Lipid Droplets are rigid and physically suppress phagocytosis, unless cell compression or stretching activates actomyosin

Abstract

As with many cell types, macrophages are sometimes filled with micron-sized lipid droplets (LD’s), but effects on phagocytosis of other cells, particulates, and microbes remain unclear. Here, we show LD’s re-structure the cytoskeleton but remain round, consistent with a high interfacial tension, and they also impair actomyosin-driven uptake regardless of target size. Engulfment starts at the apical surface, but LD’s displace apical actomyosin to the basal cortex. However, actomyosin is activated and phagocytosis rescued when applying tissue-relevant compressive stress to LD-loaded macrophages. Macrophages that are densely filled with LD’s or pre-engulfed beads likewise activate actomyosin which again rescues phagocytosis relative to sparsely loaded cells. LD’s and rigid beads also impede macrophage migration through small pores, and LD’s pressed strongly into a nucleus can even cause rapid focal rupture independent of actin. LD’s thus disrupt cytoskeleton organization and nucleus integrity, generally suppressing motility processes unless actomyosin is activated by cell compression or stretching.

Introduction

Macrophages engulf other cells or debris as well as pathogens and other foreign particles that enter tissues. Phagocytic uptake typically depends on ligand-receptor interactions and the cytoskeleton as well as target properties such as size and flexibility, (Fig. 1A) (Aderem and Underhill 1999, Gordon 2016). Immunoglobulin-G is one major phagocytic ligand that binds a target to macrophage Fcγ receptors (FcγR) that then locally trigger F-actin-rich extensions for engulfment. Whether lipid droplets (LD’s) in macrophage cytoplasm impact the cytoskeleton and phagocytosis is understudied but likely relevant to cancer and atherosclerosis among other pathologies where LD’s accumulate in macrophages (Peyron, Vaubourgeix et al. 2008, Moore and Tabas 2011, Su, Wang et al. 2020). Metabolic perturbations by LD’s are widely discussed (Wculek, Dunphy et al. 2022), but we recently reported that micron-diameter LD’s are always spherical within cells and act as rigid bodies (Ivanovska, Tobin et al. 2023). We hypothesized therefore that LD’s cause a more direct biophysical disruption of the cytoskeleton that inhibits phagocytosis.

Figure 1. Lipid droplets effects on phagocytosis.

A Macrophages engulf widely varied targets, but cytoplasmic lipid droplets (LD’s) have unclear effects on actomyosin structure and phagocytosis.

B. LD flexibility and roundness relate to cytoskeleton filament flexibility and deformation. A monolayer of mostly phospholipid on the LD gives rise to an area elastic constant that is twice the interfacial tension of the underlying hydrophobic phase (Fig S1A) (Israelachvilli 2011). An estimate of the change in energy per area is thus provided by ${\gamma }_{\text{eff}}$ when a spherical LD is transformed to another shape.

C. THP-1 human macrophages accumulate micron-sized, spherical LD’s after 24 hr oleic acid (OA) addition. ii) LD-loaded macrophages do not significantly polarize relative to control macrophages (n>185 total cells/condition; not significant ‘n.s.’ per unpaired t-test).

LD’s are phase-separated oil drops that typically consist of triglycerides surrounded by a phospholipid monolayer, and we recently measured an effective interfacial tension for LD’s which is typical of oils and which explains spherical shapes of LD’s in cells even when contacting other cell structures (Fig. S1A). As an important comparison, most cells possess a cortical tension attributable to acto-myosin, but this tension is order(s) of magnitude lower than the effective interfacial tension of LD’s. This cortical tension of macrophages nonetheless helps pull a target inward (Heinrich 2015, Jaumouille and Waterman 2020) except when the cell surface is so expanded by uptake that a macrophage is “full” and unable to engulf more (Zent and Elliott 2017). Inhibition of the acto-myosin network also suppresses engulfment (Malawista, Gee et al. 1971, Newman, Mikus et al. 1991, Tsai and Discher 2008, Barger, Gauthier et al. 2020) as does softness of a target (Beningo and Wang 2002, Sosale, Rouhiparkouhi et al. 2015, Jain, Moeller et al. 2019, Jaumouille and Waterman 2020, Vorselen, Barger et al. 2021). This all leads to the hypothesis: if cytoskeletal filaments in a macrophage are deformed and displaced by LD’s rather than vice versa (Fig. 1B), then the efficiency of phagocytosis should be impacted.

Phagocytosis can be enhanced by association of LD’s with a target bacteria or by a change in metabolic profile favoring triglyceride or macrophage differentiation over time (Bosch, Sanchez-Alvarez et al. 2020, Castoldi, Monteiro et al. 2020, Singh and Sen 2021). However, engulfment can also decrease because of energy deprivation from impaired lipolysis and LD accumulation. Likewise, impaired phagocytosis can also reflect progressive changes in phagocytosis-related gene expression (Chandak, Radovic et al. 2010, Agarwal, Combes et al. 2020, Luo, Zheng et al. 2020, Marschallinger, Iram et al. 2020). LD’s form rapidly after uptake of fatty acids and other lipids in the micro-environment (Maxfield and Tabas 2005, Yang, Qin et al. 2022). LD’s also form after engulfment of all or part of an adipocyte or other LD-laden cell (Cinti, Mitchell et al. 2005), but it is unclear whether rapidly accumulated LD’s exert biophysical effects on the cytoskeleton and phagocytosis. LD-nucleus contacts have the further potential to deform and even rupture the macrophage nucleus by imposing high curvature (Xia, Ivanovska et al. 2018, Pfeifer, Tobin et al. 2022, Ivanovska, Tobin et al. 2023). Thus, LD interactions with the cytoskeleton and nucleus can be foundational to mechanobiological origins of immune cell function or dysfunction.

In the present study, targets range from small and rigid like a bead or microbe to large and soft like a cancer cell, and we show for all that LDs in macrophages are again always round and suppressive of phagocytosis by ~30%. Importantly, the result is phenocopied by pre-loading microbeads into macrophages prior to a phagocytosis assay, which further supports a biophysical basis for suppressed engulfment. Rescue of phagocytosis deficits is also achieved by dense loading of LD’s or beads in macrophages and also by tissue-mimetic confinement of macrophages – both of which activate acto-myosin. Stress-activation and tension maintenance are known properties of myosin-II interactions with F-actin (Kovacs, Thirumurugan et al. 2007, Spinler, Shin et al. 2015, Schiffhauer, Luo et al. 2016, Park-Holohan, Brunello et al. 2021). They help explain for example the Frank-Starling mechanism of ventricle contraction upon filling. Mechanosensitive activation in phagocytosis seems comparatively understudied. Compressed macrophages further show that LD’s can drive nuclear rupture. Our results may help explain previous findings, such as an age-dependent suppression by ~30-40% of phagocytosis for monocyte-derived macrophages wherein one of the key regulatory factors (Moss, Johnston et al. 2024) associates with elevated lipid (Pajukanta, Lilja et al. 2004), although such findings are also mixed (Agarwal, Combes et al. 2020, Bosch, Sanchez-Alvarez et al. 2020, Singh and Sen 2021) and motivate alternative approaches. Equally important perhaps, inefficient phagocytosis caused by LD’s in these innate immune cells can limit downstream processes such as antigen presentation and adaptive immunity (Arango Duque and Descoteaux 2014, Gordon 2016).

Results

Lipid Droplets uniformly impair phagocytosis of diverse targets

To quantify the effects of cytoplasmic LD’s on phagocytosis, we studied multiple large or small targets that were IgG antibody-opsonized (Fig 1A). Human THP-1 macrophages were used initially because pre-treatment with oleic acid (+OA) is well-known to result in small, micron-sized droplets that fill the cytosol (Fig 1A-i; Fig S1A). We also find no significant effect on differentiation markers for ‘polarization’ of the macrophages (proinflammatory CD80 and anti-inflammatory CD206) relative to controls (Fig 1C), which is important because distinct polarizations can impact phagocytosis (Kapellos, Taylor et al. 2016, Schulz, Severin et al. 2019). LD-loaded macrophages are nonetheless impaired in phagocytic uptake of opsonized 6 μm beads (Fig 2A). The 30% reduction in uptake by +LD cells reflects phagocytic deficiencies rather than insufficient adhesion, as both populations have comparable numbers of macrophages in contact with noninternalized beads as identified by secondary antibody staining after engulfment (Fig 2A). Consistent with our recent study (Ivanovska, Tobin et al. 2023), LD’s remain round and disrupt the actomyosin network. Biological relevance is again suggested by the similar ~30-40% suppression of phagocytosis for monocyte-derived macrophages from older individuals (>50 years) compared to younger individuals (18-30 years).

Figure 2. Loading of LDs or microbeads in macrophages reduces phagocytosis similarly for a wide range of targets.

A. Brightfield and immunofluorescence (IF) images show phagocytosis of IgG-opsonized 6 μm beads by control and LD-loaded macrophages. Internalized beads (white arrows) are distinguished by secondary antibody from any external beads (magenta arrows). Nucleus indentation is evident for LD’s and for beads. +LD macrophages showed a ~30% decrease in phagocytic uptake despite similar frequency of external bead contacts. (N=3, n>400 total cells/condition, p<0.05, unpaired t-test).

B. Across a range of opsonized target diameters, LD’s impair phagocytosis. i) CD47 KO A549 cancer cells (arrows) phagocytosed by SIRPα KO THP-1 cells with or without LD’s. ii) Nanobead phagocytosis was quantified as in (A) iii) Phagocytic uptake of lentivirus by control or +LD macrophages were quantified via downstream expression of transduced GFP. iv) LD-loaded macrophages exhibit impaired uptake (>20% decrease), +LD conditions are normalized to respective controls. (N=5, n>425 total cells/condition, p<0.05, unpaired t-test)

C. Phagocytic defects in macrophages preloaded with inert LD-sized, 3 μm beads. Timeline: THP-1 macrophages were preloaded with LD-sized beads for 0.5 or 5.5 hours prior to bead washout and addition of larger 6 μm microbeads. Images of 3 μm bead (red) addition for 0.5 hours showed minimal preloading whereas extended preloading resulted in cytosol-filled macrophages similar to LD-loaded cells. As with (A), internalized microbeads beads were identified by absence of secondary antibody (magenta). Macrophages preloaded with 3 μm beads exhibit a ~33% decrease for subsequent 6 μm bead phagocytosis compared to control (duplicates of N=2, n>131 cells/condition, p<0.05, unpaired t-test).

D. Primary, murine bone marrow derived macrophages (BMDM) phagocytose less when LD-loaded. i) Images: Brightfield and IF images of macrophage uptake of opsonized microbeads in +LD and control BMDM populations. External beads are labeled with secondary antibody (magenta), internal beads are in brightfield (white arrow). Cell populations containing micron-sized LD’s (green) are present in BMDMs after 96 hour OA treatment. Graph: Presence of LDs in BMDMs diminishes phagocytic uptake (by ~30%), similar to the THP-1 model. (n1,2>145, n3>35 (confocal microscopy) cells/condition, from N=2 mice, p<0.05, unpaired t-test). ii) LD’s cause cortical actin filament displacement in highly loaded BMDMs (arrowhead).

Lung cancer A549 cells opsonized by IgG showed a similar reduction (~25%) in phagocytosis by LD-loaded THP-1 macrophages relative to controls (Fig. 2B-i). These cancer cells are somewhat larger (~10 μm) and softer than beads (~1 kPa elastic modulus for cells versus >1 MPa for bead) and so target stiffness is not a primary determinant of the LD effect. Note that to maximize cancer cell phagocytosis and better resolve differences in uptake, the CD47-SIRPα ‘Don’t eat me’ axis was disrupted by knockouts of SIRPα and CD47 in the macrophages and A549 cells, respectively (Fig S1B). Uptake analyses of both large cells (Fig S1C) and much smaller nanobeads and lentivirus (Fig 2B-ii,iii) that also undergo phagocytosis (Rodriguez, Harada et al. 2013, Sosale, Ivanovska et al. 2016, Milani, Annoni et al. 2019) provide similar results to the above microbead and cancer cell data, with LD’s significantly suppressing phagocytosis versus controls (Fig. 2B-iv). Importantly, to evaluate whether LDs exert a physical effect rather than a strictly metabolic perturbation, we preloaded macrophages with rigid beads of a similar micron-size to LDs and then quantified phagocytosis of the 6 μm beads. THP-1 cells sufficiently pre-loaded with beads indeed exhibited a phenotype similar to LD-loaded cells, with a 33% reduction in phagocytic capacity to eat larger beads within the 1-hr assay (Fig. 2C).

Primary murine bone marrow-derived macrophages (BMDMs) were also loaded with LD’s and again showed reduced internalization of IgG opsonized microbeads (Fig 2D). Despite heterogeneity in loading with LD’s, visibly loaded BMDMs showed deficits of ~30% in phagocytosis relative to cells lacking LD’s (Fig 2D-i). Cortical actin filaments also conformed to the perfectly round contours of LD’s (Fig. 2D-ii, S1D). Such observations are fully consistent with the high interfacial tension and rigidity of LD’s per our recent study (Ivanovska et al. 2023). LD’s thus deform and displace cell membranes as well as the acto-myosin cytoskeleton and they impede the phagocytic uptake process across a broad range of macrophage targets and sources.

Acto-myosin perturbations by LD’s within macrophages

To further characterize alterations to cytoskeletal organization caused by LD’s in macrophages, we quantified F-actin organization as well as overall intensity and cell morphologies. We again noted that cortical F-actin conformed to the round contours of LD’s (arrow in Fig 3A-i) and used orthogonal projections of confocal z-stacks from LD-loaded cells to discover that LD’s cause F-actin to localize more at the base of cells. In LD-containing cells, the total intensity of cortical F-actin was ~30% lower on the uppermost surfaces of the cells and ~30% higher at the substrate-cell interface (Fig 3A-i, plot). The ratio of F-actin intensities (Ctrl/LD) indeed exhibits a strong, inverse correlation across all heights measured (Fig 3A-i, inset). Importantly, the ~30% depletion of acto-myosin at the upper cortex that engages targets for phagocytosis might begin to explain how LD’s suppress phagocytosis – despite no differences in overall F-actin levels or cell spread area (Fig 3A-ii). The lack of change in cell area is additionally important given that limits to cell surface expansion inhibit uptake as a macrophage becomes “full” and unable to engulf more (Zent and Elliott 2017).

Figure 3. Rigid LD’s or mechanical squashing drives actomyosin activation and rescues phagocytosis.

A. Actin is uniquely structured and rearranged in +LD macrophages. i) Confocal orthogonal images of LD-loaded and control macrophages display differences in actin distribution, with actin more prominent at the base of +LD macrophages. Arrowhead highlights peripheral actin displacement by an LD. Peripheral F-actin was quantified at multiple z planes within the same cell. Macrophages with LD’s exhibit higher F-actin at the cell base than controls, with decreased levels at higher height.

Inset: The F-actin ratio of +LD to control macrophages across the measured heights yields a significant linear correlation (n>22 cells/condition). ii) There is no difference in F-actin levels or cell spreading between macrophage populations (N=2, n>20 and >40 cells/condition respectively). iii) Actin depolymerization by low dose Latrunculin-A (Lat-A) significantly reduces phagocytosis in controls, but not OA-treated macrophages. High dose of Lat-A eliminates phagocytosis under both conditions (N=4, n>500 total cells/condition, p<0.05, one sample t-test).

B. LD’s displace myosin in macrophages. i) Distinct holes in the myosin network are observed upon LD formation within the cell. ii) Addition of myosin inhibitor blebbistatin (blebb) only in controls (N=3, n>400 total cells/condition, p<0.05, one sample t-test).

C. F-actin and non-muscle myosin-II colocalize around rounded lipid droplets. Representative confocal z-slices of the perinuclear region (~3 μm height) and basal surface of LD-loaded macrophages (bright field) are shown (n > 50 cells). Intensity profiles correspond to the sections indicated by the green lines.

D. Mechanically-induced actomyosin activation enhances phagocytic capacity. i) Effect of LDs on phagocytosis within a physiologically relevant microenvironment was examined using a micropillar confinement system. Confined macrophages remained intact and 3 μm microbead uptake levels were assessed as per Fig 2A. ii) Confinement promoted higher phagocytosis in +LD macrophages compared to controls (N=3, n>500 total cells/condition, p<0.05, one sample t-test). Addition of blebb eliminated the phagocytic advantage of +LD macrophages in confinement (N=3, n>300 total cells/condition, p<0.05, one sample t-test). iii) LD loading impacts macrophage uptake under confinement. Cells with high LD loading uptake two-fold more than sparsely filled macrophages (N=2, N>200 total cells/condition, p<0.05, unpaired t-test). iv) Phosphorylated myosin light chain 2 (p-MLC2) is higher in macrophages undergoing 3 μm confinement compared to unconfined with 19 μm (N=1, n>25 cells/condition, p<0.05, unpaired t-test). v) Confinement of HeLa cells also increases myosin phosphorylation. Tubulin is a loading control (representative of N=2).

D. Compression stretches the cortex and activates actomyosin as does overfilling a cell with LD’s or beads. In both contexts the nucleus is also strained or deformed, which provides another organelle-scale hallmark of mechanical effects. The equation for active Acto-myosin forces is from (Zemel, Rehfeldt et al. 2010) and also applies to parts of the cell such as the cortex.

To provide more direct evidence of the importance of acto-myosin in engulfment, phagocytosis of microbeads was studied with the typical acto-myosin inhibitors (Fig S1E). Low doses of Latrunculin-A partially depolymerize F-actin and significantly decreased phagocytosis by controls but showed no significant effect on macrophages with LD’s (Fig 3A-iii). High doses of Latrunculin-A eliminates phagocytosis for both conditions (Fig 3A-iii). Both sets of results fit linear trends, but the slope is weaker for LD’s. To assess an effect of myosin-II in phagocytosis, which is also displaced by droplets (Fig 3B-i), a saturating dose of the myosin-II ATPase inhibitor Blebbistatin was added during the phagocytosis assays. Once again, only control cells showed a sufficient reduction (~15%) in bead uptake to be significant (Fig 3B-ii). Optical sections from confocal z-stacks of actin and non-muscle myosin-II at both the cell cortex and basal regions revealed extensive colocalization, including at sites where actomyosin fibers bend around and conform to the surfaces of lipid droplets (Fig. 3C). However, a more detailed understanding of their structural conformation in response to the mechanical stress imposed by the lipid droplets, will require super-resolution imaging. Lastly, we evaluated whether uptake of opsonized ~200 nm lentivirus as identified by GFP expression (Sosale, Ivanovska et al. 2016, Milani, Annoni et al. 2019) (Fig 2A-iii) was altered by myosin inhibition, and we again found a ~20% decrease only for the controls (Fig S1F). Interestingly, the suppressive ~15-20% effect of complete myosin-II inhibition is only a fraction of the 25-35% inhibition caused by LD’s. We therefore conclude that F-actin perturbations are more important.

The cytoskeleton perturbations by LD’s motivated us to seek additional conditions that impact acto-myosin and phagocytosis.

Rescue of LD-suppressed phagocytosis by compression-activation of actomyosin

Within tissues, macrophages sometimes squeeze into comparatively stiff confining spaces. Normal adipose tissue is relevant to the present studies, and we have measured an effective tissue surface tension of 30–70 mN/m (Fig.S2A,B) that reflects contributions from adipocyte LD’s plus extracellular matrix between the large adipocytes (FigS2C). Macrophages have an effective cortical tension of only ~0.1 mN/m (Rinker, Kirkpatrick et al. 2004), and so they will be compressed when they migrate between adipocytes. To investigate how LD’s impact phagocytosis under physiologically relevant 3D conditions of compression, macrophages were vertically confined to a 4 μm height using PDMS micropillars and also provided with 3 μm diameter opsonized beads for 3 hrs of phagocytosis (Fig 3C-i). We estimated the macrophage height to be approximately 6 μm from confocal z-slices (mean 6.5 μm, SEM = 0.35 μm, n = 24), so that confinement resulted in sustained strain of ~40%.

Intriguingly, confinement of the LD-loaded macrophages increased phagocytosis by ~50% relative to confined control cells lacking LD’s (Fig 3C-ii). The difference is completely abrogated with blebbistatin which again reduces phagocytosis for controls (more so than in 2D per Fig 3B-ii). This indicates that confinement synergizes with LD’s to strongly activate actomyosin. Moreover, confined cells with a cytoplasm densely filled with LD’s exhibited a two-fold increase in phagocytic activity compared to sparse filling (Fig 3C-iii). Immunofluorescence and western blot analyses of phosphomyosin light chain (p-MLC2) levels for both confined and non-confined cells revealed that confinement induces MLC2 phosphorylation (Fig 3C-iv,v, S3A). Recent studies have likewise shown that such micron-spaced confinement drives cytoskeletal rearrangement in HeLa cells which triggers sustained actomyosin contractility and recruitment of cortical myosin-II (Lomakin, Cattin et al. 2020, Garcia-Arcos, Ziegler et al. 2024).

Rescue of phagocytosis by high densities of LD’s or beads via actomyosin activation

Because external stress activates actomyosin to promote phagocytosis when combined with LD’s, we hypothesized that a strictly internal stress – without external compression – caused by especially high levels of LD’s and rigid beads could also rescue phagocytosis within standard 2D cultures. Heterogeneous loading of LD’s between macrophages was used to show by immunofluorescence that sub-populations of cells filled with LD’s have higher levels of p-MLC2 compared to sparsely loaded cells (Fig 4A). A similar trend was clear after normalization of this signal for each cell to p-MLC2’s major binding partner, the non-muscle myosin-IIA heavy chain (Fig 4A, Fig S3B). Overall averages for these cultures nonetheless show a decreased ratio of activated myosin in LD+ cells compared to control cells, consistent with earlier results showing LD’s suppress contractility and phagocytosis in standard 2D cultures (Fig 2). Furthermore, given the apical loss of cortical F-actin within LD loaded cells in the average crosssection of standard 2D cultures (Fig 3A), we quantified the ratio of activated myosin in the LD-filled cells and found a higher (p-MLC2 / myosin-IIA) ratio than control cells through a height of 3 μm (Fig S3C). This increased activation at heights associated with uptake of 6 μm beads after they settle on or near cells predicts a higher level of phagocytosis for the LD-filled cells relative to sparsely loaded cells.

Figure 4. Sparse LD’s loading alters actomyosin structure and reduces phagocytosis, but abundant loading activates myosin and rescues phagocytosis.

A. In 2D culture, quantity of LD’s influences phosphomyosin level. Images of p-MLC2 showing reduced expression in macrophages with fewer LD’s. Yellow boxes indicate macrophages containing fewer LD’s. Plot of p-MLC2 intensity across three loading conditions (ctrl, low LD, high LD) LD’s (n>50 cells/condition, p<0.05, unpaired t-test). Inset: Ratio of p-MLC2 to non-muscle myosin-IIA (NM-IIA) reveals a ‘v’ shaped trend.

B. Macrophages with low LD loading are most impaired in phagocytosis. i) Area occupied by LD’s in OA treated cells was classified as either ‘high’ or ‘low’ relative to the cell area. ii) Macrophages with a low area fraction of LD’s phagocytose significantly fewer opsonized beads than their highly loaded counterparts despite no difference in external bead contacts (N=3, n>50 cells per condition, p<0.05, one sample t-test). iii) Graphical summary of LD quantity and its correlation to phagocytic uptake. Solid lines represent individual experiments (N=3). iv) Illustration of macrophage phagocytosis as a result of LD-loading. While low loading of lipid droplets disrupts the actomyosin network and perturbs phagocytosis, high loading of lipid droplets drives myosin activation due to more severe actomyosin displacement and partially rescues uptake.

C. Preloading macrophages with inert 3 μm beads confirm that filled macrophages uptake more than their sparsely-loaded counterparts. i) Macrophages preloaded for 5.5 hrs prior to feeding the 6 μm microbeads shows the filled macrophages phagocytose more readily (white arrowhead). ii) Categorization of preloaded macrophage populations by washout time. iii) Sparsely-loaded macrophages after extended preloading are more deficient at phagocytosis relative to those more filled (N=4, n>300 total cells/condition, p<0.05, one sample t-test). iv) Schematic of bead-preloaded macrophages exhibiting ‘v-’shaped phagocytic uptake. While bead-filled macrophages phagocytose more than macrophages that are sparsely-filled, the combined effect is an overall ~33% decrease in phagocytosis compared to controls, consistent with Fig. 2C.

D. Schematic representation of how rigid lipid droplets or confinement can stretch the actomyosin cortex, leading to actomyosin activation. At first approximation, the force generated by the cortex can be described by a linear Hooke’s law–like model $F=ak(l_c-l_c^R)$, where $F$ is the active actomyosin force, $a$ is a dimensionless coefficient representing the level of contractility (depending on myosin motor polarization and/or ATP availability), $k$ is the effective stiffness of the actomyosin network, $l_c$ is the projected cortical length after stretching or compression, and $l_c^R$ is the rest length.

To assess phagocytosis by cells that were LD-filled or sparsely loaded, we again categorized each macrophage based on the LD-occupied area relative to cell area (Fig 4B-i) and found LD-filled macrophages phagocytose ~50% more than sparsely loaded cells (Fig 4B-ii). The increase aligns with the increased phagocytosis caused by compression of LD-loaded cells (Fig 3C-ii), and we also note that cell-adhered beads that were not internalized showed no difference, which argues for differences in cytoskeleton-driven uptake mechanisms rather than target attachment. Our original results showing an overall average of ~30% suppression of phagocytosis for macrophages with LD’s relative to control macrophages can now be parsed: sparsely loaded macrophages reproducibly show a ~50-60% suppression of phagocytosis, and this is partially rescued by LD-filled cells (Fig 4B-iii). Such results support our hypothesis that internal stress caused by high levels of LD’s within standard 2D cultures rescue phagocytosis, with high levels of activated myosin providing evidence for high internal stress (Fig 4B-iv).

To assess whether the partial rescue of phagocytosis by LD-filled cells has a mechanical origin, we pre-loaded macrophages with inert 3 μm beads to obtain ‘densely filled’ and ‘sparsely loaded’ sub-populations of cells which potentially phenocopy the LD filled and sparse area fractions (Fig 4C-i). Different sub-populations were obtained by washout of 3 μm beads after either 0.5 hr or 5.5-hr, which respectively gave pre-loading in many cells that was either ‘sparse’ or ‘filled’ (Fig 4C-ii). Subsequent phagocytosis of antibody-opsonized 6 μm beads showed a ~33% overall average decrease in phagocytosis by the pre-loaded cells with ‘sparse’ cells showing less phagocytosis of 6 μm beads than ‘filled’ cells (Fig 4C-iii,iv). These results with rigid and inert plastic beads (i.e. metabolically-dead) of similar size to LD’s replicate and extend the rescue of phagocytosis as a function of LD-loading (Fig 4C-iv, Fig 4A-plot, Fig 4B-iii, iv). The results also support the hypothesis that internal mechanical stress caused by high levels of LD’s and rigid beads within standard 2D cultures can rescue phagocytosis, which complements our vertical confinement experiments showing that confinement-induced mechanical stress also promotes phagocytosis when combined with LD’s. This finding aligns with the concept that, at first approximation, the mechanical stress activation of actomyosin contractility can be described by a linear Hooke’s law–like equation $F=ak(l_c-l_c^R)$, where the generated force depends on network stiffness, deviation from rest length, and a contractility coefficient determined by myosin activity and ATP availability, with $l_c$representing the projected length of the actomyosin cortex (Fig 4D).

Migration through pores is impeded by LD’s & rigid beads, which remain round

Phagocytosis is a localized and specialized motility process of macrophages that is driven by actomyosin, but cell migration is more general and requires whole-cell coordination of motility and actomyosin – especially in processes such as squeezing the cell through a narrow 3D pore (Xia, Pfeifer et al. 2019). Macrophage migration into a solid tumor or through normal tissue such as adipose tissue involves such squeezing. Thus, we hypothesized that such a motility process will also be impacted by LD’s and rigid beads. To study this in a simple model, macrophages loaded with LD’s were replated on Transwell filter membranes (with 8 μm diameter pores) at high cell density in order for cell crowding to stimulate migration through the membrane (Fig 5A-i). LD-loaded macrophages migrated ~40% less than control macrophages over 2 days (Fig 5A-ii), which is consistent with migration deficits observed in LD-loaded solid tumor cells (Ivanovska, Tobin et al. 2023) and also similar to the ~30% inhibitory effect of LD’s on phagocytosis (Fig 2). Pre-loading of 3 μm rigid beads into macrophages was anticipated to phenocopy the effect of LD’s and indeed led to a similar decrease (~ 35 %) in migration through pores (Fig 5A-ii). Images of cells in the pores show macrophage nuclei strongly deformed by LD’s or rigid beads – which always remain round despite the squeezing (Fig 5B). Such colocalization effectively reduces the pore diameter (by 50% or more in some images) and possibly explains how migration is impeded given that such 3D cell migration is further impeded by pore diameters smaller than 8 μm (e.g. 3 μm (Xia, Pfeifer et al. 2019)).

Figure 5. LD’s impair confined migration and confinement spaces can induce nuclear rupture.

A. Macrophage 3D migration is impacted by LD presence. i) Schematic: Transwell migration assays were used to quantify both migration efficiency and nuclear rupture events in THP-1 macrophages. Confocal z-slices of densely plated+LD macrophages on Tops and cells migrated to Bottoms of Transwell filters after 48 hours. ii). After 48-hour migration, LD’s or bead-preloaded macrophages show reduced migration through non-constricting 8 μm pores. (N=4, n>100 total cells/condition, square: bead loading, circle: LD loading).

B. Rigid cytosolic objects can deform the nucleus and impede cellular migration by narrowing the effective migration pathway. i) Both rigid beads (top) and LD’s (bottom) can deform the nuclei of cells undergoing pore migration. ii) Cartoon illustrating the nuclear deformation, and subsequently the smaller effective diameter, experienced by a cell nucleus caused by a LD in a migration pore. iii) Confocal z-slices of a nucleus as it enters a pore show that multiple LD’s can cooperatively squeeze a migrating nucleus to limit movement. C. In confined microenvironments, presence of LD’s increases macrophage nuclear rupture frequency. i) Schematic: Micropillar system imposed 3 μm height confinement upon macrophages to assess nuclear rupture events via nuclear accumulation of mcherry-cGAS. Confocal z-slices of live THP-1 macrophages expressing mcherry cGAS showing LD-driven nuclear rupture by accumulation of mcherry-cGAS foci at the nuclear membrane next to LDs ii) Categorization of nuclear rupture location (top, bottom, side) in LD-loaded and control macrophages after 30 mins of confinement. (N=3, n≥15 total ruptures/condition). iii) Introduction of confinement enhances nuclear rupture events in LD-loaded cells. After ~30 min, +LD cells exhibit twice as many nuclear rupture events relative to control cells and a 2.5-fold higher change in non-confining rupture levels. After 120 mins, LD-loaded cells show significantly higher nuclear rupture than control cells. The introduction of Lat-A (red) eliminates confinement-driven rupture events and reduces overall rupture. Dotted lines illustrate the best linear fits to control conditions. Middle plot illustrates that adjusting for control conditions, LD’s cause additional rupture immediately after introduction of confinement that are independent of actin depolymerization. Kinetics of LD-driven nuclear rupture are captured by the equation $y=A+B(1-e^{t∕\tau })$ with A=2, B=6, $\tau =20$. (N>2 per condition, n>200 total cells/condition, p<0.05, unpaired t-test). Images show that rupture is less frequent following the addition of Lat-A prior to confinement of control cells for 2 hours.

LD’s remain round as they rupture nuclei in compressed macrophages

The LD-nucleus interaction and the indentation that is evident within pores led us to hypothesize that nuclear rupture could occur at such sites of high Gaussian curvature as observed for cancer cells (Ivanovska, Tobin et al. 2023). The pressure exerted by a lipid droplet (FD) on a nucleus is determined by γ multiplied by the droplet’s curvature (the inverse of its radius), with higher curvature corresponding to greater pressure. To visualize any rupture, THP-1 macrophages were transduced with lentivirus to express mcherry-cGAS (Fig S4B), which is a cytosolic DNA sensor that also accumulates at sites of cancer cell nuclear rupture caused by LD’s in 2D culture (Ivanovska, Tobin et al. 2023). Interactions between the nucleus and rigid beads validated that cGAS foci occurred at the indented nuclear periphery consistent with nuclear rupture at sites of high curvature (Fig S4B-I,ii). In fixed 2D cultures, LD’s are present at sites of cGAS accumulation and associate with rupture (Fig S4B-iii), but overall rupture frequency is low (~1%). Lower frequency of nuclear rupture than previously observed in cancer cells might be expected as THP-1 macrophages have a much higher Lamin-A:B ratio that can more readily resist lamin-B dilution at high curvature (Pfeifer, Tobin et al. 2022, Ivanovska, Tobin et al. 2023). We therefore employed vertical confinement to a height of just 3 μm in order to increase the likelihood of LD-induced nuclear rupture in macrophages, and within minutes, macrophages exhibited focal accumulation of cGAS at nuclear sites of LD contact (Fig 5C-i). While most rupture events in both control and LD-loaded conditions occur at the high-curvature sides of the nucleus, LD-loaded cells exhibit a 3-fold higher frequency of LD-induced rupture on the flattened top and bottom surfaces of the nucleus after 30 min of tight confinement (Fig 5C-ii). Importantly, LD’s exhibit a roundness and diameter that is typical of the many other cytoplasmic LD’s – which again indicates the relative rigidity of LD’s.

Sustained confinement for ~120 min increases nuclear rupture in LD-loaded macrophages in parallel with rupture in control macrophages (Fig 5C-iii). Both increases depend on F-actin, as they are largely ablated by Latrunculin-A, with the difference driven primarily by the initial LD-induced rupture regime. While acto-myosin contractility dominates nuclear rupture processes during extended cell confinement, Latrunculin-A does not rescue the initial actin-independent LD-driven rupture caused by LD’s pressed against the nucleus (Fig 5C-iii, right plot). Given the substantial deformations that macrophages must endure within tissues including relatively rigid fat tissue (Fig S2), we anticipate that some rigid LD’s in the macrophage cytoplasm significantly increases the risk of nuclear rupture.

Discussion

LD’s accumulate in macrophages in multiple pathologies and tissues, with potential impacts on immunity. Although metabolic dysregulation could contribute, mechanical effects are indicated by our observations that micron-size LD’s are always round and rigid beads phenocopy LD’s across our multiple studies, starting with phagocytosis defects (Fig 2). Moreover, phagocytosis deficits are rescued by vertical compression of macrophages and by high loads of LD’s or beads in macrophages, where acto-myosin is activated by stretching (Fig 4D). In particular, we estimated cells are compressed to 60% heights, and so incompressible cells would exhibit a Poisson effect with an expanded cell diameter D given by (D/D_o)² * 0.6 = 1 that gives D = 1.3 D_o. This 30% stretching of the cortex will tend to activate myosin-II interactions with F-actin (Spinler, Shin et al. 2015, Schiffhauer, Luo et al. 2016). Such mechanosensitive activation seems understudied in phagocytosis, but mechanobiological responses are important in helping to explain for example the Frank-Starling mechanism of ventricle contraction upon filling of the heart (Park-Holohan, Brunello et al. 2021). Stretch or stress-activated actomyosin is also a basis for stiffness sensing, and phagocytosis occurs more readily for stiff targets coated with IgG (Beningo and Wang 2002, Sosale, Rouhiparkouhi et al. 2015, Jain, Moeller et al. 2019, Jaumouille and Waterman 2020, Vorselen, Barger et al. 2021).

Phagocytosis is a first line of immune defense, and its impairment may allow pathogens to accumulate and for disease to spread rapidly, thus making obesity a risk factor. Obesity is also a risk factor for diseases that develop more slowly such as cancer (Pati, Irfan et al. 2023). Furthermore, LD-induced nuclear rupture can lead to DNA damage and macrophage dysfunction or apoptosis as illustrated in aging lung (De Silva, Siewiera et al. 2023). More generally, mechanical perturbations of the macrophage cytoskeleton and nucleus by LD’s transcend individual pathologies and pose risks to general cell function. Better understanding of this biophysical type of dysfunction can perhaps lead to new strategies to help counteract pathological effects of fat on innate and acquired immunity but also in general.

Limitations of this study

Although we did not broadly study metabolic dysregulation of phagocytosis by LD’s as done by others (Guha Ray, Odum et al. 2023), ATP is central to metabolism and blebbistatin is an ATPase inhibitor of the major cellular protein myosin-II. At saturating doses of this drug, phagocytosis by LD-loaded macrophages was strongly suppressed under acute compression of cells though not in 2D (Fig 3B,C). Phosphorylation of the myosin light chain (p-MLC) also links to metabolism and is clearly modulated by LD’s as well as compression, but the non-monotonic effect of both LD’s and inert plastic beads (Fig 4) argues against LD metabolism as the specific, proximal effector of LD perturbations to phagocytosis. In other words, at least some acute metabolic effects of LD’s as well as confinement are likely to be general and upstream of the acto-myosin pathway shown here to regulate phagocytosis. Unfortunately, current methods for biochemical activation of non-muscle myosin II have multiple off-target effects, which motivates further study. In addition, the broadly studied topic of LD effects on macrophage polarization or differentiation in vitro (Strizova, Benesova et al. 2023) was not examined here. However, such polarized states are widely questioned in terms of their in vivo relevance (Bill, Wirapati et al. 2023), and no significant differences in established surface markers were detected her after LD loading (Fig 1).

Materials & Methods

Cell Culture:

THP-1 cells were cultured in RPMI 1640 medium (Gibco). Unless otherwise specified, all media were supplemented with 10% FBS (MilliporeSigma) and 1% Penn-Strep (Gibco) to obtain complete media. To obtain macrophages, THP-1 monocytes were differentiated using 100 nM phorbol 12-myristate-13-acetate (PMA, Sigma-Aldrich) for 48 hours in RPMI media. A549 cells were cultured in Ham’s F-12 medium (Gibco). HEK cells were kept in Dulbecco’s Modified Eagle Medium (DMEM, Gibco). B16-F10 cells were obtained from the American Type Culture Collection (ATCC) and cultured in RPMI. Bone Marrow Derived Macrophages (BMDMs) were collected from healthy donor mice, lysed with ACK buffer, and cultured on petri dishes in Iscove’s modified Dulbecco’s medium (IMDM, Gibco) supplemented with 20 ng ml⁻¹ recombinant mouse macrophage colony-stimulating factor (M-CSF, BioLegend) for 7 days with media refreshing every 3 days. During experiments, BMDMs were cultured in RPMI medium. All cells were kept incubated at 37°C with 5% CO₂.

Lipid Droplet Formation and Staining:

Tris HCl and Tris Base (Fisher Scientific) were dissolved in deionized water and mixed with fatty acid free bovine serum albumin (Sigma). The mixture was supplemented with oleic acid (Sigma Aldrich), mixed, and sterile filtered using a 0.2 μm filter (Fisher Scientific). The stock solution was mixed with RPMI to obtain a 400 μM final concentration. Bodipy dye (FL-C12 or 558/568) was subsequently added at a 1:2000 dilution and this media was added to differentiated THP-1 macrophages for 24 hours. Before experiments, all wells were rinsed and the media was replaced with complete RPMI. For BMDM loading, oleic acid was added at 1mM final concentration and incubated for 4 days.

Immunostaining:

Samples were fixed using 4% formaldehyde (Sigma) for 10-15 minutes, permeabilized with 0.5% Triton-X 100 (Sigma) for 10 minutes, and blocked with 5% bovine serum albumin (BSA) in PBS for 30 minutes at room temperature. Wash steps performed between all steps used 0.1% BSA. Cells were incubated with the primary antibody in 5% BSA overnight at 4°C. Antibodies used include: lamin-B1 (1:500, rabbit, Abcam), myosin-IIa (1:50, rabbit, Cell Signaling), CD80 (1:1000, rabbit, Abcam), CD206 (1:1000, mouse, BioRad), Phospho-Myosin Light Chain 2 (1:50, mouse, Cell Signaling) and γH2AX (1:500, mouse, EMD Millipore). Secondary antibody incubation (1:500; donkey anti-mouse or rabbit 488, 546, or 647; Thermo Fisher) was performed for 90 minutes at room temperature in 0.1% BSA. Addition of Alexa Flour 647 phalloidin (1:400, Invitrogen) in 0.1% BSA for 30 minutes was used to visualize actin. Cell nuclei were stained with Hoechst 33342 (Thermo Fisher Scientific) for 15 minutes in 0.1% BSA. After staining, samples were mounted for imaging or stored at 4°C in PBS.

For assays examining phagocytic uptake, secondary antibody incubation (1:500; donkey anti-mouse or rabbit 488, 546, or 647; Thermo Fisher) was performed immediately after fixation and 3x washing in PBS. Secondary antibody incubation was performed for 30 minutes. If further staining was needed, the sample would then be permeabilized and stained as described above.

Microscopy and Live Imaging:

Fixed sample imaging was performed on both epifluorescence and confocal microscopes. Lower magnification epifluorescence and brightfield images were obtained from an Olympus IX71 microscope with a 40x/0.6-NA objective and a digital EMCCD camera (Cascade 512B; Photometrics) with the exception of Fig 1B-i which was taken using a 150x/1.45-NA TIRFM oil immersion objective. Higher resolution confocal stacks were taken on either a Leica TCS SP8 system or a Zeiss LSM 980 system with a 63x/1.4-NA oil immersion objective. For mounting, fixed samples were placed between 2 glass coverslips, wet with 15 μL antifade mountant (Prolong Gold & Prolong Diamond, ThermoFisher), and allowed to set overnight in the dark. Fixed, wet samples were imaged within glass-bottom well plates (Mattek/Cellvis), dishes (Mattek), or chambered coverglasses (Ibidi). Live imaging during confinement experiments were performed at 37°C with 5% CO₂ using a Zeiss LSM 900 with Airyscan and a 40x/1.2-NA water immersion objective for 0.5-2 hrs. Images processing used Fiji software.

In vitro phagocytosis assays:

6 μm and 100 nm diameter beads:

THP-1 monocytes were differentiated to macrophages at a density of 2-3 x 10⁵ cells per well in a 6 well plate for 48 hours immediately followed by 24-hour incubation in oleic acid-containing media or complete RPMI. Differentiated BMDMs were plated at a density of 2 x 10⁵ cells per well in a 6 well plate and underwent OA incubation-or not- for 4 days before media replacement. Streptavidin coated polystyrene particles (Spherotech, SVP-60-5 and SVP01-008-5) were vortexed, centrifuged at maximum speed, and resuspended in PBS twice prior to opsonization with anti-streptavidin (1:500, mouse, Biolegend) for 30 minutes. For 6 μm beads, the beads were then washed twice, resuspended in RPMI, and added to cells at a ratio of 10:1 for a duration of 45 minutes. Cells were kept at 37°C with 5% CO₂ during the assay. After completion, detached beads were removed by rinsing twice with PBS before fixation and immunostaining as described above. Phagocytic uptake was quantified by presence or absence of secondary antibody stain on beads. For 100 nm beads, the beads were washed twice after opsonization and then stained red for visibility using PKH26 (Sigma) for 30 minutes. After washing and resuspension in RPMI, beads were then added to the cells at a 10:1 ratio and the plate was kept at 37°C with 5% CO₂ for 90 minutes. After this time, beads remaining in suspension were removed by 3x PBS rinse and cells were fixed, immunostained, and imaged as described above to quantify phagocytic uptake.

Cancer cells:

Sirpα KO THP-1 macrophages were differentiated and seeded at a density of 2 x 10⁵ cells per well in a 6 well plate for 48 hours followed by 24-hour incubation in oleic acid-containing or complete RPMI. Prior to the assay, THP-1 macrophages were labeled using Cell Tracker Deep Red dye (1:1000, Invitrogen) for 20 minutes before 1x wash and media replacement with RPMI. Td-Tomato CD47 KO A549 cells were tyrpsinized, rinsed with PBS, and antibody opsonized using anti-human CD47 (1:500, mouse, Santa Cruz) in 5% FBS in PBS for 1 hour at room temperature. After antibody washout, cells were counted and added to the plated macrophages at a 5:1 ratio. The plate was kept at 37°C with 5% CO₂ for 45 minutes before fixation and imaging.

Lentivirus:

THP-1 macrophages were differentiated and seeded at a density of 2 x 10⁵ cells per well in a 6 well plate for 48 hours. In the case of LD-loading, oleic acid-containing media was added for 24 hours as opposed to complete media. Lentivirus was produced from HEK293T cells as described below. Virus particles were added equally across all conditions and allowed to incubate for 1 hour before viral-containing media was removed and the sample was washed three times with RPMI media before media replacement in complete RPMI media. Cells were kept incubated at 37°C with 5% CO₂ for 72 hours prior to fixation and immunostaining. Phagocytic uptake was calculated as percentage of cells transduced to express GFP-sgRNA.

Bead preloading into Macrophages:

THP-1 macrophages were differentiated and plated at a density of 3 x 10⁵ cells per well in a 6 well plate for 48 hours prior to replacement of differentiation media to complete RPMI. Streptavidin-coated polystyrene particles (Spherotech, SVP-30-5) were vortexed, centrifuged at 16,000xg, and resuspended in PBS twice prior to opsonization with anti-streptavidin for 30 minutes. Opsonized 3 μm diameter beads were then centrifuged, rinsed twice, and added to macrophages for either 0.5, 4, or 7 hours to promote bead uptake. Samples were kept incubated for the duration of the bead preloading. Six RPMI washes were performed at the end of the chosen timepoint to remove most non-internalized beads. Prepared, opsonized 6 μm beads were then added to all conditions and a phagocytosis assay was performed as described above.

3 μm diameter beads under confinement:

Macrophages and 3 μm beads (Spherotech, SVP-30-5) were prepared as described in the above 6 μm diameter bead section. After resuspension of the opsonized beads, the beads were added at a high ratio (>20:1) for 15 minutes prior to the rapid confinement of macrophages to a height of 4 μm by introduction of micropillars. Cell confinement was maintained for 3 hours at 37°C and 5% CO₂ during which phagocytosis was permitted to occur. After 3 hours, confining micropillars were removed, cells were fixed, and immunostaining was performed as described above. For Blebbistatin-treated conditions, the micropillars were incubated overnight in drug-containing media for equilibration.

Production of Lentivirus for Phagocytosis:

HEK cells were plated at an initial density of 5 x 10⁶ cells in a 10 cm TC plate. Lentiviral packaging plasmids PAX2 and VSV-G were combined with the GFP-sgRNA transfer plasmid in serum free media with and supplemented with TransIT-Lenti (Mirus). After 30-minute incubation, the plasmid mixture was pipetted dropwise into the plate containing HEK cells and left at 37°C and 5% CO₂ for 72 hours. After 72 hours, the cell media was collected, and the supernatant was filtered using a 0.4 μM filter. To concentrate the lentiviral particles, PEG-it (Systems Bioscience) was used per manufacturer’s instructions. The concentrated solution was aliquoted and stored at −80 °C until used in an assay.

Preparation of Confinement Micropillars:

Polydimethylsiloxane (PDMS, RTV615-010) micropillars of the desired height were prepared and confinement was performed as previously described (Nader, et al., 2021). Briefly, plasma-treated 13mm glass coverslips were placed on a 10/1 w/w, PDMS/Crosslinker mixture on wafer molds with micropillars of the desired height. The setup was baked at 95°C for 20 minutes, and then the coverslips with the PDMS micropillars were removed from the wafer. Coverslips were passivated with pLL/PEG (SuSoS, PLL(20)-g[3.5]-PEG(2)) in 10mM HEPES buffer at room temperature for 1 hour and then incubated in medium overnight before confinement.

Confinement experiments:

THP-1 macrophages were differentiated with PMA for 48 hours at a density of 3 x 10⁵ cells per well in glass bottom 6 well plates (Mattek). Differentiated macrophages were treated-or not- with oleic acid containing media for 24 hours prior to confinement. On the day of confinement, large PDMS pillars, holding the coverslips with the micropillars, were attached to a modified 6 well plate lid. This confinement lid was then lowered onto the 6 well plate to push the micropillars on top the cells. Successful set-up to proper confinement height was confirmed visually by cell-micropillar contact using brightfield microscopy. When live imaging was performed, wells were imaged while still under confinement at noted time points. Immediately after pillar removal, cells were fixed and, when necessary, immunostained for further imaging.

THP-1 Transwell Migration:

Migration assays were performed using 24-well inserts with 3 μm or 8 μm diameter pores with polycarbonate membranes (Corning, Costar #3415). To encourage subsequent cell detachment for membrane replating, highly expressing CGAS+ THP-1 macrophages were differentiated on fibronectin-coated well plates. To construct the coating, Fibronectin (Sigma) was diluted in PBS to a concentration of 50 μg/mL and added to each well plate for 1 hour at 37°C before 2 PBS washes and macrophage plating for differentiation. For LD-loaded conditions, oleic acid-containing media was added for 24 hours in place of just complete RPMI after differentiation. In experiments using beads, preloading was performed as described above. LD-loaded and control macrophages were detached from well plates using 0.25% Trypsin (Gibco), counted, and 2 x 10⁵ cells were seeded on the tops of the membranes in complete RPMI media. Media containing 20% FBS was added to the bottom of the insert such that there was a slight nutrient gradient to encourage migration. The assay was kept incubated at 37°C and 5% CO₂ and allowed to proceed for 48 hours. At this point, cells were fixed, and membranes were detached from the insert for immunostaining as described above.

BMDM Transwell Migration:

Migration assays were performed using 24-well polycarbonate membranes with 3 μm or 8 μm diameter pores. When seeded on 8 μm pores, differentiated BMDMs were detached, and 2 x 10⁵ cells were seeded onto Transwell Tops. The macrophages were allowed to migrate for 24 hours prior to fixation, immunostaining, and mounting of the membranes for imaging. For 3 μm pore experiments, GFP-expressing B16 cells were tyrpsinized and sparsely added to inverted inserts to enable overnight attachment of ~2 x 10⁴ cells to membrane Bottoms. The low density of the B16s and constricting pore sizes eliminated cancer cell migration to membrane Tops while inverted while enabling them to serve as a chemotactic gradient for BMDMs. The next day differentiated BMDMs were detached, washed, and labeled using Cell Tracker Deep Red dye for 20 minutes before plating 2 x 10⁵ BMDMs onto Transwell Tops. Assays were permitted to run for ~24 hours before fixation, removal of membrane inserts, immunostaining, and imaging.

Cell Treatments:

For experiments requiring myosin inhibition, 25 μM of Blebbistatin (EMD Millipore) was added 1 hour prior to introduction of beads or confinement and kept in for the duration of the assay. Experiments involving actin inhibition were treated with 100 nM of Latrunculin-A (Sigma Aldrich) unless otherwise indicated (Fig 1D-iii) one hour prior to beginning the assay and kept in for the duration of the assay.

Macrophage Gene Editing:

Sirp-α knockout THP-1 cells were created using CRISPR-targeting whereas mcherry-cGAS expressing THP-1 cells were developed via lentiviral transduction.

Lentiviral transduction was performed as described in (Fischer, Boyer et al. 2022). Briefly, lentiviral particles were produced from 2x10⁶ HEK293T cells plated in antibiotic media. 24 hours after seeding, cells were transfected with 1 μg of VSV-G, 4 μg of psPAX2 packaging plasmids, and 5 μg of expression plasmid (mcherry cGAS), using Lipofectamine 2000 transfection reagent as directed by the manufacturer. After media replacement the subsequent day to remove lipofectamine, the cells were incubated at 37°C and 5% CO₂ for 48hrs to allow for viral particle production. THP-1 cells were transduced at a concentration of 1x10⁶ cells/ml in 2ml per well of a 6-well plate in THP-1 media without antibiotics. HEK293T viral media was collected, filtered using a 0.45 μm filter, and added to the THP-1 cells. Polybrene transfection reagent (EMD Millipore, TR-1003-G) was added to each well at a final concentration of 8ug/ml. Cells were centrifuged at 1250xg for 90 minutes at 30°C, then incubated for 1-2 days at 37°C with 5% CO₂. FACS was performed as described above to select for successfully transduced cells and the population of interest was expanded.

CRISPR knockout was performed using lentiCRISPR constructs as popularized by (Sanjana, Shalem et al. 2014). The lentiviral production procedure was performed as described above with the following differences: 2.5 μg of psPAX2 packaging plasmid and 5 μg of expression plasmid (Sirpa). To generate knockout, spinfection of THP-1 cells was performed in viral containing media during 2400 rpm centrifugation for 90 minutes at 30°C. Cells were resuspended in normal growth media to recover for 2 days before selection using 1 ug/ML puromycin. FACS was subsequently used to verify and collect cell exhibiting receptor knock out and the population of interest was expanded.

Actin and phosphomyosin localization analyses:

Actin analyses of LD-loaded vs control macrophages were conducted on individual slices of confocal z-stacks stained using phalloidin. The total intensity of peripheral actin was measured as the product of exterior actin area and mean intensity. The interface between the glass well and cell served as z=0 μm height and total peripheral actin was measured at ~1.5 and 3 μm heights from high z-slices. The same procedure was conducted on z-slices at heights of 0, 1, 2, and 3 μm to assess phosphomyosin and myosin-II immunofluorescence intensities.

Flow Cytometry and Cell Sorting:

For flow cytometry experiments, adherent cell lines were detached using 0.05% trypsinization, washed with PBS, and resuspended in FACS buffer (5% FBS in PBS). Cells grown in suspension were washed with PBS before resuspension in FACS buffer. If no antibody staining was required, samples and corresponding controls were run on a BD LSRII (Benton Dickinson) and subsequently analyzed using FCS Express 7 software (De Novo). If antibody staining was necessary, cells were resuspended in 5% BSA in PBS for 15 minutes at room temperature and subsequently washed and resuspended in 5% BSA mixture containing Fc block at a 1:500 dilution for 15 minutes. The samples were then centrifuged and resuspended in 5% BSA containing primary antibody (Sirp-α, 1:500, Santa Cruz) for 20 minutes. After washing using 1% FBS in PBS, the process was repeated for secondary antibody that was added at 1:500 dilution. FACS buffer was used as the final cell suspension. If fluorescence-activated cell sorting (FACS) cell sorting was required, cells were prepared the same way but kept sterile throughout the procedure. Sorting was performed on a BD FACS Aria II and the collected cells were washed with PBS and resuspended and replated in appropriate media upon return.

Western Blot:

Cells were collected and lysed. Proteins were separated using NuPAGE 4-12% bis-tris gels and transferred onto nitrocellulose membranes. Membranes were blocked in 5% BSA prepared in Tris-buffered saline with 0.05% Tween-20 (TBS-T) for 1 hour. Primary antibodies were then diluted in 5% BSA in TBS-T and incubated with the membranes overnight at 4°C. The following day, membranes were washed with TBS-T and incubated with HRP-conjugated secondary antibodies diluted in 5% BSA in TBS-T for 1 hour. Membranes were then scanned and processed.

Statistics:

Analysis and model fitting was conducted using Prism (Graphpad) or Sigma Plot. Unless otherwise indicated, plots display Mean ± SEM, and n indicates the number of cells, with N indicating the number of experiments. Statistical analysis was conducted using two-tailed Student t-tests when data passed normality tests. When data from multiple experiments were normalized to the control, one sample t-tests were conducted to test for statistical significance. Significance is indicated with a star where appropriate in a Figure if p<0.05.

Simple Physical Model for the Elastic Response of Phospholipid-Covered Interfaces:

A dense phospholipid monolayer forms at the oil–water interface of a lipid droplet (LD). Its equilibrium structure results from a balance between two competing interactions: (i) attraction between the polar headgroups and the aqueous phase, which minimizes the molecular area, and (ii) repulsion among the hydrophobic tails, which increases the area. The equilibrium molecular area, $a_0$, is established when these forces are balanced.

The interaction energy of a single molecule is

$${\mu }^i=\gamma a+\frac{k}{a},$$

where $\gamma$is the interfacial tension, $a$the molecular area, and $kis$a constant characterizing tail repulsion. Minimization with respect to $a$yields

$$\frac{d{\mu }^i}{da}=\gamma -\frac{k}{a^2}=0\Rightarrow a_0=\sqrt{\frac{k}{\gamma }}.$$

Substituting $a_0$gives

$${\mu }^i=\gamma a+\frac{\gamma a_0^2}{a}.$$

For small deviations $(a-a_0)$, Taylor expansion around $a_0$gives

$${\mu }^i\approx 2\gamma a_0+\gamma$$

where the first term is a constant offset, and the second term represents the elastic (dilational) energy of monolayer deformation. Expressed in standard quadratic form:

$$E_{elastic}=\frac{1}{2}{K}_a$$

with the area compressibility modulus

$$K_a=2\gamma .$$

Therefore, the monolayer behaves as an elastic film with a modulus twice the interfacial tension. Deviations from the spherical shape of a lipid droplet increase its surface area, storing elastic energy according to this modulus and giving rise to an effective surface tension $\gamma$.

Significance Statement:

• Macrophages accumulate lipid droplets (LD’s) which are generally considered in metabolism, but any direct physical impact of LD’s on phagocytosis is unstudied.

• LD’s have high effective interfacial tension, which makes small droplets sufficiently stiff to displace the cytoskeleton and suppress phagocytosis – except in high stress contexts of actomyosin activation.

• High interfacial tension of liquid-liquid phase separation can thus oppose cytoskeletal forces & function, but externally applied compression or internally applied stress with beads can also activate and rescue.