Oscillatory Membrane Potential Changes in Cells of Mesenchymal Origin: The Role of an Intracellular Calcium Regulating System

A number of cell types of mesenchymal origin exhibit a characteristic excitability. This excitability is distinct in some regards from that studied so extensively in nerve and muscle, but it may also share important features with that of these classically excitable tissues. The mesenchymal cells in which this excitability has been demonstrated thus far include fibroblasts growing in primary culture and continuous cell lines of fibroblastic origin (Nelson, Peacock & Minna, 1972), mouse, guinea-pig and human macrophages (Gallin et al. 1975) and guinea-pig megakaryocytes (Miller, Sheridan & White, 1978). Intracellular recordings have revealed that these cells respond to a variety of stimuli with large membrane hyperpolarizations lasting several seconds. These hyperpolarizations, which are associated with large increases of membrane conductance, have been termed hyperpolarizing activation or H.A. responses. Such hyperpolarizations may occur spontaneously or can be elicited by electrical, mechanical or chemical stimuli. In confluent cultures of L-cells the responses can be transmitted from one cell to another. Hyperpolarizing activations may also occur spontaneously as repetitive trains in an oscillatory manner, or an evoked H.A. in a previously silent cell may be followed by a series of hyperpolarizing membrane totential oscillations lasting for several minutes (Okada et al. 1977a).

In this review we will discuss the surface membrane permeability changes responsible for the H.A. responses, summarize evidence relating to the mechanisms underlying these permeability changes and will propose a model incorporating these mechanisms, which can account for both single H.A.’s and rhythmic oscillatory membrane potential behaviour in cells exhibiting this form of excitability. Finally, based on clues provided by the forms of stimuli that elicit H.A. responses, we will discuss some possible functional implications of this form of excitability in mesenchymal cells.

The hyperpolarizing increase in membrane conductance which constitutes the H.A. response is illustrated in Fig. 1. Results obtained from L-cells are generally representative, and we will describe them in most detail. L-cells are a fibroblastic cell libe initially adapted to culture in 1948 (Sanford, Earle & Likely, 1948). They normally have a doubling time of some 20 h, but can be caused to stop dividing and to increase in size by X-irradiation (Whitmore et al. 1958; Nelson & Peacock, 1973). The large cells so obtained are quite suitable for electrophysiological studies involving intracellular microelectrode recording and stimulating techniques (Nelson & Peacock, 1972, 1973). Such recordings have shown that the cells have a relatively low resting membrane potential (−15 to −20 mV) and a large membrane input resistance (20–30 MΩ) and time constant (10-15 ms). The resting membrane potential is relatively insensitive to changes in extracellular K⁺ concentration and this, in conjunction with flux measurements (Lamb & MacKinnon, 1971a, b), indicates that the permeabilities of the L-cell membrane to Na⁺ and K⁺ are roughly equal at rest. Stimulation of the L-cells by a variety of means can elicit a 20-30 mV hyperpolarization of the membrane lasting for several seconds; during this hyperpolarization the surface membrane conductance is increased 2- to 10-fold. The H.A. may also occur spontaneously and repetitively (Fig. 2) (Nelson et al. 1972) and, as described in detail by Okada and co-workers (Okada et al. 1977a, b, 1978), represents an oscillatory mechanism in L-cells.

The H.A. response is conveniently elicited by a brief (∼ 100 msec) hyperpolarizing current pulse of 10–50 nA. Even though these are large current pulses, it appears that the effective stimulus is a voltage gradient across the membrane since in macrophages which have much higher input resistances, the current required to evoke an H.A. is an order of magnitude smaller. Large depolarizing pulses may also elicit the response, but at considerably higher threshold, if at all. In macrophages, depolarizing pulses occasionally evoked H.A. in cells when hyperpolarizing pulses did not (Gallin et al. 1975). When one H.A. has been elicited, the threshold for eliciting a second response is typically increased for several seconds. However, repetitive H.A.S can occur spontaneously or in response to single stimuli as will be discussed below. When twol adjacent L-cells are impaled, an H.A. elicited in one cell frequently is followed, after a latency of a few seconds, by an H.A. in the adjacent cell. L-cells are sometimes coupled electrically, but propagation of the H.A. occurs between cells which are not directly coupled electrically (Nelson & Peacock, 1973). Gentle mechanical stimulation provided by touching the cell surface some distance away from the impaling micropipette is a very effective means for eliciting H.A. responses. On the other hand, L-cells that are spontaneously oscillating are sometimes made to stop by mechanical stimulation (Okada et al. 1977a). We have found that iontophoretic application of acetylcholine was effective in some L-cells in producing H.A. responses and that this could be blocked by atropine but not curare (Nelson & Peacock, 1972). In experiments using L-cells from another source (Okada et al. 1977a) such cholinergic chemosensitivity was not found, but these cells were responsive to other chemical stimuli (Tsuchiya, Okada & Inouye, 1978). Macrophages produce large hyperpolarizations, sometimes followed by oscillations, in response to serum components and artificial peptides known to function as chemotactic stimuli (Gallin & Gallin, 1977).

When steady polarizing currents are passed across the L-cell or macrophage membrane to alter the membrane potential and the H.A. is elicited at different steady membrane potential levels, a reversal potential for the H.A. can be demonstrated. In normal (5-4 HIM) K⁺ solutions this is about −80 mV in L-cells (Nelson et al. 1972; Okada, Tsuchiya & Inouye, 1977b) and − 54 mV in macrophages (Gallin et al. 1975). The H.A. reversal potential is markedly affected by changes in the external K⁺ concentration (Fig. 3) but is very little affected by changes in external Na⁺ or Cl⁻ concentrations (Okada et al. 1977b). Okada et al. (1977b) have also shown that oscillating H.A.S are insensitive to tetradotoxin, which blocks electrically excitable Na⁺ channels in many tissues (Narahashi et al. 1960), but are abolished by tetraethyl ammonium (TEA), which blocks certain K⁺ channels (Hagiwara & Saito, 1959). These experiments leave little doubt that the H.A. response consists of a large increase in the surface-membrane permeability to K⁺ ions.

A rise in the free intracellular Ca²⁺ concentration has been shown in a number of cell types to produce an increase in the surface membrane K⁺ permeability (Romero & Whittam, 1971; and reviewed by Meech, 1978). This appears to be the case in L-cells also; injection of calcium ions intracellularly produces a membrane hyperpolarization accompanied by an increase in surface membrane conductance (Nelson & Peacock, 1973; Henkart & Nelson, 1979; Okada et al. 1978, abstr.). Brief pulses of Ca²⁺ injections produced graded responses with a time course closely similar to the H.A. response (Fig. 4). The question of the mechanism of the H.A. response then becomes, what is the source of the calcium which produces the surface membrane permeability change?

An obvious possibility is that ingress of calcium ions from the external medium across the surface membrane accounts for the rise in intracellular calcium which generates the H.A. response. Gallin et al. (1975) showed that 1·5 mM EGTA (which chelates Ca²⁺) in the extracellular medium abolished both spontaneous and evoked H.A.S in macrophages. A23187 (an ionophore facilitating divalent ion penetration through membranes (Reed & Lardy, 1972)) by contrast, produced prolonged membrane hyperpolarizations. In this latter circumstance, if EGTA were added to the medium in the presence of the ionophore repolarization of the cell promptly occurred. Okada et al. (1978) also reported that external Ca removal abolished oscillating H.A.S in L-cells. Our results from experiments involving altered external Ca²⁺ ion concentrations, however, have been complex. We observed in L-cells that were transferred to and allowed to remain for some time in Ca-free, EGTA-containing medium before recording was begun that H.A.S appeared to be abolished. Occasionally, however, an H.A. could be elicited or even occurred spontaneously after several hours in Ca-free medium. Upon more careful examination of the results of these and other experiments we have been led to consider the release of intracellularly sequestered calcium as a source for the calcium involved in hyperpolarizing responses in L-cells (Henkart & Nelson, 1979). We summarize this evidence here.

• There is no electrophysiological evidence for obligatory Ca²⁺ ingress preceding the H.A. If an increase in Ca²⁺ permeability preceded the H.A. and Ca²⁺ ions entered the cell from the external medium, some depolarization of the membrane should be evident. Although this is occasionally seen in macrophages, megakaryocytes and L-cells, it is not generally seen in L-cells and is certainly not required preceding H.A. generation. An inhibitor of voltage-sensitive calcium channels, D-600, has no effect on the H.A.

• Cobalt ions block the H.A. response, but this requires rather high concentrations and prolonged incubations (e.g. 5 mM-Co²⁺ requires approximately 4 h to block). At a period of less than 1 h even in 15 mM cobalt, large H.A. responses were seen in some cells.

• Zero-Ca²⁺, EGTA-containing solutions reduce the proportion of cells showing H.A. responses (Table 1), but in those cells showing responses, the amplitude of the responses is not significantly reduced (Table 2). We have observed large H.A. responses in L-cells as much as h after switching cultures to balanced salt solutions containing no calcium and with 0·5–1·5 mM EGTA.

  Table 1.

  Effect of removing external calcium on H.A. responses (taken fromHenkart & Nelson, 1979)

  

  

  View Large
  Table 2.

  The effect of varying the external concentration of calcium on the amplitude of H.A. responses

  

  

  View Large

• The amplitude of the H.A. response is not increased in high (10 mM) Ca²⁺ solution (Table 2).

We feel that these results are incompatible with extracellular calcium as the sole or even predominant source for the calcium which generates the H.A. responses in L- cells. Parallel morphological studies revealed a possible basis for coupling between stimuli delivered to the surface membrane and a potential intracellular store of releasable calcium (Fig. 5). In L-cells the endoplasmic reticulum (ER) forms morphologically specialized appositions with the surface membrane which are similar to appositions between the sarcoplasmic reticulum and the transverse tubular system or surface membrane at diads or triads in striated muscle. In muscle these specialized appositions are thought to be the sites of coupling between the surface membrane and the intracellular calcium-releasing system involved in muscle contraction, the sarcoplasmic reticulum. Similar appositions between the ER and surface membrane have been described in neurones (Rosenbluth, 1962; Henkart, Landis & Reese, 1976) where the ER has also been shown to be a Ca-sequestering system (Henkart, Reese & Brinley, 1978).

Despite our conclusion that intracellular release of calcium is involved in the generation of the H.A. response it is clear that removal of extracellular Ca²⁺ has a strong impact on the H.A. mechanism as indicated by the decrease in the proportion of cells in which H.A.S can be evoked. We postulate that the functioning of the coupling mechanisms between the surface membrane and the internal calcium sequestering (and releasing) system may be dependent on or affected by extracellular Ca²⁺ and is eventually inactivated in the absence of extracellular Ca²⁺. We propose the concept of ‘permissive’ extracellular calcium to describe this phenomenon. It should be noted that the possibility of such a situation implies that caution should be exercised in interpreting the results of experiments involving low Ca²⁺ solutions. Effects seen in such experiments may be due to the removal of Ca²⁺ as a charge-carrying species crossing the surface membrane, but we feel that possible effects of low Ca²⁺ solutions on the functioning of intracellular systems should be considered as well.

The combined functional and morphological observations described above have led us to suggest the model shown in Fig. 6. The proposed sequence of events involved in an elicited H.A. response are: (i) stimulation delivered to the surface membrane is coupled by some (2) ‘permissive Ca²⁺’-dependent step to (3) release of Ca²⁺ from the subsurface cisterns and/or (4) other regions of the endoplasmic reticulum. This released calcium affects (5) the surface membrane in such a manner as to increase its potassium permeability. The cycle is terminated (6) by Ca²⁺ removal from the cytoplasm by uptake into ER or mitochondria or extrusion through the surface membrane. This is analogous to the sequence of events in muscle where myofibril contraction is the major effect of Ca²⁺ released from the SR. We discuss below other possible physiological consequences.

Evidence supporting one aspect of the above model has been provided by studies of Ca²⁺ uptake in fibroblast microsomes (Hoffman-Berling, 1964; Moore & Pastan, 1977a, b). A non-mitrochondrial subcellular fraction, probably distinct from surface membrane vesicles is capable of accumulating calcium; the source of the microsomal fraction is probably ER membranes, which suggests that the ER system would be able to maintain an elevated Ca²⁺ concentration within its lumen.

As stated earlier, L-cells and macrophages may exhibit regular oscillations in membrane potential and conductance; the cycle of these oscillations resembles in most regards the H.A. response described above. (An interesting difference is the greater sensitivity of oscillations to metabolic inhibitors (Okada et al. 1977a).) What might be the basis for such oscillatory behaviour?

The most obvious possibility would assign the oscillatory mechanism to the surface membrane. One such hypothesis would be that a relatively high resting conductance to calcium at the low resting potential of mesenchymal cells would lead to the activation of the hyperpolarizing K⁺ conductance; the hyperpolarization would inactivate the Ca²⁺ conductance and calcium extrusion would lead to a decrease in K⁺ conductance, with consequent return to the resting membrane potential and reactivation of the Ca²⁺ conductance.

Given the evidence for the involvement of internal Ca²⁺ stores, however, we suggest an alternative or additional possibility for the generation of oscillatory behaviour, based on the parameters of calcium release and uptake by the intracellular system. We postulate that the probability of occurrence of an H.A. response (that is, the large- scale release of calcium by the ER) is a function of intracellular calcium concentration. A certain level of free cytosol calcium predisposes toward release of Ca²⁺ from the ER. (This is reminiscent of Ca release by Ca from muscle SR (Ford & Podolsky, 1970; Endo, Tanaka & Ogawa, 1970). Further, as noted, an increased threshold for eliciting the H.A. occurs following the occurrence of an H.A. (Nelson et al. 1972). These properties of the system, coupled with some finite time course of calcium removal after an H.A., would be expected to generate oscillatory behaviour given appropriate kinetic values for the different processes. A schematic depiction of the system we postulate is shown in Fig. 7. It is interesting that in skinned muscle fibres, oscillatory contractions have been obtained by appropriate combinations of caffeine (affecting Ca²⁺ uptake by the SR) and cytoplasmic calcium and EGTA concentrations (Endo et al. 1970). In this situation the SR uptake and release kinetics are clearly the basis for the oscillatory phenomena. The observations that at least under some conditions calcium ions in the bathing medium can produce release of calcium from the SR, lends plausibility to our postulate (above) that certain cytosol Ca²⁺ concentrations may lead to further release of Ca²⁺ from the ER.

We have emphasized possible analogies between muscle and mesenchymal cells with reference to an intracellular compartment which sequesters calcium ions and which can be caused to release this calcium by stimuli delivered to the surface membrane. While the evidence, we feel, justifies such an anology, it should also be pointed out that there are a number of uncertainties connected with the analogy. As discussed earlier, the nature of the ‘permissive’ calcium is not well defined. Even though there are no apparent barriers to diffusion between the L-cell surfaces and the culture fluid (Henkart & Nelson, 1979), it may be that local pools of calcium near or adsorbed to the inner or outer surface of the membrane exist and their movement through the membrane or from the inner membrane surface into the cytoplasm can release intracellular calcium. A situation of this sort may occur in muscle, since Barrett & Barrett (1978) have shown that sufficiently high concentrations of EGTA can block excitation contraction coupling. It is also possible that the relative dominance of calcium entering from outside the cell and calcium released from intracellular sites may participate to various extents in different cell types in producing hyperpolarizing responses.

Our treatment of oscillatory phenomena implies that a regenerative calcium- releasing-calcium sequence contributes to oscillatory behaviour. With calcium injections, however, we have found an essentially graded relationship between the amount of calcium injected and the degree of hyperpolarization produced, rather than a threshold type of relationship. Clearly further work is required to clarify these basic issues of the cellular mechanisms involved in the hyperpolarizing response.

In general terms, the electrophysiological evidence demonstrates a system which translates stimuli impinging on the surface membrane of mesenchymal cells into substantial changes in free cytoplasmic calcium concentrations. These changes could either be transient or, in the case of sustained oscillations, of longer duration. There are a number of possible functions that may, in turn, be controlled by this cytoplasmic calcium.

The importance of intracellular calcium for microtubule assembly (e.g. Schliwa, 1976) and contractile activity of actin and myosin (reviewed by Hitchcock, 1977) have been documented in a variety of cells. This suggests that control of cell shape and locomotory activity might be a function of an H.A.-generating mechanism. Consonant with this suggestion are the observations of Gallin & Gallin (1977), who have shown that in human macrophages, endotoxin-activated serum produced more pronounced H.A. responses than did control serum and that this property of the activated sera was due to chemotactically active components in the sera (Fig. 8). They conclude that ‘ion fluxes associated with membrane potential changes are early events in macrophage activation by chemotactic factors’. Thus the changes in calcium ion concentration which are manifested in the H.A. response may also play a critical role in the regulation of cell motility involved in chemotactic responses. Tsuchiya et al. (1978), in a preliminary communication, report that serum proteins can elicit H.A. responses and propose that the H.A. is a ‘receptor potential which is closely related to cell membrane movement’. In support of this they also note that cytochalasin B, which inhibited the cell motility reversibly, inhibited hyperpolarizations in their strain of L-cells.

Megakaryocytes, as precursors to blood platelets, have been proposed to exhibit certain functional parallels with muscle cells. Chemical and mechanical stimuli to platelets may release calcium within the platelet which in turn produces contraction of thrombosthenin and platelet aggregation (Miller, Sheridan & White, 1978). The H.A. responses in megakaryocytes may reflect such Ca²⁺ release within the megakaryocyte, but detailed study of this system has not been published.

In addition to the large increase in cytosol calcium concentration, a corresponding transient decrease in Ca²⁺ concentration within the ER would be expected if release of calcium from the ER is involved in the H.A. response. Since processing of some newly synthesized proteins occurs through the ER, it seems plausible that alteration in calcium concentrations, particularly during repetitive stimuli or prolonged oscillations in this structure, might have some regulatory effect on protein translocation into the ER and, hence, on synthesis and secretion.

Finally, we would suggest that the cell biological processes revealed in the H.A. response in mesenchymal cells may have some counterpart in neurones. Increased potassium permeability is involved in a variety of post-activation hyperpolarizations and oscillatory or pacemaker potentials. Ca²+ ingress across the surface membrane is undoubtedly involved in many of these processes; subsurface cisterns are prominent in many neurones, however, and the ER has been shown to sequester Ca²⁺ in squid upon (Henkart et al. 1978) and in synaptic terminals (Blaustein et al. 1978; McGraw, Somlyo & Blaustein, 1978). The possibility that this intracellular Ca²⁺ sequestering system is a dynamic one in neurones and is accessible to surface stimuli should be entertained. Intracellular events clearly can contribute dramatically to the regulation of surface membrane behaviour.