An overview of techniques for the measurement of calcium distribution, calcium fluxes, and cytosolic free calcium in mammalian cells.

An array of techniques can be used to study cell calcium metabolism that comprises several calcium compartments and many types of transport systems such as ion channels, ATP-dependent pumps, and antiporters. The measurement of total cell calcium brings little information of value since 60 to 80% of total cell calcium is actually bound to the extracellular glycocalyx. Cell fractionation and differential centrifugation have been used to study intracellular Ca2+ compartmentalization, but the methods suffer from the possibility of Ca2+ loss or redistribution among cell fractions. Steady-state kinetic analyses of 45Ca uptake or desaturation curves have been used to study the distribution of Ca2+ among various kinetic pools in living cells and their rate of Ca2+ exchange, but the analyses are constrained by many limitations. Nonsteady-state tracer studies can provide information about rapid changes in calcium influx or efflux in and out of the cell. Zero-time kinetics of 45Ca uptake can detect instantaneous changes in calcium influx, while 45Ca fractional efflux ratio, can detect rapid stimulations or inhibitions of calcium efflux out of cells. Permeabilized cells have been successfully used to gauge the relative role of intracellular organelles in controlling [Ca2+]i. The measurement of the cytosolic ionized calcium ([Ca2+]i) is undoubtedly the most important and, physiologically, the most relevant method available. The choice of the appropriate calcium indicator, fluorescent, bioluminescent, metallochromic, or Ca2(+)-sensitive microelectrodes depends on the cell type and the magnitude and time constant of the event under study. Each probe has specific assets and drawbacks. The study of plasma membrane vesicles derived from baso-lateral or apical plasmalemma can also bring important information on the (Ca2(+)-Mg2+) ATPase-dependent calcium pump and on the kinetics and stoichiometry of the Na(+)-Ca2+ antiporter. The best strategy to study cell calcium metabolism is to use several different methods that focus on a specific problem from widely different angles.


Introduction
Cell calcium metabolism comprises several intraand extracellular compartments where calcium exchanges occur through a complex array of transporters including three types of calcium channels, two types of (Ca2+-Mg2")ATPase-dependent calcium pumps, and at least one, or perhaps two types of antiporters (Na'-Ca2" and Ca2+-H' exchangers). One of the most important parameters of cell calcium metabolism is the cytosolic ionized calcium ([Ca2+]i).
At rest, the control of [Ca2+]i is effected through the interplay of the calcium ATPase pump and the Na+-Ca2+ antiporter at the plasmalemma; calcium influx into mitochondria, driven by the inner membrane potential that is established by the proton pump; calcium efflux from mitochondria through a Na'-Ca2' *Department of Physiology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261. exchanger and/or an unidentified mitochondrial efflux pathway; and finally, a calcium ATPase pump in the endoplasmic reticulum (Fig. 1). Whether calcium cycles across the endoplasmic reticulum (ER) membrane at rest is not known.
During excitation or after hormonal stimulation, [Ca2+]i is regulated by a completely new set of transporters. The following calcium channels of several types open in the plasmalemma: voltage-operated channels (VOC), receptors-operated channels (ROC), and intracellular-messengers-operated channels (IMOC) (Fig. 2). Several subtypes of voltage-operated channels have been described (type L, T, N). At the endoplasmic reticulum membrane, inositol triphosphate or other messengers triggers and efflux of calcium, presumably through an intracellular-messenger-operated calcium channel.
Finally, [Ca2+l is modulated or influenced by a wide variety of ions or metabolites and by their specific transporters (Fig. 3). The forward or reverse mode of Voltage-operated channels (VOC), receptor-operated channels (ROC), intracellular-messenger-operated channels (IMOC) and antiporters reported to respond to specific stimuli regulating cytosolic free calcium in an excited cell. the Na'-Ca2"antiporter is thermodynamically determined by the balance between the Na+ and Ca2' electrochemical potentials. Consequently, the membrane potential, the intracellular and the extracellular Na+ activity, the (Na-K) ATPase-dependent Na pump, the Na+-H+antiporter, intracellular and extracellular pH, the Na+-glucose and Na+-amino-acid symporters, and the ATP-dependent proton pumps may all indirectly influence or modulate [Ca2+]j. With such complexity, it is not surprising to find that no single method can ever offer a comprehensive or compelling description of the control, regulation, and modulation of cell calcium metabolism.

Total Cell Calcium
Total cell calcium has been measured by atomic absorption, flame photometry, or by fluorometric titration with calcein as indicator. calcium. When the extracellularly bound calcium is removed by typsin or by EGTA, the naked cell calcium averages 0.6 mmole/kg wet weight (range 0.4 to 0.9 mmole/kg wet weight).

Cell Fractionation
Cell fractionation and differential ultracentrifugation have often been used to gain further insight into calcium distribution among various cell compartments, in spite of the fact that the methods suffer from possibility of calcium loss or redistribution among the various cell fractions. To minimize such losses or redistribution, EGTA, ruthenium red, and various anesthetics are often added to the homogenizing medium (2). The distribution of calcium among compartment varies slightly from tissue to tissue, but the overall percentage shows that most of the cell calcium is sequestered in ER and in mitochondria ( Table 2). Since the cell dry weight is approximately 20% of the wet weight and the overall naked cell calcium is 0.6 mmole/kg cell wet weight, mitochondrial calcium is about 1.3 mmole/kg cell dry weight. In liver cells, the ER calcium is about half that of the mitochondria, when expressed on a cell dry weight or cell protein basis. However, since the ER protein is 10% and the mitochondria, 33% of the total cell protein (3), the ER calcium concentration is more than twice the mitochondrial calcium per organelle mass. This is in good agreement with recent results obtained with electron microprobes (4).

Kinetic Analyses
The kinetic analyses of 45Ca uptake or desaturation curves were developed to gain some insight in the cal- cium distribution among different cell compartments in a living cell and to estimate the rate of calcium exchange between compartments (5-7). It was also an attempt to peel off kinetically the large pool of calcium bound to the glycocalyx without using enzymes or ETGA and to obviate the calcium redistribution that may occur during cell fractionation and ultracentrification. The method has many limitations that are not always recognized. First, these studies must be performed at steady state. Consequently, short-term, transient, or fastchanging events cannot be studied. Second, the kinetic pools obtained do not necessarily represent welldefined anatomical compartments. They usually reflect a collection of subpools. Indeed, unless compartments all differ materially in their behavior, groups of similar compartments will tend to act together, and the kinetic of the central compartment will be indistinguishable from the kinetics of a system of relatively few compartments (8). Identification of the pools is therefore tentative and should be supported by other independent measurements. Third, the number of kinetic components cannot be chosen arbitrarily, but must be determined by a precise formula developed by Jacquez (9) (Fig. 4). Fourth, the curves must be analyzed by the weighted, nonlinear least square method (Fig. 5) because graphical analysis is too subjective to be acceptable. Fifth, the exponential constants and coefficients of the multiexponential equation or, if one prefers, the slopes and intercepts of the kinetic components do not represent rate constants or compartments as often assumed. That would imply an improbable, pure parallel system with intracellular organelles in direct contact with the extracellular medium. A symbolic representation of the system must be chosen (series, parallel or mixed) ( Fig. 6) based on a believable anatomical description of the model under study. Sixth, steady-state kinetic analysis gives us exchange rates between compartments where influx equals efflux (pij = Jij = Jji) so that the results obtained by 45Ca uptake must match those obtained by 45Ca desaturation. One cannot assume that uptake measures calcium influx and desaturation measures calcium efflux.
45Ca uptake experiments can measure the fast and medium kinetic components fairly accurately (Fig. 7), but they cannot detect the third slow component that is less than the usual error of measurement when the curves approach isotopic equilibrium. On the other Even a careful and formally correct analysis brings a limited amount of information that is not very specific. This information reveals changes in cell calcium distribution and the rate of calcium exchange between different compartments (Table 3), but it cannot detect the primary defect in cell calcium metabolism or the sequence of events leading to the new steady state.  bBorle and Clark (27). cStuder and Borle (28). dStuder and Borle (29). eStuder (30).

Nonsteady-State Tracer Studies
To study rapid changes in calcium fluxes in and out of cells, we have used nonsteady-state tracer studies. Calcium influx is measured by zero-time kinetics during the first 30 sec of 45Ca uptake, when there is no significant backflux of tracer from cell to medium (Fig. 8). This is done by a rapid separation of the cells from the medium by filtration over a millipore filter and a rapid wash of the extracellularly bound calcium with lanthanum, EGTA, or high calcium wash.
Rapid changes in calcium efflux are measured by the fractional efflux ratio method (10). Two groups of cells labeled with 45Ca are desaturated side by side. One group is stimulated and its fractional efflux is compared with that of the unstimulated control ( Fig. 9). The fractional efflux ratio gives us the time course of the stimulation and its magnitude; it does not provide the actual rate of calcium efflux.

Permeabilized Cells
To study calcium mobilization from intracellular organelles or the relative importance of the endoplasmic reticulum and the mitochondria in regulating cytosolic free calcium, saponin or digitonin are used to make the cells plasmalemma permeable to ions. The cells are then incubated in a medium assumed to replicate the cytosol composition, i.e., high K, low Ca2', and an ATP-regenerating system. This method has been used to study the effects of intracellular messengers such as inosotol triphosphate on calcium mobilization from intracellular organelles (Fig. 10). The rise in Ca2' in the medium bathing the permeabilized cells can be measured with arsenazo III, Quin-2, or Ca2+-sensitive microelectrodes. Many investigators use mitochondrial uncouplers, Ca2' ionophores, or drugs such as dantrolene or TMB-8 to induce or block calcium release from one or the other intracellular compartments. The validity of the assumptions regarding the specificity of these agents is not always clear or compelling.

Cytosolic Free Calcium
The most significant breakthrough in the study of cell calcium metabolism in the last 8 years is, undoubtedly, the development of methods to measure cytosolic free calcium [Ca21]i in small mammalian cells. The four classes of calcium indicators are shown in Table 4. Their characteristics, relative assets, and drawbacks have already been reviewed (11)(12)(13).
To this admittedly biased investigator, aequorin is the best calcium indicator; it has a high sensitivity, a high signal-to-noise ratio, a fast response time, and it has no calcium-buffering action or no known side effect on cell function. The past difficulties in incorporating aequorin in small cells have been overcome by the development of several methods applicable to practically all mammalian cells (14,15). On the negative side, the method is not an easy one. It is tedious and time consuming. Great care must be taken to avoid calcium contamination in every solution and all laboratory ware in contact with the photoprotein. The calcium-aequorin reaction is irreversible so that aequorin is constantly being consumed. At the physiological levels of free calcium in the cell, the half-time of aequorin is several weeks (11), but above 1i0 M Ca2', aequorin is consumed in minutes.
This characteristic is actually an asset of the method. First, all the aequorin left outside the cell is immediately destroyed by the high extracellular Ca2'. Second, in the unlikely event of an accumulation of aequorin in subcellular organelles, the photoprotein would be inactivated in a matter of minutes by the high free-calcium concentration presumed to exist in these organelles. Another drawback that is shared with the fluorescent probes is the nonlinearity of the relation between calcium and the light signal. Any inhomogeneity in calcium concentrations within the cells or inhomogeneity in the cell population will distort the average measurement of [Ca2+]j.
The fluorescent probes Quin-2, Fura-2, and Indo-1 are much easier to use and therefore much more popular (16,17). They are perhaps best suited for the study of [Ca2+]i as one element in an otherwise complex system of intracellular signaling. Their best assets are their ease of incorporation as acetoxymethyl esters, the commercial availability of good fluorometers with dual excitation or dual emission, and a fairly straightforward calibration. The main drawbacks are their relatively slow response time, calcium buffering action, side effects on various cell functions, leaking from some cells, and occasional incorporation into subcellular organelles. If one is interested in cellular calcium metabolism per se, and not in cytosolic free calcium as only one of the many steps of a metabolic pathway's sequence of events, it is worth investing the time and effort required by the aequorin methodology.
The metallochromic dyes are the least popular of the four classes of intracellular calcium probes (18). They must be introduced into cells by microinjection, they have a low sensitivity and a low signal-to-noise ratio and their calibration in absolute terms is not possible because of the uncertainties regarding the stoichiometry. However, metallochromic dyes do have important qualities: they have the fastest response time of all calcium indicators (less than 10 msec), and the signal is linearly related to the free-calcium activity. Also, they are best suited to detect relative changes at high levels of [Ca2+]i in very fast events.
The Ca2"-sensitive microelectrodes use the neutral ligand ETH 1001 developed by Simon and collaborators (19). It has the broadest range of detections To determine whether the rise in cytosolic calcium results from a mobilization of calcium from intracellular organelles or from the extracellular milieu, these experiments can be performed in Ca-free media containing EGTA. Figure 12 shows that the peak [Ca21]i evoked by epinephrine is not suppressed in 0 Cao2+, but the shoulder disappears until Ca2+ is reintroduced in the perfusate, indicating that the peak results from intracellular calcium release while the shoulder is caused by an increased calcium influx. And indeed, Studer has shown that calcium influx measured by zero-time kinetics is increased by epinephrine ( Fig. 13) (R. K. Studer, unpublished). It is useful to use several methods to focus on a specific problem from different experimental angles.
For instance, the rise in [Ca2+]i observed after anoxia (21) or after lowering the extracellular Na+ by substitution with TMA (22) (Fig. 14) does not take place in the absence of extracellular Ca2" eliminating the contribution of intracellular calcium sources (21,22). The rise in [Ca2+]i could be caused by a decreased calcium efflux on the Na'-Ca2+ antiporter operating either in the forward mode (Ca2`j vs. Na+0) or by an increased calcium influx on the antiporter operating in the reverse mode (Ca2"0 vs. Na+;). Figure 15 shows that Na+o substitution by TMA stimulates calcium efflux proving that the rise in [Ca21]i is not caused by a decreased Ca efflux. On the other hand, Na+ substitution with TMA stimulates Ca2`influx when measured by zero-time kinetics (Fig. 16). The sum of these three experimental approaches suggests that lowering extracellular Na+ stimulates the reverse mode of the Na'-Ca2+ antiporter by increasing the calcium influx in exchange for an increased Na+ efflux which we have also documented experimentally (22).   (Fig. 17). As a consequence, the rise in Ca2" efflux is also depressed by adrenalectomy (Fig. 18). However, when one compares the rise in Ca efflux for a given rise in [Ca2+]i (Fig. 19), it appears that for the same increase in [Ca2+]j, calcium efflux is significantly less in adrenalectomy. This suggests a deficiency in the active calcium transport system, presumably a depressed calcium pump activity.

Isolated Organelles and Vesicles
It is also possible to study the kinetic properties of specific organelles or specific calcium transport mechanisms such as (Ca2"-Mg2+) ATPase-dependent calcium transport, Na'-Ca2" antiporters, and Ca2" channels in broken cell systems. These transporters have been isolated within plasma membrane vesicles or incorporated into artificial membranes or liposomes. The calcium transport properties of isolated mitochondria or endoplasmic reticulum have been under study for many years. Recently, patch clamps and whole cell patches have been used to study single calcium channels and intracellular signaling systems. I shall only mention these techniques since I have no expertise in any of them. In these model systems, the advantage of controlling the experimental conditions are offset by the possibility that these are not physiological conditions. Thus, the information obtained from these isolated systems may not always reflect their behavior in an intact cell and their relative role and importance in the control of cell calcium metabolism.

Calciotropic Drugs
Finally, a few words of caution should be offered regarding the various drugs or chemicals used as Ca2" ionophores, Ca2" blockers, and as agonists or antagonists of calcium transport processes. Few if any of these agents are specific, but they are often thought to be so, and their expected effects are sometimes assumed to occur without independent verification. For instance, A23187 is not only a calcium ionophore but also a proton ionophore. It increases [Ca21] . but it does not necessarily increase net calcium influx into cells, since most cells lose 50 to 70% of their calcium after the administration of A23187 (24).
Verapamil is a calcium channel blocker, but it also inhibits Na'-Ca2' exchange, and it is a powerful aadrenergic blocker (25). Nifedipine and diltiazem are both calcium channel blockers, but both increase [Ca21]i in kidney cells and in hepatocytes at concentrations of 10-9 M and 10-3 M, respectively (Borle, unpub-lished). Amiloride, dichlorobenzamil, and dimethylbenzamil are used as blockers of the Na+-Ca2' and Na+-H+ antiporters, but all three increase [Ca2+]i in kidney and liver cells at concentrations as low or lower than their KI (Fig. 20). Dichlorobenzamil also dramatically increase calcium influx in kidney cells (Fig. 21), and because of the high cytosolic calcium, it enhances calcium efflux (Fig. 22). Amiloride, dichlorobenzamil, and dimethylbenzamil are also powerful a-adrenergic antagonists (Fig. 23). The effects of calciotropic drugs on several parameters of cell calcium metabolism should be carefully surveyed before their assumed specific mode of action can be accepted.

Conclusion
In conclusion, many methods are available to study cell calcium metabolism, [Ca21]i, calcium distribution, and calcium fluxes. Beyond recognizing the assets and drawbacks of each technique, the safest approach to study a particular problem involving cell calcium is to use several different methods, focusing on it from different angles.