representative of two independent experiments from different donors (C–D). Data are presented as mean ± SEM, with statistical significance indicated (*p < 0.05, **p ≤ 0.005, ***p ≤ 0.0001)

Increased expression of the M2 surface marker CD206 confirmed polarization in response to the IL-4 stimulation under dynamic and static conditions (Fig. 4C–D). Between dynamically and statically cultured macrophages without any stimulation had no significant changes in CD206 levels (Fig. 4D).

Effect of Increased Flow Rates Without Polarization Factors

To further investigate the effect of mechanical forces on macrophage polarization, HMDMs were exposed to increased levels of flow rates of 0.5, 2.5, and 5 µl/min without any polarization factors. The M1 (TNF-α and CXCL10) and M2 (MRC1 and TGM2) marker gene expression levels did not change with the increased levels of flow rate (Fig. 5A–B). In line with these results, flow cytometry analyses showed that mechanical induction alone did not induce macrophage polarization (Fig. 5C–D). CD206 MFI exhibited a slight increase under the 5 µL/min flow rate compared to statically cultured macrophages; however, the biological relevance of this difference remains uncertain (Fig. 5D).

Fig. 5.

The influence of different flow rates (static, 0.5, 2.5, 5 µl/min) on macrophage polarization states. A–B Heatmap representation of mRNA levels for (A) M1 markers (TNF, CXCL10) and (B) M2 markers (MRC1, TGM2) in individual chips, normalized to GAPDH (2ΔCt). C Representative histograms of HLA-DR, CD86, and CD206 expression. D Median fluorescence intensity (MFI) values for M1 (HLA-DR, CD86) and M2 (CD206) surface marker expression. Each dot represents an individual chip. Data are presented as mean ± SEM and analyzed using the Kruskal-Wallis test (*p < 0.05, **p ≤ 0.005, ***p ≤ 0.0001). Results are representative of three independent experiments from two different donors. Each dynamic on-chip culture platform is labeled as ‘Ch’. M1 and M2 macrophages polarized with corresponding stimulants were used as controls (‘CTRL’)

Discussion

A more complete understanding of the different individual components of the immune system requires the generation of models of the healthy and diseased states. However, this is complicated by the fact that the available tools, such as immortal cell lines and animal models, differ greatly in terms of their response to experimental stimuli compared with the native microenvironments in humans. Moreover, the immune system consists of many different cell types throughout the body, constantly receiving feedback both chemically and mechanically. Therefore, an in-depth analysis of an immunological phenomenon requires the generation of models that incorporate more relevant cell types, such as using human primary macrophages, that would respond to the microenvironment more faithfully and mechanical cues which are often not represented in 2D cell culture systems. Such bioengineered tools are potent to promote better understanding of the role of immune mechanisms in developing new therapeutic approaches or drug screening applications [4].

The macrophage polarization behavior contributes to different physiological events across different tissues warranting a better understanding of macrophage biology. Anti-inflammatory M2 macrophages contribute to tumor progression such as glioblastoma, or M1 mediated responses are responsible for inflammatory diseases such as multiple sclerosis [29]. Therefore, modulation of macrophage polarization and studying the dysregulation of the M1 and M2 phenotypes carries high importance [30].

In addition to macrophages’ innate ability to calibrate their polarization according to the chemical stimuli in the microenvironment, there have been several new studies showing that their polarization is regulated by mechanical stimuli as well [31, 32], further highlighting the importance of incorporating mechanical forces into immunological models to get more physiologically relevant responses.

Mimicking the body’s immune response in dynamic on-chip culture platforms overcomes the poor predictability of the conventional static culture conditions, providing advanced tools for disease modeling, drug screening applications and drug response predictions [33–35]. Introduction of bioengineered on-chip platforms in immunology research has been reported demonstrating capacity of forming complex co-culture settings of macrophages with cancer and stromal cells [4, 30, 31, 36, 37].

In our study, we investigated the polarization of human monocyte-derived macrophages under dynamic conditions in designed dynamic on-chip culture platform (Fig. 1). Here we have used the primary human monocyte derived macrophages (HMDM) despite the laborious and time-consuming procedures in obtaining, generating, and storing. HMDMs are superior to macrophage cell lines (such as THP1) and mouse origin macrophages providing more relevant responses for better understanding of human macrophage polarization [38, 39].

First, we in silico evaluated the fluid flow profile within the channels of the designed dynamic on-chip culture platform (Fig. 2B–E). COMSOL simulations showed homogenous flow distribution inside the dynamic on-chip culture platform channel providing uniform reproducible mechanical stimuli to cells.

The experimental setup specifically has focused on providing comparable identical conditions for cells cultured both in static and dynamic settings. To achieve this, the total media volume used to culture HMDMs was kept the same for all conditions, preventing bias with larger amounts of stimulants (Suppl. Fig. 1H).

Later, the viability of adherent cells under dynamic conditions was validated with Calcein-PI staining (Fig 2G). Most of the stained cells were observed as green which indicates majority of our HMDM cell population stayed viable after 24 h of dynamic condition. We also observed that at a flow rate of 0.5 µl/min, cells indeed continue to adhere to the bottom of our dynamic on-chip culture platform even after 24 h of dynamic culture (Fig. 2G).

Various mechanical forces exacerbate disease progression during pathological conditions. For instance, abnormally stiff extracellular matrix and higher interstitial fluid shear stress contribute to tumor metastasis, while lower shear stress induces atherosclerotic plaque formation [40]. Macrophages, central to shaping immune responses, are directly influenced by tissue architecture due to their high mechanosensitivity [41]. Therefore, it is imperative to utilize a simple and versatile platform that would enable the induction of shear stress on the macrophages which would mimic the native microenvironment.

In this paper, we generated an dynamic on-chip culture platform that enabled the modulation of HMDMs towards either the M1 (inflammatory) or M2 (anti-inflammatory) phenotype with the incorporation of both biochemical and mechanical forces (Figs. 3 and 4). With respect to the surface marker expression and gene expression, there was no difference between static and dynamic conditions for the M1 and M2 polarization. However, our results showed that the effect of mechanical stimuli enhanced the M1 polarization in terms of TNF-α cytokine secretion response in cooperation with the LPS/IFNγ in the microenvironment (Fig. 3E). This suggests that mechanical forces act synergistically with the biochemical stimuli to determine the level of response similar to the previously published reports [20].

Recent studies suggest that both M1 and M2 macrophages are capable of sensing mechanical stimuli, but they do so via distinct mechanisms. For example, Meizlish et al. showed that macrophages can detect mechanical cues through integrin-independent, cytoskeleton-mediated pathways, which may become more prominent during the later stages of inflammation, when reparative (M2-like) macrophages are more common. At the same time, other studies have shown that lower shear stress levels, such as those seen in atherosclerosis, tend to favor M1-like polarization [42]. In our study, the applied shear stress was sub-physiological, and we believe it was below the threshold needed to activate polarization-related signaling in either direction.

The effect of shear stress on macrophage response has been highlighted as a sufficient mean to induce macrophage polarization previously [43–46]. Son et al., 2023 reported that they observed an increase in the M1 polarization markers even without the biochemical stimuli treatment [43]. However, their argument warrants further experiments with controls treated with proinflammatory factors due to their proposed levels of minuscule increase in TNF-α release. Fish et al. showed that only when providing very high levels of shear stress on cells embedded inside collagen type 1 hydrogel is sufficient for inducing inflammatory response. While we observed that it was not possible to direct macrophages towards the M1 phenotype only via increased shear stress levels, it is possible that this phenomenon only occurs when the cells are exposed to even higher shear stress levels than they are in our dynamic on-chip culture platform (Figure 5A–D).

In contrast to prior studies that investigated macrophage responses under physiological or supra-physiological levels of shear stress (typically ranging from 1 to 30 dyn/cm²) [47], our platform applies an ultra-low shear stress of approximately 3.5 × 10⁻⁵ dyn/cm², which is ~ 1000-fold lower than physiological interstitial or vascular flow. To our knowledge, the effect of such sub-physiological shear on macrophage polarization has not been previously characterized. Despite this minimal mechanical input, we observed a significant enhancement in TNF-α production under dynamic conditions, suggesting that even minute levels of mechanical stimulation may act as co-stimulatory cues in the presence of inflammatory signals. Importantly, the dynamic condition did not independently induce polarization toward M1 or M2 phenotypes, which aligns with the expectation that such low shear is insufficient to trigger transcriptional reprogramming. As to why TNF-α was selectively affected, it could be due to the low shear stress may not be sufficient to alter gene expression profiles related to M1 or M2 polarization. We also hypothesize that the increase in the TNF-α could be due to the continuous perfusion in our microfluidic system, which in turn, may reduce the local TNF-α accumulation near the macrophage surface, thereby dampening autocrine negative feedback, a mechanism known to suppress TNF-α production at high local concentrations. This altered cytokine dynamics could permit greater net TNF-α secretion under flow conditions.

Conclusion

In this study, we are reporting a perfusable dynamic on-chip culture platform that provides HMDMs a microenvironment with tunable mechanical stimuli. Our platform establishes that the mechanotransductive stimulation through shear stress during polarization have direct synergistic effects with stimulants on TNF-α production in HMDMs.

With the capability to manipulate the forces acting on cells in a spatiotemporally sensitive way, our platform can be used to decipher mechanical stimuli-related changes in macrophage function. Additionally, this platform can be modified to study immune-tissue reactions at mechanically dynamic sites of the body such as the intestine, arteries, and lungs, considering that macrophages are resident cells that can be found in all adult tissues [47]. As a future direction, further studies should investigate macrophage polarization over longer durations to determine whether dynamic culture conditions lead to sustained changes in macrophage behavior. A combinatorial approach in this manner can open new avenues for modeling and studying autoimmune diseases.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, Alperen Yılmaz, Resul Özbilgiç, Elifsu Polatlı, Sinan Güven; formal analysis, Alperen Yılmaz, Resul Özbilgiç, Elifsu Polatlı, İbrahim Halilullah Erbay, Sinan Güven; funding acquisition, Sinan Güven; investigation, Alperen Yılmaz, Resul Özbilgiç, Elifsu Polatlı, İbrahim Halilullah Erbay, Duygu Sağ, Sinan Güven; methodology, Alperen Yılmaz, Resul Özbilgiç, Elifsu Polatlı, İbrahim Halilullah Erbay, Duygu Sağ, Sinan Güven; writing-original draft, Alperen Yılmaz, Resul Özbilgiç, Elifsu Polatlı, Sinan Güven; project administration, Sinan Güven; resources Duygu Sağ, Sinan Güven; supervision Duygu Sağ, Sinan Güven; writing–review and editing Duygu Sağ, Sinan Güven. All authors contributed to the article and approved the submitted version.

Funding

This work was partially supported by TUBITAK 118S477 project.

Data availability

Data is available upon request.

Declarations

Competing interests

The authors declare no conflict of interest.

Ethical approval

Human monocyte derived macrophages (HMDMs) were generated using healthy human buffy coat under the ethical approval of Non-invasive Research Ethical Committee of Izmir Biomedicine and Genome Center, approval number 2022-006.

Consent to participate

Not applicable.

Consent to publish

Not applicable.