Single-cell assessment of the modulation of macrophage activation by ex vivo intervertebral discs using impedance cytometry

Abstract

Measurement of macrophage activation and its modulation for immune regulation is of great interest to arrest inflammatory responses associated with degeneration of intervertebral discs that cause chronic back pain, and with transplants that face immune rejection. Due to the phenotypic plasticity of macrophages that serve multiple immune functions, the net disease outcome is determined by a balance of subpopulations with competing functions, highlighting the need for single-cell methods to quantify heterogeneity in their activation phenotypes. However, since macrophage activation can follow several signaling pathways, cytometry after fluorescent staining of markers with antibodies does not often provide dose-dependent information on activation dynamics. We present high throughput single-cell impedance cytometry for multiparametric measurement of biophysical changes to individual macrophages for quantifying activation in a dose and duration dependent manner, without relying on a particular signaling pathway. Impedance phase metrics measured at two frequencies and the electrical diameter from impedance magnitude at lower frequencies are used in tandem to benchmark macrophage activation by degenerated discs against that from lipopolysaccharide stimulation at varying dose and duration levels, so that reversal of the activation state by curcumin can be ascertained. This label-free single-cell measurement method can form the basis for platforms to screen therapies for inflammation, thereby addressing the chronic problem of back pain.

1. Introduction

Effective immune systems require immune cells to perform diverse functions (Satija and Shalek, 2014), leading to heterogeneity in their cellular phenotypes. Macrophages are immune effector cells that display a high degree of phenotypic plasticity while serving multiple homeostatic functions associated with host defense, wound healing, and immune regulation, in response to activation under environmental cues (Mosser and Edwards, 2008). The spatiotemporal profiles of the stimuli on phenotypically plastic macrophages leads to heterogeneity in their phenotypes (Oyler-Yaniv et al., 2017), highlighting the need for single-cell techniques to measure macrophage activation (Spiller et al., 2010). Macrophages are the predominant cells driving inflammation in response to the degeneration of intervertebral discs (henceforth, IVD or discs) (Burke et al., 2002), as well as to biomaterial (Abebayehu et al., 2017) and cellular transplants (Langer, 2010). Emerging evidence (Jin et al., 2021) suggests that the balance of activation phenotypes of macrophage subpopulations infiltrating the degenerated disc determines outcomes, such as back and radicular pain. Back pain from disc degeneration is the most common cause of disability (Andersson, 1999), with a prevalence of over 80% in individuals between the ages of 20–50 (Hoy et al., 2014), for which no disease-modifying medications currently exist to address the long-term outcomes (Purmessur et al., 2013). Drugs designed to modulate the activation of macrophages can alleviate disc degeneration and the associated back pain, as well as immune rejection of transplant therapies. However, for in vitro screening of immunomodulation therapies, macrophage activation needs to be quantified with single-cell sensitivity, and in a dose and duration-dependent manner.

Flow cytometry after fluorescent staining of cell surface markers is commonly used for quantification of phenotypic responses at single-cell sensitivity. However, since phenotypes of macrophages are dynamic and stimulus dependent (Lu et al., 2015), (An et al., 2002), it is often not possible to capture the full spectrum of macrophage activation, using molecular markers from a particular signaling pathway (Silva et al., 2019; Vogel et al., 2014). Also, endpoint fluorescent staining assays are not compatible with longitudinal studies on activation dynamics for the same set of cells within the sample (Gordonov et al., 2016). Similarly, single-cell RNA sequencing, while valuable in identifying distinct macrophage states, is destructive and does not permit downstream studies on activated cells. Raman microspectroscopy for profiling lipid transformations is an emerging label-free technique to characterize macrophage activation with single-cell sensitivity (Feuerer et al., 2021), but it measures only limited cell numbers (~100 per sample). Label-free quantitative phase microscopy to characterize cell morphologies is especially of interest due to its ability to measure global changes in cellular phenotype to provide dose-dependent phenotypic information that does not rely on definitive markers of a particular signaling mechanism (Pavillon et al., 2018). However, due to its intensive computational needs for multivariate analysis of images, only limited cell numbers can be measured (~100–500 per sample) and the data analysis can only be performed off-line, thereby limiting the ability to select phenotypes for downstream studies.

Emerging cytometry methods using electrical, mechanical, and imaging modalities offer single-cell biophysical information for label-free stratification and cell selection based on phenotypes (Lee et al., 2021). Impedance cytometry is one such technique that can provide non-invasive, sensitive, and high throughput (300–400 events/s) biophysical information on cells based on disruptions to the magnitude and phase of alternating current flow by passing single cells or subcellular particles in a microchannel (Cheung et al., 2010), (Sun and Morgan, 2010), over several frequencies in the 0.5–50 MHz range. In this manner, biophysical phenotypes can be discerned based on cellular electrical size at low frequency (~0.5 MHz), or membrane features can be discerned based on membrane conductance and capacitance at mid-frequencies (2–10 MHz), and features in the cell interior can be followed at high frequencies (≥10 MHz) (Honrado et al., 2021). Using size-controlled co-flowing insulating beads for data normalization (Spencer and Morgan, 2020), impedance magnitude (|Z|) and phase (φZ) metrics can be quantified and compared across different biological samples over multiple frequencies, so that the dispersions can provide multiparametric information on cell phenotype (Moore et al., 2020) and drug-induced transformations (De Ninno et al., 2020), (Honrado et al., 2020; Spencer et al., 2020).

In this work, we present impedance metrics from single-cell cytometry for quantifying macrophage activation in a dose and duration-dependent manner, so that drug-induced modulation of these phenotypes can be ascertained by label-free biophysical measures for each cell, without relying on molecular markers from a particular signaling pathway. Using differing dose and duration levels of lipopolysaccharide (LPS) stimulation on macrophages (Raw 264.7) that are known to stimulate the pro-inflammatory Toll-like receptor 4 (TLR4) (An et al., 2002), macrophage activation determined by impedance cytometry is cross-validated based on imaging and secreted nitrite (NO) in media (Jacobs and Ignarro, 2001) for enabling comparison to the level elicited by degenerated discs. Impedance metrics are identified to discern the modulation of macrophage activation, including during its interaction with progressively degenerated discs, and after treatment with curcumin, an anti-inflammatory compound that can reverse activation (Zhou et al., 2015). Recent impedance cytometry work has correlated leukocyte activation to an increase in electrical diameter (Petchakup et al., 2021), but for macrophages that possess a highly migratory phenotype, intercellular interactions can limit the size increase and stress cells to create heterogeneity in their activation response (Nomura et al., 2000). Hence, using impedance phase metrics to detect the stressed subpopulation from intercellular interaction, we optimize macrophage cell density to enable facile formation of the polarized cell morphology under activation (Berghaus et al., 2010), (Taciak et al., 2018). Using multiple frequencies to aggregate the multivariate phenotypic alterations, based on electrical diameter, capacitance of the plasma membrane, and conductivity and permittivity of the cell interior, impedance metrics are identified to quantify macrophage activation by degenerated discs and its reversal by curcumin, with dose and duration dependence. We envision application of this single-cell biophysical cytometry method to screen for immunomodulation therapies that address the chronic problem of back pain.

2. Materials and methods

2.1. In vitro culture of macrophages

Macrophages Raw 264.7 cells from American Type Culture Collection (ATCC) (Manassas, VA) were cultured with Dulbecco’s modified Eagle’s medium (DMEM, high glucose 4.5 g/L, Gibco, Grand Island, NY) supplied with 10% fetal bovine serum (FBS) (Gibco), 1% penicillin (100 μg/mL), and 1% streptomycin (100 μg/mL) (Gibco) and maintained in 5% CO₂ at 37 °C. The cell seeding density was optimized using a 24 well plate including: 5 × 10⁵, 2 × 10⁵, and 1 × 10⁵ cells/mL, with 0.5 mL per well, to yield 2 × 10⁶, 1 × 10⁶, and 2 × 10⁵ cells per well after culture in complete growth media overnight and subsequently in serum-free media for 1 day.

2.2. LPS-induced activation of macrophages

In a 24-well plate, Raw 264.7 cells were pre-seeded in the complete growth media at a density of 1 × 10⁵ cells/mL (0.5 mL/well) overnight. Cells treated with serum free media were replaced with or without LPS at differing doses and durations: 10 ng/mL or 100 ng/mL for 6 or 24 h, respectively, before proceeding to impedance cytometry and secretion analysis.

2.3. Ex vivo dissection of mouse intervertebral disc and their pretreatment

Animal protocols were approved by the Institutional Animal Care & Use Committee at the University of Virginia prior to the study. Mice were euthanized in a CO₂ chamber, followed by cervical dislocation. Mouse lumbar IVD (without bony endplates) (L1-L6, 5 discs per animal) were dissected from C57BL/6 mice (8–12 weeks, 20–25 g, both male and female) (Envigo, Indianapolis, IN) under a dissecting scope following published protocol (Dai et al., 2019). In brief, dissected discs were immersed in sterile PBS with 2% antibiotic-antimycotic for 20 min, before transferring into DMEM (high glucose 4.5 g/L) containing 1% penicillin (100 μg/mL) and 1% streptomycin (100 μg/mL). After equilibrating in the incubator for 30 min, discs were randomly assigned to place 3 discs per well, in a 24-well plate. To induce disc degeneration, discs (3 discs per well) were incubated with fresh serum-free media (0.5 mL per well), with human interleukin-1β (IL-1β) (Abcam, USA) (100 ng/mL) for 2 days. To demonstrate the potential phenotype reversal effect of Curcumin, discs were incubated with fresh serum-free media with IL-1β (10 ng/mL) or IL-1β (10 ng/mL) +Curcumin (10 μM) for 1 day.

2.4. Co-culture of macrophages with degenerated mouse disc

In a 24-well plate, Raw 264.7 cells were pre-seeded in the complete growth media at a density of 5 × 10⁴ cells/well (in 0.5 mL) overnight and replaced with serum-free media. Discs subjected to various treatment (as above) were rinsed with warm DMEM twice before transferring into macrophage wells, via a sterilized fine tweezer, to allow coculture for 1 day at 37 °C in a humidified atmosphere with 5% CO₂.

2.5. Impedance cytometry

At designated time points, macrophages subjected to the described treatment and co-culture protocols were harvested by trypsin and resuspended in serum-free DMEM (1 mL per well per sample), for single-cell impedance cytometry to measure the biophysical phenotypes (Fig. 1 and Fig. S1). A syringe pump was used to introduce macrophages, with co-flowing 7 μm polystyrene beads (Sigma) at ~1.2 × 10⁵ beads/mL into the microchannel with patterned electrodes for a detection region 30 μm (width) × 30 μm (height) (McGrath et al., 2020), for measurement at ~300 events/s at four simultaneous frequencies (0.5, 2, 18 and 30 MHz) using an impedance analyzer (Amphasys AG, Switzerland). Acquisition settings were optimized for signal to noise, based on levels of signal modulation, amplification, demodulation, and the trigger voltage level. Processed signal data (impedance magnitude and phase) were analyzed using Python.

Fig. 1.

A. Biological sample preparation, including: (i) ex vivo dissection of mouse lumbar intervertebral discs, (ii) stimulation with inflammatory cytokine IL-1β, and (iii) co-culture with macrophages in a 24-well plate (iv). B. Impedance cytometry of activated macrophages.

2.6. Impedance data processing and statistical analysis

The impedance phase and magnitude for macrophages were normalized by division to those obtained for polystyrene beads to account for any temporal variations during the measurement and to enable quantitative comparison between measurements (Salahi et al., 2022). Due to normalization, impedance phase is reported in arbitrary units (with respect to impedance phase of beads indexed at zero). The macrophage cells were gated from reference beads in normalized impedance data at 30 MHz for analyzing the gated macrophages at each probe frequency (0.5–30 MHz) (SI Fig. S2). All statistical analysis was performed using a custom script in Python. Significance level was defined as p < 0.05. Comparisons between any two groups were done using a student’s two-tailed t-test, while comparisons between multiple groups were performed using a one-way ANOVA with a Tukey’s multiple comparisons test. All results are representative of at least three repetitions; with error bars indicating standard deviation between sample triplicates.

2.7. Validation of macrophage activation with real-time RT-PCR and Griess assay

After incubation with LPS or coculture with discs, the NO in culture media was measured with Griess assay (Promega, Madison, WI) following standard protocols (Li et al., 2017). Total RNA of macrophages was isolated with Trizol Reagent (Invitrogen) and quantified with NanoDropTM 1000 Spectrophotometer (ThermoFisher Scientific, Waltham, MA). One microgram of total RNA was used for cDNA synthesis with iScript cDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA) following the manufacture’s instruction. Real-time reverse transcription polymerase chain reaction (RT-PCR) was performed with RT^2 SYBR Green Fluo FAST Mastermix (Qiagen Sciences, Germantown, MA) on a QuantStudio3 Real-time PCR system (Applied Biosystems, Waltham, MA). The relative mRNA levels of target genes were calculated by the comparative CT method (also known as the 2–ΔΔCT method) using 18S as an internal control. The mRNA expression of inos, and tnf-α were evaluated. The 18S rRNA were used as internal controls. Sequences of primers listed in Table S1.

2.8. Macrophage cell morphology analysis

After coculture with discs (pre-treated with 10 ng/mL IL-1β or 10 ng/ mL IL-1β+10 μM Curcumin) for 1 day, blight-field images of macrophages (200 × magnification) were assessed for morphology (3 region-of-interests per sample per SI: Figs. S9–S10). Cell boundaries were traced by ImageJ for automated analysis of cell area, perimeter, circularity, and aspect ratio. In each condition, ~100 cells were analyzed to obtain mean values of these morphological parameters.

3. Results and discussion

3.1. Optimizing macrophage seeding density to control for heterogeneity

The measured impedance magnitude (|Z|) represents amplitude reductions and impedance phase (φZ) represents temporal alterations to the input sinusoidal signal, as created by each passing cell suspended in its culture media, which is measured simultaneously at four frequencies of interest. These include: 0.5 MHz due to its dependence on particle size, 2 MHz due to its dependence on membrane properties, 18 MHz due to its dependence on cell interior and 30 MHz to allow for gating of cells versus co-flowing beads. After normalization of the data from each cell to data from co-flowing beads, the metric of electrical diameter (or $\sqrt[3]{{\left|Z\right|}_{0.5\text{MHz}}}$) is with respect to beads of known size, opacity (|Z|_{plobe freq.}/|Z|_{0.5 MHz}) is with respect to beads at opacity of unity, and impedance phase is with respect to beads at a phase level of 0. Per dielectric model simulations using typical biophysical cellular parameters, cell size differences sharply alter the |Z_0.5MHz| (SI: Fig. S3) and plasma membrane conductance alterations can be detected based on φZ in the 0.5–3 MHz range (SI: Fig. S4B), while its capacitance alterations modify the φZ response in the 0.5–10 MHz range (SI: Fig. S4A) and the opacity in the 2–8 MHz range (SI: Figs. S4C–S4D). Similarly, conductivity alterations to the cell interior modify the φZ and opacity in the >10 MHz range for cells in the 0.8–1.2 S/m range (SI: Fig. S5A & Fig. S5C), while permittivity alterations to the cell interior modify only the φZ in the >10 MHz range. Many of these simulated impedance metrics have been experimentally validated using modified red blood cells (Salahi et al., 2022).

Intercellular interactions strongly influence the phenotype of highly migratory cells, such as macrophages (Shim et al., 2011), thereby creating a degree of heterogeneity in their polarized morphology under activation. Using three different macrophage seeding densities per well (# cells/mL): high (5 x 10⁵), medium (2 x 10⁵), and low (1 x 10⁵), we observe that higher cell densities limit cell spread to cause successively smaller cell sizes (SI: Fig. S6A & Fig. 2A–C). The resulting alteration in size and membrane conditions cause the histograms φZ_{0.5 MHz} to show distinct subpopulations (Figs. 2D and SI: Fig. S6B), as indicated by arrows. While the majority population (green arrow) is relatively unshifted, a downshifted minority subpopulation (red arrow) is apparent at all three cell densities within the scatter plots (Fig. 2A–C) and the histograms (Fig. 2D and Fig. S6B). The emergence of this minority subpopulation is not due to loss of cell viability, which would cause much lower φZ_{0.5 MHz} levels, as observed for heat treated cells in Fig. 2D. Instead, we attribute it to a stressed subpopulation with arrested migration ability for cells seeded at higher densities. To ensure that intercellular interactions do not influence the polarized morphology of activated macrophages, we maintain a low cell density level in all subsequent work, wherein this subpopulation is highly reduced. Decrease in cell size arising from higher cell density would lower electrical diameter (or $\sqrt[3]{{\left|Z\right|}_{0.5\text{MHz}}}$) and φZ_{0.5 MHz} (Fig. S3B), but the latter impedance metric is also a sensitive indicator of the cell membrane condition (De Ninno et al., 2020); particularly alterations its RC time constant (product of resistance and capacitance per SI: Figs. S4A–S4B). At higher frequencies, the impedance phase no longer exhibits a systematic dependence on cell density (SI: Figs. S6C–S6D), indicating minimal alterations to the cell interior. Future work uses high frequencies: φZ_{2 MHz} and φZ_{18 MHz} that are sensitive to the phenotypes in the cell interior (SI: Figs. S5A–5B) for assessing activation-induced alterations to macrophages.

Fig. 2.

Heterogeneity in macrophage (Mac) phenotypes due to intercellular interactions as a function of cell seeding density. Scatter plot of impedance phase at 0.5 MHz (φZ_{0.5 MHz}) versus electrical diameter $\left(\sqrt[3]{{\left|Z\right|}_{0.5\text{MHz}}}\right)$ of macrophages is shown at different cell seeding densities: A. High (5 × 10⁵ cells/mL), B. Medium (2 × 10⁵ cells/mL), and C. Low (1 × 10⁵ cells/mL), with 0.5 mL per well. D. Histogram of φZ_{0.5 MHz} for control (untreated) macrophages at different cell seeding densities is compared to heat treated samples, to show stress rather than viability loss.

Fig. 5.

Macrophage activation by degenerated discs is reversed by curcumin treatment of discs by comparing control macrophages (Mac), Mac with discs pre-treated with IL-1β 10 ng/mL for 1 day (Mac+(IL-1)disc), and Mac with discs pre-treated with IL-1β 10 ng/mL and curcumin 10 μM for 1 day (Mac+(IL-1+Cur)disc) based on: A. Scatter plot (and histograms) of φZ_{2 MHz} vs. φZ_{18 MHz} show downshifting with degenerated discs and its reversal in presence of curcumin. B. Histograms of electrical diameter; Activation reversal: C. Not significant based on mean electrical diameter; but indicated by images D(i)-D(iii) and associated image analysis in Supplementary Information (Fig. S10); E. Not significant based on mean φZ_{2 MHz}; F. Significant based on mean φZ_{18 MHz} differences between samples (*p < 0.05 level, **p < 0.01). G. Nitrite secreted in cell culture supernatant measured using the Griess assay indicates reversal of the activation phenotype (*p < 0.05), albeit not at single-cell sensitivity. Cross-validation based on mRNA expression is shown in SI: Fig. S8.

3.2. Dose and duration-dependent LPS-activation

Activation of immune cells is known to increase their electrical diameter (Petchakup et al., 2021) and this murine macrophage cell line (Raw 264.7) exhibits increases in cell size with activation under LPS stimulation, as determined by Coulter cell size analysis at high dose and duration levels (100 ng/mL, 1 μg/mL, and 10 μg/mL for 2 days) (Saxena et al., 2003). Herein, we focus on a lower dose and duration range (100 ng/mL for 24 h, 100 ng/mL for 6 h and 10 ng/mL for 6 h). Also, since electrical diameter only measures changes to the cell exterior, whereas macrophage activation can influence morphologies of the cell membrane and its interior region, we explore impedance phase (φZ) at various frequencies, as a multiparametric index to quantify the activation in a dose and duration-dependent manner. While alterations in impedance magnitude opacity do not exhibit systematic differences with LPS stimulation (SI: Fig. S7), the φZ levels at 2 MHz and 18 MHz show systematic downshifting (Fig. 3A), while the electrical diameter ($\sqrt[3]{{\left|Z\right|}_{0.5\text{MHz}}}$) shows systematic increases (Fig. 3B) as a function of LPS stimulation dose and duration, as apparent from the significance plots in Fig. 3C–E. The activated phenotype measured by impedance is cross validated based on nitrite in the media, using the Griess assay that reflects levels averaged across the entire macrophage population. This confirms secretion of high levels of nitric oxide in media, especially after LPS stimulation at a dose of 100 ng/mL for 24 h (Fig. 3F), but the assay cannot detect significant differences at the lower dose or shorter duration. However, the impedance metrics reported here can detect the activation over lower LPS dose and duration levels, and the sensitivity of the respective metrics are compared in a later section.

Fig. 3.

LPS dose and duration-dependent macrophage activation based on comparing control (untreated) macrophages to those treated with LPS doses of 10 ng/mL and 100 ng/mL for 6 h, and 100 ng/mL for 24 h. A. Scatter plot (and histograms) of φZ_{2 MHz} vs. φZ_{18 MHz} shows downshifting with activation. B. Histograms of electrical diameter; C. Mean electrical diameter shows significant differences between samples (***p < 0.001); D. Mean φZ_{2 MHz} shows significant differences between samples (**p < 0.01 level; ***p < 0.001); E. Mean φZ_{18 MHz} shows significant differences between samples (*p < 0.05 level; **p < 0.01 level; ***p < 0.001); F. Cross-validation based on secreted nitrite measured in culture media using the Griess assay. Polarized morphology of macrophages after 6 h LPS treatment is in SI: Fig. S9B.

3.3. IL-1 induced disc degeneration leads to macrophage activation

The pro-inflammatory cytokine, interleukin-1β (IL-1β), is highly expressed in degenerative disc tissues and is a major mediator of disc degeneration and inflammation (Le Maitre et al., 2005). For instance, direct injection of IL-1β promotes progression of degeneration in a rat lumbar disc herniation model (Kim et al., 2021) and IL-1 contributes to disc disease through degradation of extracellular matrix (Johnson et al., 2015), leading to activation through the MAPK, NF-kB and C/EBP signaling pathways (Wang et al., 2013). The resulting macrophage phenotypes have been measured based on their polarized morphology (Silva et al., 2019) and cytokine secretion profile (Takada et al., 2012). Herein, the phenotype of macrophages is compared to that after co-culture with invertebrate discs, before and after IL-1 pretreatment to induce disc degeneration. The scatter plots of single-cell macrophage events based on φZ_{2 MHz} and φZ_{18 MHz} metrics (Fig. 4A) show downshifting after co-culture with degenerated discs. The electrical diameter metric shows an increase for the macrophage control sample after interaction with both, untreated and IL-1 treated discs (Fig. 4B), but with no statistical significance between the two disc conditions (Fig. 4C). On the other hand, downshifting of the φZ_{2 MHz} and φZ_{18 MHz} metrics is systematic and statistically significant between the two disc conditions (Fig. 4D–E), indicating sensitivity to the changes in macrophage phenotype after interaction with untreated versus degenerative disc. It is also clear that in comparison to the naive macrophages (control mac), IL-1 pretreated discs significantly promote (*p < 0.05) the mRNA expression of iNOS (Fig. 4F) and TNF-α (Fig. 4G) in macrophages, and this is also apparent based on the nitrite level in the culture media (Fig. 4H). Hence, we infer that the φZ_{2 MHz} and φZ_{18 MHz} metrics can detect alterations in the activation state of macrophages due to degenerative discs, which resembles that obtained after LPS stimulation.

Fig. 4.

Degenerated discs stimulate macrophages (Mac) towards a pro-inflammatory phenotype based on impedance metrics, NO secretion, and inflammatory gene expression; Mac with control discs, and Mac with degenerated disc (IL-1β 100 ng/mL, 2 days). A. Scatter plot (and histograms) of φZ_{2 MHz} vs. φZ_{18 MHz} shows downshifting with degenerated discs to indicate activation. B. Histograms of electrical diameter; C. Mean electrical diameter shows significant differences between samples (***p < 0.001); D. Mean φZ_{2 MHz} shows significant differences between samples (*p < 0.05 level; ***p < 0.001); E. Mean φZ_{18 MHz} shows significant differences between samples (*p < 0.05 level,***p < 0.001). PCR analysis of inos (F) and tnf-α, (G) mRNA levels of control macrophages (Mac) and Mac with degenerated discs (IL-1 stimulation) (*p < 0.05); H. Nitrite contents in culture media using the Griess assay. The polarized macrophage morphology after co-culture with IL-1 treated discs in shown in SI: Fig. S9D.

3.4. Curcumin reverses degenerated disc-induced macrophage activation

Treatment with curcumin is known to significantly reduce IL-1α and IL-6, iNOS, COX-2 and MMP-9 levels in rats with surgically induced lumbar intervertebral disc (Hu et al., 2017) and it regulates macrophage polarization through altering cytokine expression (Chen et al., 2021). Intraperitoneal injection of curcumin is also effective in attenuating mouse lumbar radiculopathy by reducing neuroinflammation, oxidative stress and nociceptive factors (Xiao et al., 2017). The φZ_{2 MHz} vs. φZ_{18 MHz} scatter plots from single-cell macrophages events (Fig. 5A), confirm downshifting of their respective levels after co-culture with IL-1β pretreated discs and recovery to their original level after curcumin treatment. Also, while an increase in cell electrical diameter is observed for the macrophages after co-culture with IL-1 pretreated discs to indicate activation (Fig. 5B), the recovery after curcumin treatment is not statistically significant (Fig 5C). Images of cell morphologies (Fig. 5D(i)–5D(iii)), alongside the associated image analysis in Supplementary Information (SI: Figs. S9–S10) confirm that IL-1β induced disc degeneration promotes macrophage polarization to cause larger cell sizes (area and perimeter), higher aspect ratios and less circularity levels (*p < 0.05 vs naive macrophages). With curcumin + IL-1β pretreated discs (M0+ (IL-1+Cur) disc), these morphological changes are greatly diminished (*p < 0.05 vs mac+(IL-1) disc group). The activated macrophages (M0+(IL-1) disc) show an increase in size in comparison to the naive macrophages (Mac), which is in accordance with the increase in electrical diameter noted earlier through impedance-based analysis (Fig. 5B). The activated macrophages are more dendritic/spindle shaped, resulting in an increased perimeter and aspect ratio due to the stretched cellular extensions, and therefore a concurrent decrease in circularity in comparison to naïve macrophages (Mac). Upon treatment with curcumin (Mac+(IL-1+Cur)disc), the morphology of the macrophages more closely resembles the naive macrophages (Mac) indicating an amelioration or inhibition of activation, but these results are based on just 100 cells. On the other hand, alterations in the φZ_{18 MHz} levels determined from automated analysis of ~10,000 events are statistically significant (Fig. 5F), but respective alterations in φZ_{2 MHz} levels are not significant (Fig. 5E). Due to specificity of the φZ at high frequencies (>10 MHz) to interior conductivity and permittivity (SI: Figs. S5A–5B), we suggest that macrophage activation induces alterations to the intracellular regions (mitochondria, endoplasmic reticulum, and polarization of interior morphologies), which are more sensitively transduced to significant changes for φZ_{18 MHz}, but not for φZ_{2 MHz}, wherein cell membrane alterations have a role (SI: Figs. S4A–S4B). Nitrite secretions confirm the ability of curcumin to reverse the activation of macrophages by IL-1;β pretreated discs (Fig. 5G), while cross-validation based on mRNA expression is shown in SI: Fig. S8.

3.5. Impedance metrics indicate dose and duration-dependent macrophage activation

Macrophage activation under different environmental cues can follow a variety of signaling pathways that may involve different subcellular regions. Hence, we consider the sensitivity of our approach to assess macrophage activation in a stimulation dose and duration-dependent manner, using impedance metrics of φZ_{2 MHz} (Fig. 6A), φZ_{18 MHz} (Fig. 6B) and electrical diameter or $\sqrt[3]{{\left|Z\right|}_{0.5\text{MHz}}}$ (Fig. 6C).

Fig. 6.

Macrophage activation at varying doses of IL-1 pretreatment and curcumin treatment to reverse the phenotype is benchmarked versus that obtained after stimulation with varying doses and duration levels of LPS, using the metrics of: A. φZ_{2 MHz}, B. φZ_{18 MHz}, and C. electrical diameter or $\sqrt[3]{{\left|Z\right|}_{0.5\text{MHz}}}$ to highlight the dynamic range of the sensor for quantitative comparisons (*p < 0.05 level, **p < 0.01, ***p < 0.001). The respective significance levels after LPS treatment are in Fig. 3.

All three metrics can use LPS stimulation over a range of dose and duration levels as the basis to create a progression of macrophage activation states to benchmark versus that obtained under stimulations from disc degeneration due to IL-1β. For instance, IL-1β treatment at 10 ng/mL and 100 ng/mL levels causes activation intensities that compare with LPS stimulation for 6 h at levels between 10 ng/mL and 100 ng/mL, with a good degree for correspondence for all three metrics. Systematic increases in electrical diameter are easily deduced under progressive activation with LPS (Fig. 3B–C) and this physical property can possibly be used to enrich for subpopulations with the highly activated phenotype. However, impedance phase information, which follows the RC time constant behavior of macrophage cells at each frequency is needed for subtle distinctions, such as between LPS stimulation at 10 ng/mL versus 100 ng/mL after 6 h stimulation (Fig. 3D for φZ_{2 MHz}) and between untreated and IL-1 treated discs (Fig. 5F for φZ_{18 MHz}). Similarly, while the metrics of electrical diameter $\sqrt[3]{{\left|Z\right|}_{0.5\text{MHz}}}$ in Fig. 6C) measuring the cell exterior, and φZ_{2 MHz} in Fig. 6A measuring plasma membrane alterations can significantly detect differences in macrophage activation state under the two studied IL-1β doses used to cause disc degeneration (10 ng/mL and 100 ng/mL), only φZ_{18 MHz} can detect recovery from the activated state at the lower dose (10 ng/mL) after curcumin pretreatment (Fig. 6B), likely due to the effect of macrophage polarization on organelles in the cell interior. To better understand this, we consider the dielectric simulations of φZ as a function of cytoplasmic conductivity changes (σ_cyto in SI: Fig. S5A), as would occur due to ion pile-up in the endoplasmic reticulum, or as a function of cytoplasmic permittivity changes (ε_cyto in SI: Fig. S5B), as would occur due to alterations in the mitochondrial phenotype. Based on this, φZ_{18 MHz} is sensitive to σ_cyto at levels of >0.8 S/m, whereas φZ_{2 MHz} is sensitive to σ_cyto at levels of <0.8 S/m. Furthermore, φZ_{18 MHz} is sensitive to σ_cyto, whereas φZ_{2 MHz} and opacity over the full frequency range (SI: Fig. S5D) do not show any sensitivity to ε_cyto. The φZ_{2 MHz} and opacity levels over the full frequency range also depend on membrane properties (ε_mem and σ_mem); hence, φZ_{18 MHz} is preferred if specificity to cell interior is required. These observations highlight the importance of the multifrequency approach to determine the appropriate set of impedance metrics that capture the full spectrum of activation for each type of stimulation, especially for cell types or stimulations with unknown activation response.

4. Conclusions and outlook

For screening immunomodulation therapies that can arrest inflammation due to macrophage infiltration for addressing chronic back pain, there is a need to quantify macrophage activation phenotypes under the relevant stimulations, with single-cell sensitivity, and in a dose and duration dependent manner. Since cytometry after staining for molecular markers relies on a particular signaling pathway to discern activation, we explore biophysical cytometry based on single-cell impedance metrics to quantify the activations in a multivariate manner over several frequencies that probe different subcellular regions. We infer that while the metric of cell electrical diameter can detect activation based on systematic increases, impedance phase metrics, particularly at high frequencies that detect phenotypes in the cell interior are needed for dose and duration dependent strategies to monitor macrophage activation by degenerated intervertebral discs and assess the effects of curcumin to reverse this activation. We envision the application of impedance cytometry to follow activation of macrophages under relevant biomimetic stimulations in a microfluidic device for in vitro screening of immunomodulation therapies to address the chronic problem of back pain and to assess immune responses to emerging transplants.