(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Macrophages self-generate and refine chemotactic gradients during migration towards complement C5a [1]

['Abhimanyu Kiran', 'Cruk Scotland Institute', 'Bearsden', 'Glasgow', 'United Kingdom', 'Division Of Cell', 'Developmental Biology', 'University College London', 'London', 'Peter A. Thomason']

Date: 2026-04 

Macrophages rely on efficient chemotaxis to locate sites of infection and tissue damage. One general strategy that enhances chemotactic accuracy is the use of self-generated gradients, where cells locally deplete attractants to create or sharpen guidance cues. Here, we show that mouse bone marrow–derived macrophages (BMDMs) migrate toward the complement component C5a using this strategy. Cells actively deplete C5a from their surroundings, establishing local gradients as fresh attractant diffuses inward. We visualized this process in real time with fluorescent C5a and reproduced its dynamics using computational models. C5a depletion is mediated primarily by C5aR1-dependent endocytosis. This mechanism produces complex responses, with different C5a concentrations inducing temporally distinct waves of migration, and maximal chemotaxis occurring below 10 nM C5a. As expected, increasing C5a concentrations recruit more cells. In contrast, human macrophages inactivate C5a mainly through carboxypeptidase-mediated degradation, yielding a higher optimal concentration (~30 nM) and distinct migratory dynamics. Both species also deplete externally-imposed C5a gradients, sharpening them and enhancing guidance. These findings identify C5a degradation as a critical mechanism by which macrophages extract directional information from their environment. Self-generated gradient formation, despite different mechanisms across species, emerges as a conserved and versatile strategy for immune navigation.

Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: RHI is on the Editorial Board. The other authors have declared that no competing interests exist.

Funding: This work was supported by the Wellcome Trust ( https://wellcome.org ; grant 221786/Z/20/Z to RHI), UK Medical Research Council ( https://www.ukri.org/councils/mrc/ ; grant MR/X000702/1 to RHI) and Cancer Research UK ( https://www.cancerresearchuk.org ; grant A31287 to the CRUK Scotland Institute and A1920 to EWR). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Copyright: © 2026 Kiran et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

We now demonstrate that mouse bone marrow-derived macrophages (BMDMs) employ self-generated gradients to steer toward C5a. The cells endocytose or degrade the C5a ligand while they respond to it, creating gradients even in initially uniform environments and sharpening preexisting gradients around the responders. This represents, to our knowledge, the first demonstration of self-generated chemotaxis in macrophages, and fills a missing piece in our picture of host defence.

Although C5a is known as a dominant chemoattractant for macrophages, it is unclear how macrophages navigate when C5a is abundant but not spatially structured. We therefore asked whether they, like other systems, could generate their own navigational cues. C5a biology makes it an especially suitable context to test self-generated guidance. It is generated rapidly and diffusely during complement activation [ 4 ], often creating high but poorly directional concentrations [ 17 ]. Macrophages express high levels of its receptor, C5aR1. At the same time, C5a is inactivated by multiple macrophage-expressed mechanisms, including endocytosis, protease-mediated degradation, and enzymatic trimming. These features suggest that macrophages are not merely passive responders but may actively remodel C5a fields to create navigational cues.

However, the gradients of these chemical signals at infection sites are not always well-defined. Attractants are often initially uniform or spread into shallow gradients by tissues. Consequently, to navigate effectively, cells must locally modify or degrade these constant concentrations to create steep gradients, facilitating their precise migration towards the source of infection. This phenomenon, known as self-generated chemotaxis, has been observed in various cell types, including various cancers, dendritic cells, and model organisms such as zebrafish and Dictyostelium [ 13 – 16 ].

Within tissues, macrophages constantly sample their environment using pattern recognition receptors (PRRs) to detect pathogen-associated molecular patterns. Once activated, they migrate directionally toward infection sites, guided by chemokines from other immune cells and complement factors such as C5a and C3a [ 11 , 12 ]. C5a is one of the most potent chemoattractants known for macrophages and also neutrophils. It is produced at sites of complement activation, a clear hallmark of tissue damage, by proteolysis from the much larger C5 molecule.

For immune cells, chemoattractants fall into two main categories: those produced by the target (e.g., formyl peptides like fMLP from bacteria) and those produced by the host, such as chemokines or complement-derived peptides like C5a [ 4 , 5 ]. Immune cells detect these gradients and efficiently migrate toward their sources [ 6 – 8 ].

Cell migration is a fundamental biological process involved in embryo implantation, development, cancer metastasis, wound healing, and immune cell function [ 1 , 2 ]. While cells can migrate randomly, it is extremely inefficient over most physiological scales; efficient migration in vivo is usually guided by directional cues from the environment. Among the best-understood guidance cues are diffusible chemicals—chemoattractants—which form gradients interpreted by receptors on the cell surface. The direction and steepness of the gradients provide steering cues that are interpreted through receptors, most often seven transmembrane receptors that couple to G proteins, yielding chemotaxis-cell migration directed towards the source of the chemoattractant [ 3 ].

Results

Effects of lipopolysaccharide activation Activation in the presence of bacterial or pathogen traces is a fundamental feature of macrophage biology. We therefore examined self-generated chemotaxis in LPS- activated and inactivated macrophages. The LPS binds and activates the Toll-like receptor 4 (TLR4) on the surface of the macrophage, triggering a signaling cascade that alters gene expression and leads to the production of inflammatory cytokines and mediators [24]. LPS stimulation was not essential for self-generated chemotaxis. Both activated and inactivated macrophages respond effectively to homogeneous C5a, and start to migrate at the same time. However, the inactivated macrophages chemotax less accurately than LPS-activated ones. Clearly, both activated and inactivated macrophages express a high enough level of C5a receptors to steer. The differences, therefore, more likely reflect changes in the pathways that connect gradients to migration and steering, or proinflammatory responses after activation (Fig 2). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. LPS stimulation of self-generated chemotaxis of BMDMs to C5a. BMDM migration in response to homogeneous C5a in the absence (A) or presence (B) of lipopolysaccharide (LPS) stimulation. In both conditions, cells established self-generated gradients and migrated across the bridge, but LPS-activated macrophages showed greater accuracy and coherence in their directional movement. Red dashed lines outline the regions occupied by migrating cells on the bridge. https://doi.org/10.1371/journal.pbio.3003728.g002

Dose dependence of self-generated C5a chemotaxis In other cell types, we have shown that the number of cells in the front wave of a self-generated gradient is approximately proportional to the attractant concentration [15,21]. We tested whether this holds for BMDMs using self-generated gradients formed from initial C5a concentrations of 0.3 nM, 3 nM, and 30 nM (S5 Movie). Figure 3A shows responses of the same initial cell density after 6 h. As expected, few cells entered the bridge at 0.3 nM C5a, because that is all that is required to inactivate or remove all of the C5a as it diffuses from the source well. With 3 nM C5a, the number of cells in the front wave was clearly greater (Fig 3A), but only by about 2-fold despite the 10× increase in chemoattractant. The same was true at the yet higher concentration of 30 nM C5a; after an initial delay for the cells to deplete the pool (see above), a larger wave still was formed (Fig 3A), but again with less than 3-fold further increase in cell number despite a 10× increase in attractant. Again, negative controls with no C5a gave negligible migration under these conditions (Fig 3B). Note also the slower response at higher concentrations as described earlier. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Dose dependence of self-generated chemotaxis to C5a. (A) Representative images showing BMDM migration after 6 h in initially homogeneous C5a concentrations of (i) 0.3 nM, (ii) 3 nM, and (iii) 30 nM. Red dashed boxes highlight the front wave of cells. (B) Negative control (no C5a), showing negligible migration. The red dashed line indicates the region assessed for cell migration. (C) Quantification of cells in the front wave at different C5a concentrations after 6 h. Data are shown as mean ± SE. Statistical significance was assessed using Student t test; *** indicates p < 0.0001. The data underlying this figure can be found in https://doi.org/10.5522/04/31211095. https://doi.org/10.1371/journal.pbio.3003728.g003 We would expect an approximately linear relationship between attractant concentration and the number of cells in the wave. The actual result suggests an extra level of complexity. One possibility is that cells are increasing their rate of breakdown at higher C5a concentrations, for example, proposed induction of extra breakdown enzymes at high ligand occupancy as a mechanism to increase robustness [21]. Another is that cells are communicating with one another through additional mechanisms than simple C5a depletion. To follow this, we examined hybrid agent-based computational simulations. These were sufficient to recreate the experimentally observed dose responses (S6 Movie). We also examined the effect of changes in initial cell density. These supported earlier observations that the number of cells in the front wave depends only on the attractant concentration [15]; cell density only affects the time before the wave is initialized (compare the time required for cells to migrate upon the bridge for the same C5a concentration in S2A, S2B, and S2C Fig).

BMDMs deplete C5a from their surroundings using endocytosis To perform self-generated chemotaxis, macrophages must get rid of the C5a in their local environment to form locally steep gradients. Two general mechanisms for this have been described—first, using enzymes that convert the chemoattractant into an inactive form [19,25], and second, by receptor-mediated endocytosis of the chemoattractant [13,26]. The enzymes may be membrane-bound with their catalytic domains exposed to the extracellular medium, like LPP3, or secreted, like phosphodiesterases. One strong candidate enzyme, carboxypeptidase M, is known to inactivate C5a in a complicated fashion by trimming off the C-terminal arginine [27], and is specifically expressed in human macrophages [28,29]. However, mouse macrophages such as our BMDMs do not express carboxypeptidase M (see, for example, the Immunological Genome Project–https://rstats.immgen.org/Skyline/skyline.html). We also examined C5a by mass spectrometry after incubation with BMDMs. The trimmed C5a (C5a-des-R) was not detected when mouse cells were incubated with C5a, whereas it was detected in all samples using THP-1 cultured human cells (cultured cells with macrophage-like gene expression and behavior, differentiated from a monocytic leukaemia line, as primary human cells were unavailable; S3 Fig). Interestingly, a hitherto undescribed double-trimmed form (C5a-des-GR) was also seen in the human and not mouse experiments. Mouse macrophages express a range of other secreted and membrane-bound proteases, but CPM is the only enzyme known to degrade C5a. To further distinguish endocytic from enzymatic degradation, we used commercially available C5a labeled with fluorescent AF647 at its N-terminus. Extracellular enzymes such as proteases do not change the fluorescence of this molecule when they attack it—trimming the C5a into an inactive form does not alter its fluorescence, so active and inactive molecules fluoresce the same. However, endocytosis should be visible as a loss of fluorescence from the medium and a sharp gain in vesicles inside the cell [30]. We visualized both medium and intracellular detail in BMDMs migrating up a self-generated gradient of fluorescently labeled C5a-AF647, examining both using a confocal microscope, after the cells had moved some distance across the bridge. The results are striking (Fig 4)—BMDMs internalize the labeled C5a, and remove it from the medium, creating a gradient. The chambers, which were set up with homogenous C5a, show a clear gradient after a few hours’ incubation with cells—low in the immediate vicinity of the BMDMs, becoming higher by 200 µm away. Soon after the chamber has been assembled, the macrophages contain punctate fluorescence, reflecting the endocytosed C5a and its breakdown products. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Receptor-mediated uptake of C5a by BMDMs. (A) Distribution of fluorescent C5a-AF647 after incubation without (i) or with (ii) BMDMs. (B) Quantification of fluorescence intensity along the bridge in the presence or absence of BMDMs, showing depletion of C5a near the cells. (C) Time course of C5a uptake. BMDMs were exposed to 5 nM C5a-AF647 for 1 h, washed, and imaged at 0, 1.5, and 3 h. Fluorescence localized to intracellular puncta and accumulated near the perinuclear region, consistent with endocytosis and lysosomal processing. Images were captured and processed using identical acquisition parameters. (D) Magnified view showing C5a-AF647 in vesicular structures 1.5 h after exposure. Cells prepared as in C. The data underlying this figure can be found in https://doi.org/10.5522/04/31211095. https://doi.org/10.1371/journal.pbio.3003728.g004 Thus, BMDMs set up their own chemoattractant gradients by endocytosing C5a from their surroundings, and retaining it, presumably for lysosomal breakdown. Endocytosed C5a could be recycled and re-secreted or targeted to lysosomes and broken down. To explore this mechanism, cells were seeded in Petri dishes at a similar density to the chemotaxis chambers, exposed to 5 nM of C5a-AF647 for 1 h, then washed with fresh medium and examined after different intervals (Fig 4C). This clearly shows the cells retain the C5a they take up. The fluorescent signal appeared to be in endocytic vesicles and eventually localized in the perinuclear region (Fig 4C), and was not released in significant quantities over three hours.

Comparison between human and mouse macrophages We hypothesized that the presence of CPM in human macrophages might alter the way they made and responded to self-generated gradients. Human macrophages could use the surface bound enzyme to get rid of C5a in their local environment, in contrast to the receptor-mediated endocytosis employed by the mouse macrophages. We tested this hypothesis several ways. First, we observed the change in the extracellular imposed gradient of fluorescent C5a-AF647 by the macrophages over time. Interestingly, although both types of macrophage could perform self-generated chemotaxis, only the mouse cells were able to affect the fluorescence profile of the C5a gradient (Fig 5A and 5B), demonstrating a role of receptor-mediated endocytosis in depleting C5a from the environment. For human macrophages (Fig 5C and 5D), the gradient of fluorescence remained similar, suggesting that the C5a was mainly being inactivated by the surface bound enzyme rather than endocytosed. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Different C5a inactivation mechanisms in mouse and human macrophages. (A) A gradient of fluorescent C5a (C5a-AF647) was established in an Insall chamber with and without mouse BMDM cells. (B) Measurement of fluorescent intensity along the bridge with and without BMDM cells, (C) Imposed gradient of fluorescent C5a with and without THP-1 cells, and (D) Measured intensity along the bridge with and without THP-1 cells, (E) Dose response of chemotaxis. The number of cells that had migrated 200 μm along the bridge was measured after 18 h for different C5a concentrations. (F) Spider plot of mouse and human macrophages in response to uniform 3 nM C5a. (G, H) Comparing the self-generated chemotaxis of mouse and human macrophages to 3 nM C5a showed an insignificant difference in their cosθ (G), but a significant difference in the mean speed (H). Bar in G and H = SE. ** represents p-value < 0.001. The data underlying this figure can be found in https://doi.org/10.5522/04/31211095. https://doi.org/10.1371/journal.pbio.3003728.g005 The different mechanism of establishing self-generated gradients would be expected to lead to different behaviors. In particular, the K d of receptors is typically much lower than the K m of enzymes, so receptor-mediated mechanisms typically work at a lower ligand concentration than enzyme-mediated ones. The dose response of both types of macrophages was assessed by counting the number of cells crossing a line 200 μm along the bridge after 18 h for various C5a concentrations. As predicted, the human cells steered most effectively at a higher C5a concentration (>30 nM versus 10 nM, Fig 5E). Tracking the chemotaxis of both types of macrophage, using 3 nM C5a so as to be below the saturation for both types of macrophage (Fig 5F), showed that human macrophages performed much better at the higher end of the gradient. This presumably reflects their greater ability to degrade the chemoattractant in their vicinity, forming steeper and more robust gradients compared to mouse macrophages, which can only internalize one C5a molecule at a time through receptor endocytosis and are thus limited by receptor availability at the cell surface. The collective behavior is also different—mouse macrophages show synchronized motion among themselves to form a clearly visible front wave (Fig 3A), whereas the human macrophages chemotax more independently (S7 Movie). Interestingly, the human macrophages can operate within a broader range of chemoattractant concentrations (Fig 5E), showing a clear response at very low concentrations as well as a better one at higher concentrations. This robustness is a predicted outcome of systems that use chemoattractant-degrading enzymes as well as receptor-mediated uptake [15,21]. Mouse and human macrophages behaved dissimilarly when performing self-generated chemotaxis to 3 nM of C5a (Fig 5F). Although they had a very similar time independent chemotaxis index (cosθ—see Fig 5G), the average speed of the human macrophages was twice that of mouse macrophages (Fig 5H). The speed of the chemotactic wave depends predominantly on the interaction between the Vmax of C5a degradation and its diffusion, rather than the inherent speed of the cells, so fast migration of human cells is consistent with rapid degradation of C5a, again consistent with the involvement of surface bound enzymes.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3003728

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/