The Properties of an Intercellular Matrix and Its Relation to Electrolytes

The possession of a relatively tough surface layer appears to be a characteristic of all animal cells including the Protozoa. If this layer be violently disturbed the whole cell rapidly becomes instable and may disintegrate; slight injuries can, however, be repaired by regeneration from the exposed cytoplasmic surface. In young cells this surface layer is probably to be regarded as an integral part of the living cell, but as the cell increases in age the more superficial parts of the surface layer become converted into inanimate secretions of various types. In the case of certain tissues it can be shown that the individual cells are bound together by a matrix which is derived from this clear surface layer of the young individual cells.

The existence and origin of the matrix is particularly clear in the case of the segmenting eggs of Echinus esculentus (Gray, 1924). About twenty minutes after fertilisation, the cortical region of the egg is differentiated from the interior as a distinct hyaline layer. When the egg divides, each blastomere is separated from its fellows by a layer of this hyaline substance derived from and continuous with the original surface. In the earlier cleavages the continuity between the surface layer and the intercellular matrix is quite obvious, but when the blastula stage is reached the cells are closely pressed against each other, so that it is difficult to detect any true intercellular matrix except at the distal surfaces of the cell where a clear hyaline layer can be identified. In the case of the ciliated epithelium on the gills of Mytilus edulis the intercellular matrix has a similar distribution (fig. 1), except that in the case of the lateral epithelium the continuity between the surface layer and the intercellular laminæ can be shown by appropriate staining with silver nitrate (fig. 2). The stability of the epithelium depends on the presence of this intercellular complex, since if it be removed adjacent cells separate easily from their neighbours. Similar membranes have been described in the case of other tissues, and quite recently Galtsoff (1925) has shown that, when dissociated cells of Microciona join together to form restitution bodies, the cells fuse by their superficial “hyaloplasm” and that the “granuloplasm “of the cells remain as separate entities. In all these cases the volume of the surface layer or intercellular substance is comparatively small in comparison to that of the cell itself, but in the case of connective tissue and of cartilage this is not the case. In the connective tissue of young vertebrates the intercellular substance appears to consist of mucoid (with fibrils of collagen and elastin), and the cells themselves form only a small fraction of the tissue.

The stability of all these types of intercellular substance appears to be associated with the presence of divalent cations. Herbst (1900) showed that in calcium-free sea-water the intercellular matrix of echinoderm larvæ is disorganised and the cells separate from each other. In the case of Mytilus tissue the matrix is disorganised in the absence of magnesium. Galtsoff (1925) found that calcium and magnesium are both necessary for the complete reunion of dissociated sponge cells. The intercellular matrix of connective tissue also contains considerable quantities of calcium and magnesium. It is significant to note that these divalent ions are also closely associated with the maintenance of the osmotic properties of the living cells in both plants and animals.

The generally accepted explanation of these facts is that both the intercellular matrix and the physiological cell surface consist of a calcium compound which, in the absence of calcium, breaks down to give a “soluble” substance which is dispersed in the medium. Thus, in the presence of pure sodium chloride, the calcium compound reacts to form a soluble sodium salt which is soluble. No direct evidence is available as to the organic constitution of the physiological cell-surface, although the higher fatty acids are at times suggested. In the case of the intercellular matrix the evidence is slight, except in the case of vertebrate connective tissue where mucoid is the chief constituent.

In the present paper the relationship between the intercellular matrix of Mytilus tissue to electrolytes will be discussed in some detail, and it will be shown that the presence of divalent ions is only essential for stability under certain specified conditions. Further, it is important to know in the case of a contractile tissue (where a wave of activity can pass across the intercellular junctions), whether an ionic change in the external environment operates on the intercellular matrix or on the individual cells.

A study of the intercellular matrix of echinoderm larvæ (Gray, 1924) shows that it consists of a plastic substance whose superficial layer is not elastic. In the presence of hydrogen ions it readily loses water, and in the presence of hypertonic sea-water it absorbs water. Chambers has shown that when the blastula stage is reached, the surface membrane (derived from the matrix) is tough and resistant. In the case of Amœba verucosa the ectoplasm is coherent and to some extent elastic (Howland, 1923).

The only compounds which appear to have these properties are the proteins and possibly some of the polysaccharides. The calcium salts of most fatty acids do not as a rule form diphasic gels, but when precipitated in the form of a membrane they are inelastic and brittle; they are not plastic, nor are they transparent. They do not absorb water by imbibition. The only exception to this is the type of gel described by Schryver (1915) in the case of calcium cholate, where the conditions of stability are much nearer those to be described in this paper than those of most fatty acids. The substances whose physical properties most nearly approximate to those of the intercellular matrix are such proteins as mucin or gluten. If these substances in the powdered form are rubbed with a little water they readily yield a tough coherent mass which is practically insoluble in water, and which can be pressed out into a more or less transparent membrane. The factors which influence the stability of the intercellular matrix of Mytilus tissue are strikingly similar to those which affect the stability of these protein gels when on the alkaline side of their isoelectric point.

Unfortunately such systems do not appear to have been extensively studied, so that it is necessary to discuss them in some detail before describing their biological application.

1. Membranes of gluten or mucin when exposed to low concentrations of an alkali such as sodium hydroxide gradually lose their coherence and pass into the dispersed condition. The more highly ionised is the membrane the more readily is it dispersed.

The dispersion by means of alkali is, however, markedly reduced by the presence of neutral salts (Hardy and Wood, 1908). It is primarily the cations of the salts which affect the system, and the efficiency of any cation largely depends on its valency, the divalent ions being much more effective than the monovalent. This stabilising action of cations against the dispersive power of alkali can properly be regarded as an electrostatic effect. It is important to note that the anions present in the system play little or no part in this respect, but that all cations on the other hand antagonise the action of hydroxyl ions. Robertson (1916) showed that salts affect the dispersal of casein by alkali in essentially the same way, and figs. 3 and 4 show the inhibiting effect of cations on the solubility of sodium mucinate.

2. Although the dispersal of a membrane most readily takes place when it is ionised, it usually only does so if the constituent particles have an affinity for water. The attraction between the water and the particles, in conjunction with the electrostatic repulsion of the latter are together sufficient to overcome the cohesion between adjacent particles. Thus the sodium salts of the proteins are readily dispersed because their molecules have a high affinity for water—they are in fact soluble. Under certain conditions, however, the divalent metals form compounds which have little or no affinity for water, so that membranes of such substances do not disperse unless very strong forces are used, and particles of such compounds can only remain dispersed if electrically charged.

The formation of these basic salts does not occur unless protein ions are present If the ionisation of a protein system is reduced by the presence of an excess of monovalent cations the formation of basic insoluble salts is inhibited (see p. 182). The capacity of such metals as calcium to form insoluble compounds will be referred to as the chemical effect of ions.

A lyophilic effect can, however, also be exerted by undissociated molecules or by non-electrolytes, e.g. the “salting out” of proteins and the inhibition of glycerine on the dispersal of casein by hydroxyl ions (Robertson). Whereas, therefore, the electrostatic and chemical action of salts is restricted to the cations and only operates on ionised particles, lyophilic effects on the other hand operate on unionised particles and can be exerted by molecules or by ions of either sign. Since lyophilic effects are only operative if the colloid is other-wise free to disperse, it is not surprising to find that these effects are all but completely masked when the powers of dispersion are strongly reduced by electrostatic action. Thus in physiologically balanced solutions containing calcium and magnesium the lyophilic effect of anions almost entirely disappears (Gray, 1922).

Although the factors controlling the stability of plastic colloidal membranes are still far from clear, it is certain that unless such metals as calcium have entered into intimate chemical union with the colloid, the stability of the latter will be determined not by the presence or absence of any particular ion but by an equilibrium in which many factors may be involved. The two main factors are the degree of ionisation of the colloid, and its ability to combine with water to form a soluble electroneutral particle.

Under normal conditions the ciliated cells on the gill are in contact with sea-water on their outer surface and indirectly with the blood on their inner surface. When a portion of the tissue is excised and placed in sea-water, both sides of the tissue are exposed to the same medium, and the conditions are therefore abnormal. As far as the stability of the inter-cellular matrix is concerned, however, this fact can be ignored since excised tissue in contact with sea-water on both the inner and outer surfaces is stable for many days. The material used in the experiments was thoroughly washed in sea-water before use. Owing to the small amounts involved the quantity of matrix dispersed cannot be estimated directly, and consequently the efficiency of a solution to maintain the stability of the matrix as a continuous phase was either judged by the time which elapsed before the cells at the apical ends of the gill filaments begin to float freely in the solution, or by an approximate estimate of the percentage of cells separated from the whole epithelium.

Those conditions which allow the matrix to disperse also induce an absorption of water by the cytoplasm of the cells, and the final separation of the cells may be hastened by this process. Dispersion of the matrix and swelling of the cells are two correlated phenomena.

If portions of a gill are placed in pure water the cells rapidly swell (fig. 5b) and may even burst. This type of swelling must be regarded as an osmotic effect and brought about by all osmotically active substances inside the cell. It is entirely different to the imbibitional swelling which is only concerned with electrolytes. The osmotic swelling can be compensated by the presence of non-electrolytes in the external medium.

In all isotonic solutions of non-electrolytes dispersion of the matrix and swelling of the cells occurs if the solutions are more alkaline than pH 4. The rate at which the matrix is dispersed depends on the nature of the non-electrolyte and on the concentration of hydroxyl ions present.

Different non-electrolytes obviously have very different powers of bringing about dispersion. Further, whereas urea is a much more powerful dispersion medium than the electrolyte NaCl, yet glycerine and dextrose are much less powerful. This fact alone makes it unlikely that the matrix is simply composed of some insoluble calcium salt. The effect of hydroxyl ions still further indicates the inadequacy of such a view since nearly all calcium salts are more soluble in lower hydroxyl ion concentrations than in higher. It will be noted that the effect of the different non-electrolytes is parallel to their effect on the stability of gelatin gels. The addition of magnesium or calcium greatly retards the dispersion of the matrix in all non-electrolytes (see p. 181).

The dispersion of the matrix is always accompanied by an uptake of water by the cells, and the two processes go on together. Both processes are much more rapid in urea than in glycerine. Observations with a large number of different solutions both of electrolytes and non-electrolytes leave no that the rate of water absorption by the cell is proportional to the rate at which the matrix goes into the dispersed state. The first effect of such an imbibition of water is to render the cell spherical (fig.5c), and a small round vacuole soon becomes visible in the cytoplasm. As imbibition proceeds the vacuole increases in size at the expense of the cytoplasm. In this condition small granules are often visible embedded in the cyto-plasm, and in this position show no visible Brownian movement, but as soon as the granules are included in the vacuole they show active movement. If the cell is stained with neutral red, the granules of the dye are red in colour as long as they are motionless in the cytoplasm, but become orange on liberation into the vacuole.^* It is difficult to avoid the conclusion that the factors involved in the dispersion of the intercellular matrix involve also a corresponding change inside the cell. Swelling and dispersion of the cytoplasm into a vacuole are correlated just as are the two phenomena in such colloidal systems as have been discussed.

A difference in the physiological properties of different non-electrolytes is of some importance. It was observed by Carlson (1906) in their action on the heart, and Denis (1906) concluded that it is due to the effect of their varying viscosity on the rate of diffusion of calcium and magnesium ions from the tissue. Attempts to obtain direct proof of this hypothesis have failed in the case of Mytilus tissue. The loss of electrolytes from the tissue was followed by measuring the electrical resistance of the surrounding non-electrolyte. As far as possible equal quantities of tissue, and equal volumes of non-electrolytes were used. It was, of course, impossible to ensure that the quantities of tissues were exactly equal and had an equal quantity of sea-water in the interstices of the filaments.

In both solutions exosmosis of electrolytes appeared to be complete in twenty minutes; the matrix, however, was dissolved by this time in urea, and not for nearly three hours in glycerine. It seems likely that the differential effect of urea and glycerine is largely due to their differential effect on the actual rate of disintegration of the matrix after the electrostatic change due to the removal of magnesium and calcium is complete. This view is strengthened by two experiments. First, if the tissue is washed in pure for half an hour some at least of the divalent ions can be shown to have left the tissue, if it be now transferred to the time of dispersal of the matrix is practically the same as that of tissue transferred direct from sea-water to the alkaline NaCl. The second experiment refers to sodium caseinate; the following figures show the differential effect of urea and glycerine with and without the presence of magnesium.

Although the difference in the physiological action of different non-electrolytes such as urea and glycerine may be associated with the effect of their viscosity on the rate of exosmosis of electrolytes, in this particular case it is, to a much larger extent, due to their direct effect on the stability of the colloid itself^*

Within a well-marked range of concentrations ethyl alcohol inhibits the rate of dispersal of the matrix. Thus in ⁠.

The precipitating action of alcohol on hydrated particles is well known, and Robertson has shown that in not too high concentrations it markedly reduces the rate of solution of sodium caseinate. The effect of higher concentrations is probably due to dispersion in a denaturated state. It also occurs in casein systems. In a subsequent paragraph it will be shown that the dispersive power of all non-electrolytes is very greatly reduced by the presence of divalent cations, and it is interesting to note that R. S. Lillie showed that the absence of such ions can be partially compensated by alcohol in the case of sea-urchin eggs.

The temperature coefficient of the dispersal of the matrix in urea solution is low. At pH 7.0 it is about 1.5 for the range 4° to 14°. Similarly low coefficients appear to be characteristic of the dispersal of proteins.

Isotonic solutions of such salts as lithium, sodium, ammonium, and potassium chlorides bring about dispersion of the matrix at about the same rate. At pH 8.0 and at 15 °C. complete dispersion occurs after about one hour, although material from different animals differs some-what.^* In comparison to urea (the only non-electrolyte whose viscosity is approximately the same as that of the electrolyte) the process of dispersion is markedly slower in all the alkaline chlorides. As before, the hydroxyl ion is a very important factor.

Fig. 7 shows that from pH 5.0 to pH 10.3 the rate of dispersion is proportional to the concentration of hydroxyl ions. The same thing is true for the dispersion of casein by sodium hydroxide (Robertson).

So far, the results are in harmony with the view that dispersion is due to hydroxyl ions and that the sodium or other ions inhibit this electrostatically, but in $M2$ concentration different sodium salts also exert a well-marked lyophilic action.

The anions fall into well-marked lyophil series as was first pointed out by R. S. Lillie (1906). The series is practically the same as that found for protein dispersals. The fact that the lyophil series is all but entirely unobservable in the presence of divalent cations (see Gray, 1922) is entirely in keeping with the hypothesis that the lyophil action can only apply to those parts of the matrix which are electrostatically free to disperse. By increasing the concentration of NaCl to twice the isotonic value, the inhibition of dispersion is greatly increased, and approximately equal to that of ⁠.

The dominating effect of the monovalent hydrogen ion has already been illustrated. It is by far the most powerful inhibitor of dispersion which exists, as is of course the case in protein systems.

— It is the distinctive property of magnesium that it is the only metal which at the normal pH of sea-water will prevent the dispersal of the matrix, and at the same time maintain the translucency and healthiness of both the matrix and the cilia. The effect of magnesium is seen in fairly low dilutions.

The stabilising effect of magnesium in the presence of sodium chloride has been described in a previous paper (Gray, 1922). It should be noted, however, that although magnesium alone, or better still, a mixture of sodium and magnesium prevents the dispersion of the matrix (and at the same time the uptake of water by the cells), yet such solutions do exert a definite dispersive action after about forty-eight hours. Complete stability is only assured in the additional presence of calcium, and even potassium appears to have some slight action.

Again, whereas the addition of magnesium to a solution of sodium chloride very greatly assists in the stabilisation of the intercellular matrix, the addition of magnesium to potassium chloride solutions has little or no effect. Calcium, on the other hand, does effect stabilisation in the presence of excess of potassium.

All other divalent cations resemble magnesium in exerting a marked stabilising action on the matrix, but they all involve a change in the matrix which leads to its precipitation in a granular form. This invariably occurs also with trivalent cations. This effect is seen very clearly in the case of calcium and it only takes place in the absence of magnesium.

In the absence of magnesium from sea-water, the higher the concentration of calcium and the higher the hydroxyl ion concentration, the more intense is the coagulation of the matrix. In the presence of 0.08 mol. calcium chloride at pH 9.0, both the cilia and the matrix rapidly become opaque and break up into minute granules. This phenomenon has already been described elsewhere (Gray, 1922).

In order that calcium ions should react with a protein (on the alkaline side of its isoelectric point) to form an insoluble basic compound, the protein must be ionised. Thus, if calcium chloride is added to a solution of sodium caseinate at pH 7.3 a marked opacity readily appears and eventually a definite precipitate is formed. If, however, the ionisation of the sodium caseinate is reduced either by the addition of acid, or by the presence of an excess of metallic cations, the opacity caused by the addition of calcium ions is partially or altogether inhibited. The action of sodium chloride is quite well marked, and magnesium chloride may inhibit the formation of a precipitate altogether. It is therefore reasonable to conclude that in the presence of magnesium the intercellular matrix of Mytilus cells is very slightly, if at all, ionised, and that it can only react with calcium to form an insoluble compound if the magnesium be removed and the solution be sufficiently alkaline to allow of ionisation of the membrane surface.

The distinctive effect of magnesium on both protein and living systems appears to be its capacity of exerting an electro-static effect on negatively charged surfaces without seriously affecting the capacity of the particles to exist in a hydrated condition. Very high concentrations of magnesium are usually required for the precipitation of dispersed hydrated proteins, whereas an electrostatic effect on charged particles is readily effected by low concentrations. All other divalent ions are liable to react with the protein or intercellular matrix to give a product which has no affinity for water and has therefore little or no coherence.

The whole of the experimental evidence shows that the intercellular matrix of the cells on the gills of Mytilus consists of a colloidal substance whose stability as a continuous phase is influenced by the same factors which influence the stability of certain proteins. By far the most powerful dispersive agent is the hydroxyl ion, and all salts and some non-electrolytes inhibit its action. Under normal conditions in sea-water the dispersive action of the hydroxyl ions is largely but not entirely inhibited by the magnesium ions, and these ions also prevent the reaction which would otherwise occur between the matrix and the calcium in the water.

The matrix is only completely stable when sodium, potassium, magnesium, and calcium are present, and it is therefore necessary to conclude that each of these ions enter into the final equilibrium, and that one ion cannot completely compensate for the absence of another. It follows that the analogy to inanimate systems such as the proteins must not be pushed too far. The living system probably involves some more delicate factor than the simple maintenance of an “insoluble” matrix. It may be. that the matrix has some essential texture which is essential for complete stability as is the case in certain soap systems. Further, there is as yet no simple explanation of the fact that although magnesium ions in the presence of sodium chloride exert a well-marked stabilising action, yet in the presence of potassium chloride they are distinctly inefficient.

It is of interest to note also that although the adhesive properties of the isolated cells of Microciona largely depend on the presence of either Ca or Mg, yet for complete adhesion, sodium, magnesium, and calcium are all necessary (Galtsoff, 1925). The delicate adjustment between cations required for the complete stability of the intercellular matrix of Mytilus, Echinus (see Gray, 1924), and Microciona indicates possibly that that region of the matrix nearest to the cell is an integral part of the latter. It also illustrates the fact that a complete physico-chemical analysis of the problem is not yet available.

In the particular case of the tissue here investigated all the evidence points to the conclusion that in an environment of sea-water the surface of the tissue is very slightly, if at all, ionised, and for this reason it does not react with the calcium in the external medium. It is not unreasonable to suppose, however, that in other cases the surface is of such a nature that it is capable of bearing a definite electro-negative charge even in the presence of the magnesium in sea-water; in this case calcium ions in the sea-water will react with the matrix to form a stable compound. In other words, a matrix of a calcium compound would form of the nature of those which apparently exist in echinoderms and in Fundulus. Further, this type of membrane might well occur in the case of tissues normally in equilibrium with waters containing little or no magnesium. The calcium type of matrix can thus be very simply derived from the more generalised colloided type such as exists in Mytilus. It might be suspected that the calcium type of matrix would occur in the case of freshwater organisms and this is supported by Benecke’s (1898) experiments with Spirogyra. Such membranes in the case of marine animals must, however, resemble gluten in that when in combination with calcium they still retain a certain degree of coherence. In the case of animals living in more acid environments calcium would not form basic compounds so readily, but would tend to do so if the waters became more alkaline.

Considerable stress has been laid upon the similarity between the colloidal properties of the intercellular matrix of Mytilus with those of certain proteins, and it is therefore of interest to consider any evidence which points to the actual presence of a protein in the matrix. When a considerable amount of matrix is dispersed in a suitable medium and the solution evaporated to dryness, the residue gives a faint Millon’s reaction indicating that a protein is present. Since, however, there is always a certain amount of extraneous mucilage on the surface of the gill it is impossible to be certain that the colour reaction is really due to the dispersed matrix. The most characteristic reaction of the matrix is its capacity to reduce silver salts. After impregnation with silver nitrate the outlines of each cell are clearly marked out, and the intercellular matrix is seen in some cases to be in communication with the surface membrane of the tissue. In this respect, the matrix of all cells including that of vertebrate connective tissue seems to be the same. The bases of the ciliated cells of Mytilus rest on the skeletal tubes of the filament. These tubes are sometimes described as consisting of chitin, but as they give very clear biuret and Millon’s reactions they must be protein in nature (? conchiolin). They differ from the intercellular matrix in being practically insoluble in cold dilute alkalies, although they stain in silver preparations. All the available facts receive a reasonable explanation on the assumption that the undifferentiated cell matrix is a protein of the mucoid type, and the evolution of intercellular secretions may be depicted as follows:—

It is doubtful to what extent these results throw any light on the constitution of the system responsible for maintaining the osmotic properties of the whole cell. Although Osterhout (1906) has shown that in certain cases calcium is probably an essential constituent of the physiological structure which produces the high electrical resistance of normal cells, yet the absence of calcium can partially at least be compensated by the presence of hydrogen ions, and one of the most powerful destructive agents is the hydroxyl ion. This would appear to show that the maintenance of the normal permeability of the cell depends on the stability of some organic substance, which although normally maintained by calcium, can by altering the concentration of hydrogen ions be maintained by other means. The present series of experiments indicates that in the absence of divalent cations the water content of the cells ceases to be controlled by osmosis but is controlled by inhibition.