Evidence That Calcium-Dependent Cellular Processes are Involved in the Stiffening Response of Holothurian Dermis and That Dermal Cells Contain an Organic Stiffening Factor

Echinoderms have neurally regulated collagenous tissues (Takahashi, 1967; Maeda, 1978; Wilkie, 1978a,b, 1984, 1988; Motokawa, 1984, 1988; Morales et al. 1989). Resident neurosecretory cells are thought to control the stiffness and strength of their connective tissues through secretions that alter the stress-transfer properties of the matrix between the spindle-shaped collagen fibrils (Matsumura, 1973, 1974; Smith et al. 1981; Hidaka and Takahashi, 1983; Motokawa, 1984, 1988; Wilkie, 1984, 1987, 1988, 1992; Trotter and Koob, 1989; Trotter et al. 1994). The identities of the secreted molecules and the mechanisms by which they affect the interfibrillar matrix are unknown.

A number of experiments have demonstrated that fresh echinoderm tissues containing viable cells are plasticized by lowering the concentration of calcium ions in the bathing medium (Hidaka, 1983; Diab and Gilly, 1984; Motokawa, 1984, 1987, 1988, 1994; Byrne, 1985; Wilkie, 1984, 1988, 1992; Hayashi and Motokawa, 1986; Shadwick and Pollock, 1988). They are stiffened by subsequent restoration of normal calcium concentrations. These experimental results are consistent with two distinct hypotheses. In the ‘extracellular calcium regulation’ hypothesis, the neurosecretory cells control the mechanical properties of the tissue by regulating the free Ca²⁺ concentration in the extracellular matrix; the strengths of the binding interactions of some macromolecular constituents of the interfibrillar matrix are postulated to be Ca²⁺-dependent. In the contrasting ‘cellular calcium regulation’ hypothesis, the role of Ca²⁺ is postulated to be in the regulation of one or more Ca²⁺-dependent cellular processes, such as secretion (Motokawa, 1988). The two hypotheses are not mutually exclusive, since the secretion of Ca²⁺ could be Ca²⁺-dependent.

Recent experimental results have been interpreted to provide support for both hypotheses. Szulgit and Shadwick (1994) found that the spine ligament (‘catch apparatus’) of the sea urchin Eucidaris tribuloides was plasticized by the chelation of extracellular Ca²⁺ but was stiffened by the addition of the non-ionic detergent Triton X-100. Their results were offered in support of the idea that Ca²⁺ chelation had changed one or more cellular activities, which secondarily resulted in a decrease in tissue stiffness. The addition of the non-ionic detergent was thought to have increased tissue stiffness by releasing one or more stiffening factors from lysed cells. The continual presence of a calcium chelator was thought to have eliminated the possibility that the stiffening factor could be Ca²⁺. Motokawa (1994) treated specimens of the dermis of a sea cucumber, Holothuria leucospilata, with Triton X-100 in combination with freezing and thawing, and subsequently estimated their viscosities from the results of creep (tensile) tests in the presence of widely varying cation concentrations.

He also concluded that Ca²⁺ acted at the cellular level, but argued in addition that Ca²⁺ acts directly as a stiffener of the interfibrillar matrix. His results were thus taken to support the hypothesis the dermis stiffening is caused by extracellular Ca²⁺-dependent Ca²⁺-secretion.

The experiments reported here were undertaken to discriminate between these hypotheses by testing the following specific predictions of the ‘extracellular calcium regulation’ hypothesis. (1) The stiffness and strength of the extracellular matrix should be Ca²⁺-dependent even in the absence of viable cells. (2) The stiffness and strength of the extracellular matrix should remain high in the presence of drugs that interfere with calcium-dependent cellular events when the extracellular concentration of Ca²⁺ is maintained at a normal level.

The results contradict both predictions and therefore fail to support the ‘extracellular calcium regulation’ hypothesis. However, the results are consistent with the ‘cellular calcium regulation’ hypothesis. In addition, evidence has been obtained for the existence of a cell-sequestered organic ‘stiffening factor’.

Portions of the data included in this report have been published previously in abstract form (Trotter and Koob, 1992, 1994, 1995).

Adult sea cucumbers, Cucumaria frondosa Hyman, were obtained by dredging Frenchman Bay, an inlet of the Gulf of Maine, and were maintained in mesh cages submerged beneath the floating dock of Mount Desert Island Biological Laboratory, Salsbury Cove, ME, USA. All experiments on fresh specimens were carried out at Mount Desert Island Biological Laboratory.

Test specimens were prepared from the two ventral interambulacra, which lack tube feet. The body wall musculature was stripped away from the dermis. To prepare uniform specimens, the safety shields of single-edged razor blades were glued together, either directly or with an intervening metal shim, using commercially available cyanoacrylate adhesive. This cutting apparatus was used to produce equivalent specimens from the white inner dermis that were approximately 3 cm long, 0.9 mm thick and either 1.7 or 1.8 mm wide. The long axis of each specimen was parallel to the longitudinal axis of the animal. The 0.9 mm thick side was in its radial dimension, and the wider side was in its circumferential dimension.

Creep tests were carried out in a chamber constructed from acrylic containing an epoxy-coated copper tube (Fig. 1). Natural sea water from the pumped seawater system of Mount Desert Island Biological Laboratory was passed through the tube in order to maintain the temperature of the test solutions at that of the water in which the animals lived, approximately 12–15 °C. The epoxy coating prevented the test solutions from being contaminated by copper salts. The ends of the test specimen were secured in acrylic clamps by the use of stainless-steel screws and cyanoacrylate adhesive. A 3 mm wide by 2 mm deep groove was milled in one of the two clamp members, and a complementary ridge was milled into the other. The specimen was retained in the groove by the glue and the pressure created by tightening the screws. Both the glue and the screws were found to be necessary to hold specimens firmly enough to be tested without either slipping or breaking at the clamps. After a specimen had been secured in both clamps, a digital caliper was used to determine the distance between the two clamps at which the specimen began to resist further extension. This was taken as the undeformed length (L₀). The bottom clamp was retained by shelves in the test chamber. A threaded steel rod extended from the top clamp. This rod was gripped by a collet clamp attached to a braided stainless-steel wire, 0.6 mm in diameter. The wire passed over two stainless-steel bearing-mounted nylon pulleys at either end of the horizontal member of a T-shaped aluminum stand. The other end of the wire was attached to a metal container which was weighted with lead shot. The container rested on a small platform. The specimen chamber was bolted to the top of a small platform jack. Between the two pulleys, the stainless-steel wire was glued to the core of a linear variable differential transformer (LVDT: Schaevitz model 100MHR, Lucas Schaevitz, Pennsauken, NJ, USA). The hollow cylinder of the LVDT was held in place by a V-shaped acrylic piece that could be positioned anywhere between the two pulleys. The holder and LVDT were held in place by tape. The output of the LVDT was passed from a digital transducer readout (Lucas-Schaevitz model DTR451) to a chart recorder. The system was calibrated by recording the deflection produced on the pen of the chart recorder by specific length changes in a micrometer. Once calibrated, all the settings of the signal conditioner and the chart recorder were left unchanged. Prior to each test, the weight was lowered until the bottom clamp made contact with the retaining shelf, and the specimen became straight. The LVDT was then positioned and taped. The creep test was begun by lowering the platform jack, leaving the weight hanging free, suspended only by the specimen. The lengthening of the specimen was recorded for a period of 30 min, or until the specimen broke or the limit of the LVDT was reached. When the test solutions contained only inorganic salts, buffer and chelating agent, the specimens were fully submerged during the tests. When drugs were included in the test solutions, they were dripped onto the specimens during the tests. The weight used in these tests was 200 g (producing a force of 1.96 N) above the weight required to balance that of the specimen clamps and specimen. The specimens were 0.9 mm×1.8 mm in cross section, and the initial load was therefore 1.21 MPa. The extension values for each time point were divided by the L₀ value for the specimen to obtain creep plots, in which strain (ε=Δ L/L) is expressed as a function of time.

Bending tests were carried out on 0.9 mm×1.7 mm×30 mm specimens. The specimens were glued at their midpoints to 0.5 mm tungsten wire which had been bent into a Z shape with two 90 ° angles. The wire containing the specimen was mounted in a device consisting of a vertical plate of acrylic mounted in an aluminum stand (Fig. 2). A groove had been cut from the top of the plate. The top of the groove was closed by a slip of stiff foam in which a slit was cut to accommodate the wire and to hold it firmly in place. The bottom of the groove had a shallow slit for the wire containing a small amount of Plasticine. A piece of millimeter-interval graph paper was glued to the front of the acrylic plate. A small ledge of acrylic was glued to the front of the plate, to one side of the groove, so that its upper surface would be at the same vertical position as the bottom of the specimen. When the wire containing the specimen was in position in the device, 15 mm of specimen extended beyond the wire as a cantilever beam, while the other half of the specimen was supported by the acrylic shelf. The testing device was mounted in a flowing seawater tray, with the specimen positioned approximately 4 cm above the surface of the water. This was to maintain the temperature and humidity at relatively constant values throughout all of the tests. During the mounting process, the specimen was held in a slightly upwardly bent position using the handle of a pair of tweezers to support it gently. To begin a test, the free end of the specimen was released, and a stopwatch was used to measure the time required for the specimen end to move a vertical distance of 4 mm, beginning at a position no lower than 1 mm above the horizontal position of the specimen, and no higher than 3 mm above the horizontal. The time was rounded up to the nearest second. Very plastic specimens therefore had a uniform deflection time of 1 s. Because most tests were completed in less than 2 min, and no tests took longer than 5 min, the potential problems arising from specimen desiccation were not thought to be serious. Specimens from the dermis of C. frondosa were suitable for bending tests because they retained their shape even after lengthy incubation in test solutions.

Initial experiments were carried out using creep tests. Some difficulties inherent in this test modality, however, led to the introduction of bending tests, which have several distinct advantages. One major problem with the creep test was that it was difficult to find clamping conditions that prevented specimens from either slipping or breaking at a clamp. This was especially true with plasticized specimens, which yielded readily as the clamps were being tightened. A second problem was that the time required for each creep test made it difficult, with a single test apparatus, to conduct many tests within the same time frame. This was a significant limitation, because the animal-to-animal variability in viscosity measured on specimens tested in artificial sea water (ASW) made it necessary to conduct each set of experiments on specimens from a single animal. For example, it can be seen in Fig. 4 that the bending times of specimens from animals 7 and 9 averaged more than 70 s, whereas those of specimens from animal 3 averaged less than 30 s. These difficulties partially offset the inherently greater information content of creep tests when compared with bending tests. One advantage of the bending test was that the tissue was not distorted or damaged by being clamped. A second advantage was that the rapidity of the tests allowed many tests to be conducted within a short time, and hence the number of repeated tests of each experimental variation was sufficient to obtain statistical information. These advantages partially compensated for the limitation that each test produced only a single time point and thus lacked the more complex information content of a creep test. Both test modalities were applied to all of the experiments described in the present paper, with the exception that the evaluations of the stiffening effects of tissue extracts were made exclusively with bending tests. In all cases where both test modalities were used, comparable results were obtained.

The principal test solutions used in these studies were as follows: (1) Mops-buffered artificial sea water (ASW), which consisted of 0.5 mol l⁻¹ NaCl, 0.05 mol l⁻¹ MgCl₂, 0.01 mol l⁻¹ CaCl₂, 0.01 mol l⁻¹ KCl and 0.01 mol l⁻¹ 3-(N-morpholino)propane sulfonic acid (Mops), pH 7.8–8.0; and (2) EGTA–ASW in which the CaCl₂ was replaced by 0.0072 mol l⁻¹ EGTA. Standard specimens were cut from the dermis and incubated in the test solutions at ambient seawater temperature, with intermittent gentle agitation. Each experiment comparing the effects of different treatments used specimens from a single animal, and each experiment was repeated using specimens from at least three different animals. Approximately 30 specimens could be obtained from a single animal. Each specimen was tested only once. Unless otherwise noted, all incubations took place at seawater temperature (12–15 °C), and the pH was maintained at 7.8–8.0.

Free Ca²⁺ concentrations were estimated using the B&S stability constants (Brooks, 1992) in the computer program Maxchelator v6.5, obtained from Dr Chris Patton, Stanford University, USA. The total calcium contents of specimens were determined using atomic absorption spectroscopy (model 460 AA spectrophotometer, Perkin-Elmer Corp., Norwalk, CT, USA) on tissue extracts made in 1 mol l⁻¹ HCl. Preliminary experiments showed that 1 mol l⁻¹ HCl extracted 100 % of the calcium, as judged by the inability of sequential treatment with 5.25 % sodium hypochlorite (household bleach) and 70 % HNO₃ to extract more. Specimens from the dermis of C. frondosa were suitable for direct measurements of the calcium associated with soft tissue elements because the adults of this species lack calcareous ossicles (Hyman, 1955). This fact was verified by microscopic examination of the contents of tissues that had been dissolved in sodium hypochlorite.

Tissue extracts were prepared from the same region of the dermis as that used for mechanical tests. The dermis was minced into pieces of about 1 mm³, and the mince was then extracted at seawater temperature in 5 volumes of EGTA–ASW for 90–180 min. A direct freeze–thaw extract was prepared by freezing the mixture of tissue and EGTA–ASW at –60 °C for at least 2 h, followed by incubation at seawater temperature until the liquid was completely thawed. These steps were repeated for a total of five freeze–thaw cycles. An EGTA–ASW extract (without freezing) was made by separating the tissue from the liquid after 180 min of incubation at seawater temperature. The tissue from this procedure was then resuspended in the same volume of fresh EGTA–ASW, and was frozen and thawed five times as described above. A direct freeze–thaw extract was also made from the longitudinal muscles of the body wall. In all cases, the extracts were clarified by centrifugation at 27 000 g for 30 min and were stored frozen.

In preparation for transmission electron microscopy, fresh and treated tissues were fixed for 36 h at 20 °C in 2.5 % glutaraldehyde, 0.1 mol l⁻¹ Mops, 0.41 mol l⁻¹ NaCl, 0.05 mol l⁻¹ MgCl₂, 0.01 mol l⁻¹ CaCl₂, 0.01 mol l⁻¹ KCl, pH 7.9. They were then rinsed for 24 h in several changes of the same solution lacking glutaraldehyde, post-fixed for 2 h in 1 % OsO₄, 0.1 mol l⁻¹ sodium cacodylate, pH 7.3, stained for 2 h in the dark in 0.5 % uranyl acetate in water, dehydrated in increasing concentrations of ethanol, and embedded in Spurr’s resin. Ultrathin sections were stained with uranyl acetate and lead citrate and examined in Hitachi H-600 electron microscopes.

All chemicals were reagent grade or better. Verapamil, 3,4,5-trimethoxybenzoic acid 8-(diethylamino)octyl ester (TMB-8), Mops and Triton X-100 were from Sigma Chemical Company, St Louis, MO, USA.

Fresh specimens incubated in 7.2 mmol l⁻¹ EGTA–ASW were markedly plastic in comparison with those incubated in ASW. The plasticization caused by EGTA was fully reversed by incubation in complete ASW (Figs 3, 4). This tissue thus shows the same mechanical responses to calcium removal and replacement that have previously been demonstrated in other echinoderm collagenous tissues. In the presence of 7.2 mmol l⁻¹ EGTA, the maximal free Ca²⁺ concentration in 1 ml of solution containing a standard specimen 0.09 cm×0.18 cm×3 cm, assuming that the specimen originally contained 10 mmol kg⁻¹ total calcium, was calculated to be 52 μ mol l⁻¹.

To determine whether the calcium-dependent stiffening effect would occur in tissues which did not contain viable cells, test specimens that had been incubated in EGTA–ASW for at least 90 min were subjected to three separate treatments that were predicted to lyse cellular membranes by different mechanisms. Some specimens were exposed to 1 % Triton X-100 in EGTA–ASW, others were exposed to deionized water for 30 min, and others were frozen (–60 °C) and thawed (seawater temperature) five times in EGTA–ASW. They were all subsequently evaluated in EGTA–ASW, using both creep and bending tests. All three treatments resulted in viscosity increases comparable to or greater than those seen when the tissues with living cells were returned to ASW. Qualitatively similar results were obtained using the two test modalities (Figs 3, 4). In bending tests, the water and freeze–thaw treatments consistently produced greater stiffening responses than did detergent treatment. The cause of these differences is not clear. It suggests that osmotic shock and freeze–thaw treatments might be more effective at lysing cell membranes.

Three principal types of electron-dense granules were seen in the cells of the deep dermis (Fig. 5A). The first type was round and small (approximately 200 nm in diameter) and very electron dense. The second type was larger (approximately 500 nm in diameter), frequently ellipsoidal and less electron-dense. The third type was larger still (greater than 1 μ m in diameter) and highly variable in electron density, even within a single granule. No two granule types were observed in the same cell, and they may therefore represent three separate cell types. Cell processes containing granules were enclosed by a continuous external lamina. The plasmalemmae of the granular cells were lysed by all three treatments (Fig. 5B–D). Some granules of all three types remained in the treated tissues, but the number of granules was markedly decreased. These treatments thus resulted in exposure of the extracellular matrix to the contents of the cells, including the contents of the granules, and coincidentally the tissue became stiff, even in the continual presence of the calcium chelator EGTA. These results suggested that the cells contain one or more substances capable of stiffening the extracellular matrix and that Ca²⁺ is not one of these substances.

To determine how effective the EGTA treatments were in removing calcium from the specimens, the total calcium content was determined by atomic absorption spectroscopy of identically treated specimens from five different animals. The results (mean ± S.D., N=5) showed that EGTA alone reduced the total calcium concentration from 8.9±1.2 to 0.58±0.22 mmol kg⁻¹ wet mass. Water, detergent and freeze–thaw treatments in the presence of EGTA reduced the total calcium further to 0.22±0.18, 0.14±0.04 and 0.08±0.10 mmol kg⁻¹, respectively. Hence, more calcium was chelated in those specimens in which the cells had been lysed. Presumably this was cellular calcium that had been released and made available to the chelator by cell lysis. The observation that the specimens which had been frozen and thawed had both the highest viscosities and the lowest total calcium contents made it very improbable that the release of calcium from lysed cells participated in the stiffening response.

The results just described suggested that the mechanical effect of Ca²⁺ replacement on live tissues is on one or more cellular processes, such as secretion, rather than directly on the extracellular matrix. To determine whether calcium-dependent cellular processes are involved in the stiffening responses of live tissues, specimens that had been plasticized by incubation in EGTA–ASW were subsequently incubated in ASW or in ASW containing verapamil or 3,4,5-trimethoxybenzoic acid 8-(diethylamino)octyl ester (TMB-8). Verapamil is known to block certain voltage-dependent Ca²⁺ channels in a number of different cell types (Triggle, 1981), while TMB-8 is known to have pleiotropic effects on calcium-dependent cellular processes (Malagodi and Chiou, 1974; Gordon and Chang, 1989; Ishihara and Karaki, 1991). At 1 mmol l⁻¹, both drugs blocked the stiffening effect of Ca²⁺ replacement, and their effects were reversed by returning the specimens to ASW without drugs. Both drugs were equally effective at plasticizing tissues that had not been incubated in EGTA–ASW, but had instead been placed directly into ASW containing the drug (Figs 6, 7). They were both maximally effective at doses approaching 1 mmol l⁻¹ and were ineffective at doses lower than 0.1 mmol l⁻¹. Neither drug affected the stiffness of tissues in which the cells had been lysed by freezing and thawing (not shown).

These results indicated that experimental manipulation of the extracellular Ca²⁺ concentration affects tissue viscosity indirectly, possibly through an effect on cell secretion, and that cell lysis causes tissue stiffening, even in the presence of a calcium chelator. It might be expected, then, that cell lysis would result in the release of a soluble stiffening factor that acts on the extracellular matrix. Evidence for the presence of a cell-derived stiffening factor was obtained from experiments in which fresh tissues were incubated for 90 min in EGTA–ASW to plasticize the extracellular matrix, followed by incubation in one of four distinct tissue extracts, made in the same solution. The extracts were obtained by freezing and thawing minced dermis (five cycles) in EGTA–ASW (freeze–thaw extract: FTE in Fig. 8), by extracting minced dermis in EGTA–ASW without freezing and thawing (EGTA extract: EgE in Fig. 8), by freezing and thawing the dermis residue from the previous step in fresh EGTA–ASW [EGTA–ASW-extracted tissue subsequently extracted by freezing and thawing in EGTA–ASW: (Eg)FTE in Fig. 8], and by freezing and thawing body wall muscles in EGTA–ASW (MuscE in Fig. 8). The results showed that the lysed cells released a factor that caused matrix stiffening in the presence of EGTA (Fig. 8). No stiffening effect was obtained in extracts of dermis made without freeze–thaw cycles or in freeze–thaw extracts of muscle. The stiffening factor was detected in four different preparations. It was stable to repeated freezing and thawing and was non-dialyzable (using dialysis tubing with a molecular mass cut-off of 14 kDa), but was destroyed by boiling for 15 min.

Previously published data have shown that the tensile stiffness of sea urchin spine ligaments (Hidaka, 1983; Diab and Gilly, 1984; Shadwick and Pollock, 1988), crinoid intervertebral ligaments (Wilkie, 1983), ophiuroid intervertebral ligaments (Wilkie, 1988, 1992), sea cucumber dermis (Eylers, 1982, 1989; Motokawa, 1984, 1988, 1994; Hayashi and Motokawa, 1986) and several other echinoderm collagenous tissues (Byrne, 1985) can be experimentally altered by manipulating the concentration of free Ca²⁺ in the bathing medium. It has been pointed out that the plasticizing effect caused by depleting the tissues of free Ca²⁺ could have resulted either from an inhibition of one or more Ca²⁺-dependent macromolecular associations in the extracellular matrix or from an inhibition of one or more cellular events, such as cell secretion (Motokawa, 1988; Wilkie, 1988). Nevertheless, the former explanation has been favored, and models have been published which hypothesize that calcium plays an essential role in the self-association of the highly sulfated glycosaminoglycans (GAGs), which are known to be present in the tissues (Wilkie, 1988, 1992). The increased self-association of these GAGs has been suggested to be the molecular mechanism underlying the increased stiffness or viscosity observed when the free calcium concentration is increased (Wilkie, 1988, 1992; Kariya et al. 1990; Motokawa, 1994). However, most of the previously published experiments showing a calcium-dependence of tissue stiffness have been performed using fresh tissues, containing viable cells, and it has thus been impossible to decide between the two different explanations of the results.

Motokawa (1994) used a combination of 1 % Triton X-100 and freeze–thaw treatments to lyse cells in the dermis of H. leucospilata. He compared the creep rates of specimens in which the cells had been lysed with those of intact specimens from the same animals. If the intact specimens were incubated in the presence of EGTA, they showed the same responses to large changes in the concentrations of monovalent cations as were shown by detergent-treated specimens. He interpreted these results as evidence for a cellular effect of extracellular Ca²⁺, which is in agreement with the present findings. However, he also found that Ca²⁺ chelation by EGTA significantly reduced the viscosity of both intact and detergent-treated specimens, and interpreted this to be evidence that Ca²⁺ plays a specific role as a stiffener of the extracellular matrix. This finding conflicts with the present results and with those of Szulgit and Shadwick (1994). The latter authors found that sea urchin spine ligaments responded to Triton X-100 by stiffening in the continual presence of EGTA. Moreover, the spine ligaments have also been observed to stiffen after freezing and thawing, using a protocol similar to that described in this report (G. K. Szulgit, personal communication). The causes of these different findings are not clear. Species variation might be important since Motokawa found that water (osmotic lysis) and freeze–thaw treatments of H. leucospilata were less effective in cell lysis than was Triton X-100.

The three methods employed in the present study to destroy cellular functions were all predicted to produce membrane disruption and cell lysis, leading to a release of those cytoplasmic, granular and nuclear contents that are soluble in artificial sea water containing EGTA. Triton X-100 produces chemical lysis by dissolving membrane lipids (Helenius and Simons, 1975). Freezing and thawing lyses membranes, probably through the physical damage caused by ice crystals (Mazur, 1966). Water produces osmotic lysis because of the high solute concentrations within cells and cell organelles. Electron microscopy confirmed that cell and organelle lysis resulted from all three treatments. That all three methods of lysis caused the stiffness of the dermis to increase even in the presence of EGTA strongly suggested that the cells contain one or more stiffening factors that affect the extracellular matrix independently of the extracellular Ca²⁺ concentration. Because the stiffening effect is observed in the presence of EGTA, it is clear that released intracellular Ca²⁺ alone could not be the stiffening factor. This conclusion is supported by the observation that the tissues with the highest viscosities were those that had the lowest total calcium concentration.

The plasticizing effect of EGTA and the stiffening effect of Triton X-100, water and freeze–thaw treatments were qualitatively similar in creep and bend tests. Creep tests of intact specimens in EGTA exhibited a rapid initial rate of elongation which, in some animals, became slower after 5–10 min. Fig. 3B shows an example of this phenomenon, which was not seen in specimens in which the cells had been lysed. This loading response could indicate that the effector cells in the dermis possess one or more mechanotransduction mechanisms that operate – independently of the concentration of extracellular Ca²⁺ – to stiffen the tissue in response to internal shearing strains. These effects could not be observed in bend tests, because these only measured what would have been equivalent to the initial deformation rates seen in the creep tests. Because of this difference, the results of the creep and bend tests were quantitatively different, although they showed the same qualitative effect of the different treatments. It might be thought that the different ionic treatments used (EGTA–ASW and ASW) could have produced their effects by artifactually causing either swelling or shrinking of the specimens. Such effects could potentially be large, because the tissue is a discontinuous fiber-reinforced composite material (Smith et al. 1981; Trotter and Koob, 1989; Trotter et al. 1994), in which the viscosity (or shear strength) of the interfibrillar matrix as well as the distance between fibrils both strongly affect the resistance to tensile loads. The effects on bending could be greater than those on creep rates because the rate of bending would change with the mass of the beam, which would change with the amount of water in the tissue. Although these considerations would carry considerable weight in any effort to deduce from the data the absolute material properties of the specimens, they are of much less concern in the comparative studies reported here. This is because the comparisons from which the main conclusions have been drawn were made between treatments in which the ionic conditions were identical. Thus, verapamil-and TMB-8-treated specimens were much more plastic in ASW than were untreated specimens in ASW; and Triton X-100-, water-and freeze–thaw-treated specimens were much stiffer in EGTA–ASW than were untreated specimens in EGTA–ASW. Finally, and perhaps most significantly, specimens in EGTA–ASW containing a freeze–thaw extract of dermis were much stiffer than were specimens in EGTA–ASW containing either an EGTA–ASW extract of dermis or a freeze–thaw extract of muscle. These results cannot have been caused by differential water contents in tissues exposed either to ASW or to EGTA–ASW.

A major prediction of the extracellular matrix calcium hypothesis is that the tissue ought to remain stiff in the presence of pharmacological agents which inhibit or disrupt cellular calcium-dependent events, provided that the calcium concentration in the extracellular matrix is kept within a normal range. Although the tissues remained stiff for many hours after they had been dissected from the animals and stored in ASW, they became plastic after 90 min in ASW containing EGTA and they became stiff again when the normal calcium concentration was restored. Significantly, the stiffening observed when calcium was restored was inhibited by the presence of verapamil or TMB-8. Neither of these drugs would be predicted to interact directly with components of the extracellular matrix, and neither had a plasticizing effect on tissues that had been frozen and thawed. The plasticizing effects of both drugs were reversed by washing them out of the medium. They are, therefore, unlikely to have affected cell viability. Importantly, both drugs also caused a reversible plasticization of tissues continuously maintained in ASW. This result showed that interference with cellular signalling pathways not only prevented the stiffening response to the restoration of normal Ca²⁺ concentrations, but also caused a reversible plasticization of tissues in which the Ca²⁺ concentration was unchanged. These results are inconsistent with the prediction of the extracellular calcium regulation hypothesis and strongly suggest that the effects on tissue stiffness caused by experimental modulation of the calcium concentration are cellular effects.

Although the concentrations of verapamil required to affect the dermis of C. frondosa were much greater than those required to inhibit voltage-sensitive Ca²⁺ channels in mammals (Triggle, 1981), they were similar to the concentrations of organic channel-blockers required to block the voltage-sensitive Ca²⁺ channels in echinoderm nerves (Berrios et al. 1985). Thus, a Ca²⁺ channel might be involved in the stiffening response. It should also be noted, however, that high concentrations of verapamil may inhibit the activity of protein kinase C by interfering with its interaction with phosphatidylserine-containing membranes (Mori et al. 1980). TMB-8 affects multiple cellular pathways, including Ca²⁺ movement across membranes and phosphatidylserine-stimulated protein kinase C activity (Malagodi and Chiou, 1974; Kojima et al. 1985; Gordon and Chiang, 1989; Ishihara and Karaki, 1991). Therefore, no specific site of action is indicated by the experimental results. The results with both drugs are consistent with an active stiffening mechanism that involves the translocation of Ca²⁺ across the cell membrane and/or with the activation of protein kinase C.

Taken together, the results suggest that, in this experimental model, the dependence of extracellular matrix stiffness on experimentally manipulated changes in the extracellular Ca²⁺ concentration is cell-mediated. Diab and Gilly (1984) and Szulgit and Shadwick (1994) have come to the same conclusion concerning the role of Ca²⁺ in the spine ligaments (‘catch apparatus’) of sea urchins, and Motokawa (1994) has found that Ca²⁺ affects cellular activity in sea cucumber dermis. Diab and Gilly (1984) also found that the polyamines putrescine and cadaverine, at a concentration of 10 μ mol l⁻¹, inhibited the stiffening of sea urchin ligaments in response to mechanical agitation. They suggested that transglutaminase-mediated protein crosslinking might be part of the stiffening response. More recent studies have shown, however, that polyamines are potent inhibitors of cation – including Ca²⁺ and K⁺ – channels in a number of cell membranes (Ficker et al. 1994; Scott et al. 1993) and also affect other membrane functions (Schuber, 1989). Diab and Gilly’s results are thus also consistent with the cell membrane being a site for stiffening inhibition in the presence of normal extracellular Ca²⁺ concentrations. In the sea urchin spine ligament (‘catch apparatus’) as well as in the sea cucumber dermis, then, the depletion of extracellular Ca²⁺ or interference with normal cellular membrane functions causes the tissue to become plastic, while restoration of the Ca²⁺ or reversing the inhibition of membrane functions stiffens the tissue.

It is generally believed that the granules of the cells observed in the dermis, and in other mutable collagenous tissues of echinoderms, secrete the substances that regulate the stiffness of the interfibrillar matrix (Wilkie, 1979, 1984, 1988; Holland and Grimmer, 1981; Smith et al. 1981; Motokawa, 1982a, 1984, 1988; Hidaka and Takahashi, 1983). Since all of the cells were lysed by the treatments used in the present studies, it is impossible to attribute the stiffening effect to any specific granule, nor was that an objective of the experiments. It is clear, however, that the stiffening effect of cell lysis is not due to the release of Ca²⁺, since the total calcium contents of tissues in which the cells had been lysed were lower than those of tissues that had not been exposed to lytic agents. Thus, the calcium identified histochemically in certain granules of the cells in brittlestar ligaments (Wilkie, 1979) and in the dermis of the sea cucumber Stichopus chloronotus (Matsuno and Motokawa, 1992) is likely to have a function other than as a secreted stiffening agent. Because the quantity of calcium identified by pyroantimonate precipitation in the vacuoles of the ‘vacuole cells’ in the dermis of S. chloronotus seemed to be reduced in tissues stimulated to become stiff, it was thought possible that this cell regulates dermal stiffness by secreting and sequestering calcium (Matsuno and Motokawa, 1992). If S. chloronotus dermis is physiologically similar to C. frondosa dermis, the loss of Ca²⁺ from the vacuoles may be related to a secretory process, but the calcium per se is probably not the stiffening agent. No ‘vacuole cells’ were observed in C. frondosa dermis and they were not observed in an earlier study of S. chloronotus dermis (Motokawa, 1982b).

The results reported here are consistent with the existence of an organic stiffening agent that acts on the matrix in calcium concentrations that are at least two orders of magnitude below that of sea water. The need for cell lysis and the failure to obtain stiffening activity in lytic extracts of muscle support the idea that the stiffener is a specialized dermis cell product. The observations that the activity of the stiffener is non-dialyzable and is destroyed by boiling suggest that it is macromolecular, possibly protein, and unlikely to be a neurotransmitter. Although stiffening was assayed in tissues containing living cells and the activity theoretically could have been a cellular or an extracellular effect, it is more likely to have been extracellular because it occurred in the presence of EGTA, which consistently plasticizes tissues containing living cells. In these respects it differs markedly from the ‘stiffening factor’ prepared from sea cucumber coelomic fluid (Motokawa, 1982a), which was thought to act via the nervous system. Experiments are under way to purify and characterize the macromolecule(s) responsible for the stiffening activity, to determine the site or sites of action and to determine whether a plasticizer can also be identified and characterized.

This research was supported by grants from the National Science Foundation and the Office of Naval Research. The authors thank Robert Shadwick, Greg Szulgit, Gillian Lyons-Levy, Stephen Andrews and three anonymous reviewers for their helpful comments on the manuscript.