1. Diffusion of Chemical Signals
An important concept in understanding the mechanism of localised intracellular signalling is
diffusion. Diffusion is, of course, the net movement of ions or molecules from a zone of higher concentration to one of lower concentration. Thus, if an intracellular signal is generated within the cell at the contact point between the phagocytic target and the plasma membrane of the phagocyte, its intracellular concentration will be higher at that point than elsewhere in the cell and diffusion of the signal molecules away from the point of contact will occur. Diffusion is, of course, the result of the random Brownian motion of individual ions or molecules, causing the initial concentration of signals to be “diluted out” as molecules move away from a source. Although random at an individual ion/molecule level, the net effect occurs in a predictable way and the concentration can be calculated at any point distant from the source as it declines with passing time
[1]. The decrease in concentration of a molecule depends on its distance from the source of the signal (the contact between the target and the cell), the time elapsed (t), and its diffusion constant (D, μm
2/sec).
The equations which govern these relationships, especially in two and three dimensions, can look daunting
[1], but the main parameter which governs the process is the diffusion constant, D, a measure of “diffusivity” of the molecule or ion in a given environment. A simple, but useful, approach for estimating how a localised focal signal may be restricted is to calculate its “diffusion length” given by:
2. Theoretical Basis of Localisation of Ca2+ Signalling
Phagocytosis by opsonised particles is accompanied by Ca
2+ influx from the extracellular environment generating intracellular Ca
2+ signalling
[2][3]. The earliest report of phagocytosis by neutrophils showed that Ca
2+ in the medium was an absolute requirement for phagocytosis
[4]. However, an intracellular Ca
2+ signal may not always be obligatory, and there are early reports using nonopsonised particles in which no Ca
2+ signals were found
[5][6]. Indeed, the phagocytic cup can partially form without any Ca
2+ signal
[7][8]. However, when a critical number of opsonin receptors have been engaged (often about 50% of the cup), Ca
2+ influx is triggered, causing a large cytosolic Ca
2+ signal
[7]. This Ca
2+ signal accelerates the rate of completion of phagocytosis and probably accounts for the increased phagocytic rate caused by opsonisation observed in cell populations. The flattening and spreading of neutrophils onto surfaces, which may be considered as “frustrate phagocytosis”, is also accompanied by obligatory large Ca
2+ signals
[9][10]. The Ca
2+ signal can be thought of as being the permissive signal for efficient phagocytosis.
As suggested earlier, in considering the question of whether this Ca
2+ signal could be localised to the site of phagocytosis, a useful parameter is the “diffusion length” which indicates how far from the source of the signal Ca
2+ ions would travel by a defined time. The diffusion length can be calculated from knowledge of the diffusion constant, D, as 2√(Dt. The diffusion constant for Ca
2+ in water, D
Ca, has been measured as 7.7 × 10
−6 cm
2/s
[11]. i.e., approximately 800 μm
2/s. The diffusion length of Ca
2+ ions after 100 ms (0.1 s) would thus be 2√(800 × 0.1) = 18 μm. As this is larger than the diameter of many phagocytes (e.g., neutrophils are about 10 μm), it suggests that Ca
2+ ions are unlikely to act as a localised phagocytic signal. However, the diffusion constant for Ca
2+ (D
Ca) in extracted cytosol is lower and has been measured, using Ca
45 at 530 μm
2/s
[12], 223 μm
2/s in cytosolic extract from Xenopus oocytes
[13], or in living cells using fluorescence correlation spectroscopy at 650 μm
2/s
[14]. Thus, in the cytosol, the diffusion length of free Ca
2+ is about 9.2–16 μm, which is still too large to generate localised Ca
2+ signals. However, in the cytosol, Ca
2+ ions bind reversibly to Ca
2+ buffering agents, such as proteins and other molecules. It is estimated from the association rate (“on rate” or K
on) and buffer concentration that the lifetime of a free Ca
2+ ion would be in the order of 10
−5s
[13], meaning that the free unbuffered Ca
2+ signal would be extremely localised. However, the buffer molecule is also mobile and will diffuse, carrying Ca
2+ with it until the ion is released. It is, thus, more appropriate to consider the diffusion of buffered Ca
2+ ions. The measured diffusion constant for buffered Ca
2+ is higher than for free cytosolic Ca
2+ and the diffusion length of the buffered Ca
2+ is extended from d1 to (d1 + d2). For Ca
2+ -binding proteins with the EF hand-motif (e.g., calmodulin, calbindin, etc.), the dissociation rate (k
off) is about 10 s
−1, giving a “lifetime” of the Ca
2+–buffer complex of 0.1 s (100 ms). The diffusion constant of the buffered Ca
2+ has been measured in cytosol extract to be 13 μm
2/s. After 100 ms (when the Ca
2+ ion is released), the diffusion length is only about 2.3 μm
[13]. Assuming that the released Ca
2+ ion is free for about 1 s before being taken up into immobile Ca
2+ stores, the “effective range” of buffered Ca
2+ is about 5 μm
[14]. Although these measurements were made using cytosol extract with a Ca
2+-buffering content of about 300 μM
[13], the Ca
2+-buffering ratio of neutrophil cytosol is similarly high, about 1000-3000:1
[15][16], meaning that for every free Ca
2+ ion in the cytosol, there are 1000-3000 bound Ca
2+ ions, so the conclusion can be extrapolated to neutrophil cytoplasm. The diffusion length predicted can also be experimentally confirmed in neutrophils by osmotically lysing Ca
2+-containing lysosomes to release a small localised Ca
2+ signal in the cytosol. This is achieved by the use of a dipeptide glycyl-l-phenylalanine 2-naphthylamide (GPN), which is cleaved within cathepsin C-containing lysosomes and is widely used to osmotically rupture the lysosomal membrane and release its Ca
2+ content into the cytosol. In neutrophils, this procedure generated “Ca
2+ puffs” with diameters of about 3 μm
[17], similar to those expected from the theoretical diffusion length calculation for buffered Ca
2+. This suggests that the localisation of Ca
2+ signals during phagocytosis is at least theoretically possible in neutrophils and probably other phagocytes. However, as the influx of Ca
2+ from the extracellular environment is the important Ca
2+ source for phagocytosis, understanding the mechanism for generating the Ca
2+ signal is crucial.
The diffusion length of free Ca2+ is very small (d1), as Ca2+ ions are rapidly bound to proteins (buffers). However, the buffer also diffuses (d2), carrying the Ca2+ ion to be released at a different location. The time between the binding and release of the Ca2+ ion depends on the dissociation rate (koff or off rate), during which time the buffered Ca2+ travels the distance d2. Ca2+-binding proteins have koff rates of about 10 s−1, giving a “lifetime” of 0.1 s. Hence, the diffusion length of the buffered Ca2+ is extended from d1 to (d1 + d2).
The classic initiation mechanism for Ca
2+ signalling is via IP
3, which is produced by the action of a phospholipase C on phosphatidylinositol 4,5-bisphosphate (PIP
2)
[18]. The two classes of “opsonin receptors” both activate PLCγ through tyrosine kinase receptors. These include the Fc receptors, triggered by antibody-opsonised targets and β2 integrin, triggered by complement (C3bi)-opsonised targets
[19]. PLC β in neutrophils is also activated by G protein-coupled receptors for soluble agonists, such as formylated peptides. The activation of either PLC generates IP
3, which diffuses from the plasma membrane away from the initiation site to engage IP
3 receptors on Ca
2+ storage sites throughout the cell. This releases Ca
2+ from the Ca
2+ stores, which in turn signals plasma-membrane-channel opening through a mechanism known as “store operated Ca
2+ channel” (or SOCC) opening
[20][21]. Unlike Ca
2+, IP
3 has a long effective range
[13] of 24 μm (having both a higher diffusion constant, 283 μm
2/s, and having no buffers). Thus, for cells smaller than 24 μm, IP
3 will fill the cell and act as a global messenger, thus losing any directional information as to the locus of its formation at the phagocytic site.
Route (a) shows the effect of direct coupling of the stimulus to the opening of Ca2+ influx channels (here supposed to be the transient receptor potential cation channel, subfamily M, member 2, or TRPM2). Ca2+ influx from that site, with limited diffusion distances, would retain information as to its source. In contrast, (b) is the signal for Ca2+ via PLC. The generation of IP3 with its extensive diffusion would result in the opening of Ca2+ channels from remote Ca2+ storage sites in the cell, which would open store-operated Ca2+ channels (SOCC) in the plasma membrane at all loci and so lose directional information.
The mechanism by which store-operated Ca
2+ Channels (SOCC) open when Ca
2+ is depleted from the endoplasmic reticulum may involve the Orai1 and STIM1 system
[22][23][24]. There is considerable evidence that the Orai–STIM system also operates in phagocytes
[2][25][26] and is responsible for hotspots of high Ca
2+ at the formed phagosome–cytosol boundary
[27].
A Ca
2+ influx mechanism which may not involve the depletion of Ca
2+ stores or IP
3 generation is the opening of TRPM2 channels (transient receptor potential cation channel, subfamily M, member 2), Ca
2+-permeable channels which are highly expressed in phagocytic immune cells
[28][29][30]. These plasma membrane channels, previously called LTRPC2 or TRPC7, are permeable to both Na
+ and Ca
2+ and may be involved in immune-cell behaviour including adhesion and migration
[31]. TRPM2 is activated or directly opened by ADP-ribose (ADPR) as well as by oxidants (e.g., H
2O
2)
[32]. There is evidence that neutrophils (and probably other phagocytes) utilise these Ca
2+-signalling pathways
[33][34][35]. Although the mechanism of activation of the channel or of the generation of ADP-ribose is not fully established, oxidants are known to open the channel directly and also to enhance the opening by ADPR
[32]. As oxidants are generated locally within the phagosome, they may play a role in local TRPM2 opening through H
2O
2 signalling. The key point, however, is that since TRPM2 activation does not involve IP
3 generation, there is at least a theoretical possibility that an opsonin-triggered Ca
2+ signal could remain localised to within a few microns of the phagocytic contact site if TRPM2 channels were involved.
3. Localisation of Ca2+ Signalling of Phagocytosis
An early paper
[36], using a single-wavelength Ca
2+ reporter, Quin2, reported that Ca
2+ signalling was highest in the pseudopodia and periphagosomal cytoplasm. Although fura2 was reported to give a similar result
[37], persistent, localised, elevated Ca
2+ during pseudopodia formation and the onset of phagocytosis cannot generally be confirmed using modern dual-wavelength or confocal imaging of a number of probes. The reported Quin2-elevated fluorescent signals in pseudopodia (i.e., prephagocytic structures) were probably the result of an optical effect whereby fluors are more efficiently excited and emission-detected in the less light-scattering cytoplasm of pseudopodia
[38]. When followed in time, the Ca
2+ signal which accompanies phagocytosis, although presumably triggered locally, causes a global elevation in cytosolic Ca
2+ (14,16, 45). In one study
[39], the question was addressed of whether the presence of the Ca
2+ indicator itself affected the Ca
2+ distribution by being an additional Ca
2+ buffer with high mobility. The Ca
2+ indicator (fluo4) was, therefore, chemically coupled to a large protein with low mobility and microinjected into neutrophils before phagocytosis. This indicator also failed to detect localised Ca
2+ signalling. How, then, can Ca
2+ signal the localised events that precede and trigger phagocytosis?
With this in mind, it is relevant that one of the many cytosolic targets of an elevated Ca
2+ level has a very low affinity for Ca
2+, the Ca
2+-activated protein calpain. There is increasing evidence that calpain activity is involved in the formation of phagocytic pseudopodia
[40]. For example, pharmacological inhibition of calpain slows phagocytosis by arresting the rapid phase of internalisation
[7][9]. The outcome of calpain activation is the cleavage of ezrin, a protein which “staples” the cortical actin network to the plasma membrane
[41], maintaining cell-surface wrinkles, which are a reservoir of extra membrane required for phagosome formation
[41][42][43], and also regulates the cell-surface membrane tension
[42][43][44]. After ezrin cleavage, there are fewer cell surface wrinkles as measured by SEM
[45] or sub-domain FRAP (fluorescence recovery after photobleaching)
[46] and, consequently, more membrane available for the formation of the phagosome
[39]. Using constructs of ezrin with fluors attached to either side of the cleavage site (each with different excitation and emission properties), the cleavage of ezrin could be monitored by separation of the two fluors
[47]. This method demonstrated that ezrin cleavage occurred in the phagocytic pseudopodia as it encroached around the target for internalisation
[47]. The release of the plasma membrane from the actin cytoskeleton has also been observed in other cell types forming “cell protrusions”. Membrane-proximal F-actin, i.e., actin, held close to the membrane by crosslinking proteins was reduced at the loci of these cell protrusions in RPE-1 and HUVEC cells
[48] and ezrin was decreased at similar protrusion sites in osteosarcoma cells
[49]. Furthermore, in response to a photoactivatable ezrin phosphatase used experimentally to reduce the effectiveness of ezrin linking, the cell protrusion rate was increased, suggesting the release of a brake of the protrusion system
[49].
While most Ca
2+-activated cytosolic signalling proteins have kds between 0.1 μM and 1 μM, calpain-1 requires a Ca
2+ level of 50–100 μM for activation
[50]. For example, calmodulin, the ubiquitous Ca
2+-regulated protein, has a kd for Ca
2+ of about 0.3 μM, meaning that at the resting cytosolic Ca
2+ level of 0.1 μM, calmodulin is not active, but when cytosolic Ca
2+ rises to 1 μM, calmodulin becomes almost fully active. In contrast, calpain-1 has a kd for Ca
2+ of about 30 μM, so that calpain in the bulk cytoplasm will remain inactive during the Ca
2+ rise to 1 μM. The consequence of this is that calpain can only be activated very closely to the open Ca
2+ channel or in a cell location where such a high Ca
2+ level may be reached. One such location is within wrinkles of the plasma membrane, structures such as microridges. Phagocytes by necessity have such structures as these provide the membrane reservoir needed for the formation of the phagosomes
[39][45][46]. It has been estimated that the neutrophils’ surface area has an excess of membrane, two to three times that required simply to enclose the cell volume
[39][44]. From the modelling of Ca
2+ influx into a confined volume of cytosol, it has been estimated that the Ca
2+ concentration in the cytosol within the wrinkled membrane can reach hundreds of micromolar
[51]. This is because open Ca
2+ channels allow Ca
2+ to enter the small volume of cytosol within the wrinkle at a faster rate than it can diffuse into the bulk cytosol through the “mouth” of the wrinkle
[41][51]. Once the local Ca
2+ buffers in the wrinkle space are overwhelmed, the free Ca
2+ concentration can rise to very high levels. The intra-wrinkle Ca
2+ has been recently measured using a construct, called EPIC3, from a low-affinity genetic Ca
2+ indictor, CEPIA3 (kd for Ca
2+ ≈ 11 μM), coupled to ezrin, which locates within the wrinkles and near the plasma membrane
[52]. Upon phagocytosis, the intra-wrinkle Ca
2+ was estimated to rise to 30–80 μM
[52], sufficient to activate calpain, release ezrin, and uncouple the plasma membrane from the underlying actin cytoskeletal network
[47][52].
4. Localisation of Ca2+ Signals after Phagocytosis
In contrast to the role played by Ca
2+ in initiating phagocytosis, after a closed phagosome has formed, elevated periphagosomal Ca
2+ may occur. It is possible that the phagosome itself provides the localised source of Ca
2+ for this effect. This periphagosomal Ca
2+ signal has been reported using the dual-excitation-wavelength probe, fura 2
[53], and more recently using the genetic Ca
2+ reporter, GCaMP3, where a long-lived halo of elevated Ca
2+ was detectable lasting over 6 min after phagocytic-cup formation
[54]. Reported focal hotspots of Ca
2+ (Ca
2+ puffs) near phagosomes have also been visualised
[27], which aid lysosome–phagosome fusion and also “boost phagocytosis” by recruiting Ca
2+ stores near phagosomes
[55].
The ion channels through which Ca
2+ ions leak into the cytosol from the phagosome are probably TRPM2. These have been shown to be on the phagosomal membrane
[56] and can be experimentally opened via the intraphagosomal microinjection of ADPR or H
2O
2 [56]. Since the intraphagosomal environment is oxidising (as a result of the NADPH oxidase) and contains H
2O
2 (the substrate for phagosomal myeloperoxidase), it seems likely that the channels are open and generate the elevated periphagocytic Ca
2+ and periphagosomal Ca
2+ hotspots observed. The roles of TRPM2 opening on the plasma membrane and on the phagosome are unclear, but since TRPM2 deletion increases susceptibility to polymicrobial sepsis without reducing phagocytosis itself
[57], it would seem more likely that they have a role in postphagocytic events (e.g., lysosome fusion), rather than in signalling phagocytosis. It is obvious, in any event, that since these are postphagocytic events, they cannot be involved in signalling the phagocytosis event.