Synthesis and secretion of proteoglycans by cultured chondrocytes

Experimental Cell Research 148 (1983) 493-501

Synthesis

and Secretion of Proteoglycans Cultured Chondrocytes

Effects of Monensin,

K. MADSEN,

Colchicine

S. HOLMSTRGM

by

and P-D-Xyloside

and K. OSTROWSKI”

Depurtment of Histology, Kurolinska Instituiet, S-10401 Stockholm, Sweden

Chondrocytes, isolated from elastic ear cartilage of young rabbits, were grown in monolayer cultures in Ham’s F-12 medium. Synthesis and secretion of macromolecules were monitored by labelling with radioactive precursors and the effect of monensin and other experimental agents was investigated. Monensin caused an inhibition of the incorporation of precursors into macromolecular material and a moderate intracellular accumulation when used in higher concentrations. The effect was more pronounced for 3’S04 than for ‘H-labelled glucose or proline. p-Nitrophenyl-/?-o-xyloside alleviated this inhibition to some extent, but there was a concomitant increase in the amount of intracellular labelled material. Colchicine and monensin together caused a severe inhibition of the incorporation of 35S04 and a marked shift of the label to the intracellular compartment. Colchicine also increased the sensitivity of the cells to monensin, lowering the minimal effective concentration about one order of magnitude. The latter results are consistent with the idea that cytoplasmic microtubules have a stabilizing function on the secretory pathways and, that their removal by colchicine. causing a ‘randomizing’ of the Golgi complex, makes these pathways more vulnerable to monensin.

According to the current model [l, 21 cartilage proteoglycans are complex macromolecules in which chondroitin sulfate and keratan sulfate chains are attached to a protein core, giving a monomer molecular weight (M,) of about 1-2x 106. Extracellularly, the proteoglycan monomers form large aggregates with hyaluronic acid and link glycoproteins. Primary cultures of rabbit ear chondrocytes produce similar proteoglycans [3]. The synthesis and secretion of proteoglycans by chondrocytes are believed to occur by the general pathway suggested by Palade [4]. Thus, synthesis of the core protein and the first steps in glycosylation occur in the rough endoplasmic reticulum. The product is then transported via the Golgi complex to the cell surface, where it is released by exocytosis. During this transport, glycosaminoglycan chain prolongation and sulfation occurs [I. Experiments with colchicine and other anti-microtubular drugs have shown that cytoplasmic microtubules are important for the synthesis and secretion of proteoglycans from chondrocytes [6-B] and for the structural integrity of the Golgi complex [9, 101. For example, the secretion of 35S-labelled proteoglycans was inhibited by 30-50 % and there was a moderate intracellular accumulation of labelled material when chondrocytes in culture were treated with 10 uM colchi* On leave from Department of Histology and Embryology, Medical School, Warsaw, Poland. Copyright 0 1983 by Academic Preat. Inc All right\ of reproductwn ,n any form rewrved ,X,14-4x27’83 $03.00

494 Madsen, Holmstrbm und Ostrowski tine [7]. Further experiments with colchicine and B-D-xylosides [ll, 121 have indicated that the role of microtubules in secretion might be facilitatory rather than obligatory, and/or that there might be alternative, non-microtubuli-dependent pathways for the secretion of proteoglycans. There was also an inhibition of the synthesis of proteoglycans, which could be the effect of the disturbances in the Golgi complex. Recently there has been a considerable interest in the effect of monensin on the secretion of proteins from various cells. Monensin is a carboxylic ionophore with affinity to monovalent cations, especially Na+ and H+ (for a review. see Pressman [13]). This drug produces characteristic changes in the Golgi complex, viz. an abundance of large irregular vacuoles and an absence of normal cisternae [14--161, presumedly because of a partial NaiK equilibration. Monensin also severely inhibits the secretion of a number of proteins, including immunoglobulins [14-161, procollagen [17, 181, fibronectin [l&20] and virus envelope glycoprotein [21]. The effect on proteoglycans appears to be more complex. Tanzer and co-workers [22, 231, using cultures of chick sternal chondrocytes, found that monensin caused a marked decrease in sulfation of the glycosaminoglycan side chains and only moderate inhibition of glycosylation and secretion. However, in a recent paper [24], evidence was presented that monensin in rat chondrosarcoma cells caused an intracellular accumulation of proteoglycan core protein which was both underglycosylated and undersulfated. It has also been demonstrated that hyaluronic acid synthesis in rat chondrosarcoma cells was not inhibited even by high concentrations of monensin, whereas proteoglycan synthesis was [25]. In the present paper, monensin was used as a tool to further investigate the role of the microbubules and the Golgi complex for secretion of proteoglycans from rabbit ear chondrocytes in culture. MATERIALS

AND METHODS

Chemicals and Isotopes Monensin was obtained from Calbiochem AG. Lucerne, Switzerland, and colchicine from Serva Feinbiochemica, Heidelberg, West Germany. p-Nitrophenyl-/I-o-xylopyranoside was purchased from Koch-Light Laboratories Ltd, Colnbrook, Bucks., UK. Solutions of guanidine-HCI (Gu-HCI) were prepared from a ‘practical grade’ reagent (Sigma Chemical Co.. St. Louis, MO.) and filtered with charcoal prior to use. “SO4 (carrier-free), o-[6-3H]glucose (12.8 Ciimmol) and L-[S-‘Hlproline (35 Ciimmol) were obtained from Amersham International Ltd, Amersham, Bucks.. UK. Safefluor liquid scintillation cocktail was obtained from Lumac B.V.. Schaesberg, The Netherlands. Other chemicals used were of reagent grade.

Cell Culture Four week-old, New Zealand white rabbits were obtained from a local supplier. The animals were killed with sodium pentobarbital and the ears were cut off at the base. After soaking the ears in 0. I % chlorhexidine for 20 min, they were rinsed in PBS and chondrocytes were isolated from the auricular cartilages as described previously [3]. In brief, the procedure consists of cleaning the cartilage from skin and most of the perichondrium, digestion in 0.1% clostridium collagenase (Sigma, type IA) for 4540 min at 37°C. and further cleaning of the cartilage under the dissecting microscope. Finally. the Exp Cd

RPS 148 11983)

Effect of monensin on proteoglycan secretion

495

cartilage was cut into small pieces and digested overnight in 0. I % collagenase, dissolved in complete culture medium including 10% (vol/vol) serum. The liberated chondrocytes were washed three times with culture medium, counted in a hemocytometer, and plated in 35 mm plastic Petri dishes, 24-well tissue culture plates. or in 25 cm’ culture flasks (Falcon Plastics, Los Angeles, Calif.) at an initial density of 1.5~ IO5 cells/cm’. The cells were cultured in F-12 medium [26] supplemented with IO% heat-inactivated newborn calf serum (Flow Labs Ltd, Irvine. Scotland). The medium was further supplemented with Hepes (IO mM), Tes (IO mM), gentamycin (Sigma. 100 ugiml), and L-ascorbic acid (50 ugiml). The medium contained 6 uM SOi- and 10 mM glucose. The chondrocytes were cultured at 37°C in air containing 5 % COz. Experimental incubations were performed in fresh medium on the day after plating.

Experimental Procedure Stock solutions were made of monensin and colchicine in ethanol, and of the xylosides in DMSO. The experiments were initiated by changing the culture medium and adding the appropriate drugs. To control cultures an equal amount of ethanol and/or DMSO without the drug was added. The cultures were preincubated with the drugs for 90 min. Radioactive precursors, “SO4 in combination with [3H]glucose or [‘Hlproline, IO &i/ml of each, were then added. The cultures were incubated-for 4-6 h at 37”C, being then terminated by removing the medium and rinsing the dishes with cold PBS. In the quantitative experiments, the pericellular matrix was removed by collagenase digestion and the chondrocytes were then lysed in I M KOH at 60°C for 60 min [27]. Subsequently, the amount of proteoglycans was measured separately in medium, pericellular matrix and cell lysate. The cell lysate was also used for protein determination [28].

Analytical Procedures The amounts of macromolecular “S and ‘H activity were determined by desalting on small columns of Sephadex G-25 (PD-IO, Pharmacia Fine Chemicals, Uppsala) eluted with 4 M Gu-HCI [I I]. 0.5 ml aliquots of the desalted material were then mixed with 0.2 ml ethanol and 4 ml Safefluor and counted in a liquid scintillation spectrometer. Proteoglycans were isolated by isopycnic CsCl centrifugation under associative conditions [3]. Briefly, the culture medium was decanted, lyophilized and then dissolved in 0.4 M Gu-HCI, 0.05 M NaAc buffer, pH 5.8, containing CsCl to a density of 1.67 g/cm3. The cell layers were extracted with 4 M Gu-HCI, 0.05 M NaAc buffer, pH 5.8, at 4°C for 24 h and then dialyzed against 9 vol of buffer. The density of the extracts were adjusted with CsCl and the pH of both media and extracts was corrected with 5 M acetic acid containing CsCI. In order to minimize proteolytic degradation, protease inhibitors (EDTA, 50 mM; 6-amino-hexanoic acid, 100 mM; benzamidine-HCI, 5 mM: PMSF, 1 mM) were included in all solutions. Centrifugation was performed in a Beckman 50Ti rotor at 48000 rpm for 48 h at 15°C. Size distribution of the material was determined by chromatography on a 0.6~ I10 cm column of Sepharose CL 2B (lot 7589), before and after reduction and alkylation [29]. The column was eluted with 0.5 M NaAc, pH 7.0, at 3.5 ml/h.cm’, collecting 0.5 ml fractions. Void volume (V,) and included volume (V,) was determined with proteoglycan aggregate (bovine nasal septum, Al) and ‘HZ0 respectively. For analysis of the degree of sulfation of the chondroitin sulfate side chains, aliquots of the material were digested with papain, precipitated with CPC in the presence of carrier chondroitin sulfate. and digested with chondroitinase AC-II (Seikagaku fine Biochemicals, Tokyo. Japan, lot KE4802). The digests were then separated by paper chromatography [27]. A few samples were also analyzed by ion exchange chromatography on Aminex A-25 [30]. For determination of hydroxyproline/proline, samples were hydrolyzed in 6 M HCI at 110°Cfor I6 h and analyzed by ion exchange chromatography on Aminex Ql5OS [7].

RESULTS Effect of Monensin on 35SO4 and [3H,lGlucose incorporation Monensin caused an inhibition of the total incorporation of both 35S04 and [3H]glucose into macromolecular material in the rabbit chondrocyte cultures (fig. 32-838337

Exp Cell RPS 148 (1983)

496

Madsen, Holmstriim and Ostrowski

E

2 -I!0

Monensin

cont.

(PM)

Monensin

cont.

10-a

10-7

16

(MI

Fig. 1. Relation between the concentration of monensin and the amount of ‘5S and ‘H activity incorporated in chondrocyte cultures, labeled with 3”S0, and o-[6-3H]giucose. Radioactivities were determined separately in c, medium; h, pericellular matrix (collagenase digest); and (I, cell lysate. The values are the means of five dishes. Fig. 2. Influence of 10 FM colchicine on the relation between monensin concentration and effect on radioisotope incorporation. Cultures were labeled with j5S04 and L-[5-‘Hlproline. Each point is the mean of six determinations. (A) 35Sactivity; monensin. (B) “S activity; monensin and colchicine. (0 ‘H activity; monensin. (D) 3H activity; monensin and colchicine. 0, Cell lysate; n , matrix; 0, medium; 0, mathematical sum of the other three.

1). The drug affected 3sS incorporation much more than that of 3H. The inhibitory effect showed an apparent maximum at a monensin concentration of 1 \tM. with values of -89 % and - 16% respectively for the two tracers. For 3sS, the inhibitory effect was most pronounced in medium and matrix. At 10 PM there was an increase in the amounts of intracellular radioactivity, again most evident for 35S. The different effects of monensin of 35S04 and [3H]glucose incorporation are in agreement with the findings of Tajiri et al. [22]. The increase in radioactivity in the cell lysates at 10 VM of monensin could conceivably be the result of intracellular accumulation of proteoglycans or precursors. Intracellular accumulation due to monensin has been reported for a number of secretory proteins [14-211 and, recently, for proteoglycan core protein 1241. Influence of /?-D-Xyloside on Monensin-Treated Cultures In order to localize the level of the synthesis or transport block induced by monensin, p-nitrophenyl-B-D-xyloside was used as a probe. This drug acts as an Exp Cdl Res 148 (1983)

Effect of rnonensin on proteoglycan secretion

7

0

497

0.1 uM

E loE 4 3. Gel chromatography on Sepharose CL-2B of proteoglycans (Al fractions) isolated from unexposed chondrocytes and chondrocytes treated with monensin.

Fig.

exogeneous acceptor for the tirst galactosyl transferase, thereby bypassing the need for proteoglycan core protein [3 I]. If monensin affects mainly the availability of core protein and/or the initiation of glycosaminoglycan side chains, one would expect the synthetic block to be alleviated by the xyloside. On the other hand, if the synthetic block is concentrated to glycosaminoglycan chain elongation and sulfation, the effect of xyloside would be minimal. Finally, delayed intracellular transport would show up as a partial alleviation of the block by xyloside, accompanied by an accumulation ‘proximal’ to the transport obstacle. As shown in table 1, the xyloside partly overcame the inhibition by monensin. Cultures exposed to monensin were able to increase the total incorporation of j5S04 by 106% when also treated with xyloside. Furthermore, there was a concomitant intracellular increase of the tracer by 49%. This indicates that. although an artificial increase in the concentration of primers for the glycosaminoglycan chain-elongation enzymes results in an increased glycosaminoglycan synthesis, the capacity of the secretory pathways is lowered irrevocably by monensin. In the xyloside-treated cultures, there was also a marked shift in the distribution of labelled material from the matrix to the medium. This is consistent with the mechanism of xyloside action, i.e. production of free glycosaminoglycan chains [3 I]. tnteraction between Monensin and Colchicine When chondrocyte cultures were treated with both monensin and colchicine, the impairment of secretion was severe (table 1). Whereas the decrease in total sulfate incorporation was moderate (25%) and within experimental errors, there was a dramatic shift of 35S activity to the intracellular compartment. Thus, intracellular 3sS activity increased by 23% even when compared to untreated controls, and by 134% when compared to monensin-exposed cultures. The ability of colchicine to influence the inhibitory effect of monensin was Exp Cdl Res 148 (1983)

498

Madsen, Holmstriim

Table 1. Injluence

and Ostrowski

of B-D-xyloside

and colchicine

on monensin-treated

Cell lysate

A Control

B Monensin C Xyloside + monensin D Colcicine + monensin C vs B D vs B

cultures

Matrix

dpm

SD

A%

dpm

SD

A%

2 075 I 088 I 624 2 546

180 390 215 335

-4Xh -22” + 23”

65 688 5 062 2 752 I 326

II 370 3 024 182 7s

-92” -96’ -98’

+49” + 134h

-46 -74’

Each value is the mean of five dishes. Statistical significance was tested with Wilcoxon’s rank sum tes n p
further investigated by incubation of chondrocyte cultures with various concentrations of monensin in the absence or presence of 10 uM colchicine. For this experiment the chondrocytes were plated into 24-well tissue culture plates with the same initial density as before. The cells were plated in 1 ml medium and the experiment was done with 0.5 ml of medium per well. The cultures were preincubated with the drugs for 90 min, after which 20 uCi/ml each of 35S04 and L-[S3H]proline was added. Incubation was continued for a further 6 h and the amount of macromolecular radioactivity was then determined as described. Selected samples were also taken for hydroxyprolineiproline determination. As shown in fig. 2A, monensin inhibited 35Sincorporation in a dose-dependent manner with half-maximal effect at about 5~ 10-s M. The effect on [3H]proline incorporation was less marked, and mainly seen in the extracellular material (fig. 2 C). Analysis of the extracellular material by ion exchange chromatography gave a 4-hydroxyproline/proline ratio of 0.8.5-0.90, indicating that most of this material was collagen. In contrast, the intracellular material contained a very low proportion (<0.02) of 3H-labelled hydroxyproline, indicating that only a small part of the intracellular radioactivity represents hydroxylated collagen. The proportion of hydroxyproline varied only slightly with changes in monensin concentration. Colchicine potentiated the inhibition of 3sSincorporation caused by monensin, particularly in the low concentration range (fig. 2B). Consequently, the lowest concentration of monensin used (5~ lo-’ M) inhibited sulfate incorporation by 60% when colchicine was present, whereas this concentration of monensin was without effect in cultures not containing colchicine (fig. 2A). The potentiating effect by colchicine on monensin inhibition of [3H]proline incorporation was less marked (fig. 20), but clearly evident in the extracellular fractions. The proportion of hydroxyproline in the different fractions was not affected by colchicine. As shown in fig. 2B, D, the proportions of intracellular 35S and 3H-activity were both raised by colchicine, indicating an increased intracellular accumulation of macromolecular material. Exp Cdl Res 148 (1983)

Effect of monensin on proteoglycan secretion 499

Protein

Total

Medium dpm -

SD

A%

dpm

SD

A%

w

SD

b.5 574 3 565 15 601 3 434

3 635 I 763 6 872 437

-92h -66” -92h

113 337 9 71s 19 916 1 305

14 509 4 858 6 677 287

-91h -82’ -94h

242 226 246 230

21 17 9 21

+106” -25

+338’ -4

Qualitative Analysis of Proteoglycans and Glycosaminoglycans from Monensin-Treated Cultures For these analyses the chondrocytes were plated in 25 cm2 culture flasks at 1.5~ IO5 cells/cm’ and later labelled with 50 uCi/ml of D-[6-3H]glucose and “SO4 for 6 h. Proteoglycans were isolated from culture media and cell layer extracts utilizing isopycnic CsCl centrifugation, as described earlier, the Al fractions being analyzed by gel chromatography on Sepharose CL-2B. The elution profiles (fig. 3) indicated that the proportion of aggregated proteoglycans decreased when the cultures were treated with increasing concentrations of monensin. Similar results were reported by Mitchell & Hardingham [25]. In addition, the results of the disaccharide analysis by paper chromatography (table 2) and ion exchange chromatography (not shown) of chondroitinase digests indicate that the degree of sulfation was lowered, and that the 4S/6S ratio was increased, in the monensintreated cultures. These findings are similar to and corroborate those of other investigators [22-2.51. DISCUSSION Our results on the effect of monensin on proteoglycan synthesis and secretion in cultured chondrocytes derived from rabbit ear cartilage, i.e. a normal, mammaTable 2. Composition of chondroitinase AC-II digests

Control Monensin Monensin

0.1 uM I nM

Origin

ADi4,6diS

ADi4S

ADi6S

ADiOS

4Si6S

9 760.4 413.4 97.1

3 135.0 21.8 10.3

24 994.2 I 355.3 489.7

17 279.9 523.2 156.7

2.2 50.8

I.446 2.SYO 3.128

Al fractions isolated from the culture medium were digested with chondroitinase AC-II in the presence of 5 mg carrier bovine nasal septum, AI-DI. The digests were separated on Whatman No I paper in butyric acid: 2 M NH,OH 5 : 3 for 24 h. The figures are given as dpm ‘H.

500 Madsen, Holmstrijm and Ostrowski lian, elastic cartilage, are in good agreement with the findings of others 122-251 working with chondrocytes from other sources. The drug inhibits sulfate incorporation more than glucose incorporation and, in higher concentrations, causes a moderate intracellular accumulation of labelled material. The effect of monensin on the sulfation and size of the proteoglycans synthesized is also in accordance with previous reports. The finding that the amount of intracellular ‘H-labelled material was not lowered by monensin indicates that the drug was not toxic to the cells. Furthermore, the effects of monensin and P-D-xyloside in combination suggest that monensin blocks intracellular transport of proteoglycans, corroborating the findings of Nishimoto et al. 1241.Transport blocks by monensin have also been reported for a number of other secretory products [ 14-211. Colchicine potentiated the inhibitory effect on proteoglycan synthesis by monensin and the two drugs in combination caused a marked intracellular accumulation of labelled material, which neither did alone. The latter finding seems to indicate a more severe inhibition of the secretory process, although a part of the effect may also be due to a decrease in the intracellular turnover of proteoglycans or proteoglycan precursors. On the basis of the moderate effect of colchicine on proteoglycan synthesis and secretion from cultured chondrocytes, we have previously suggested that cytoplasmic microtubules may have a facilitatory, rather than obligatory. role in the secretion of proteoglycans or that alternative pathways exist [I 1, 121. The experiments described in this paper indicate that removal of the cytoplasmic microtubules by colchicine makes the secretory pathways more vulnerable to the effect of monensin. The major site of monensin action in the cell appears to be the Golgi complex [32]. Earlier experiments with colchicine 19. IO] and other anti-microtubular agents [33] have also demonstrated that functional microtubules are necessary for the structural integrity of the Golgi complex. Thus, disruption of cytoplasmic microtubules leads to the spreading of individual dichtyosomes more or less randomly in the cytoplasm. It is conceivable that such ‘randomizing’ of the Golgi complex decreases its efficiency and reduces its capability to handle additional strains on the secretory pathways. However, further experiments are required to decide whether this is the mechanism behind the observed synergism between colchicine and monensin or whether there is another common target for the two drugs. This work was supported by the Swedish Medical Research Council (proj. no. 03355), the M. Bergvall Foundation. the King Gustaf V 80th Birthday Fund and by the funds of Karolinska Institutet. We acknowledge the skilful technical assistance of Ms G. Carlsson and MS M. Engstrom. The manuscript was revised by Dr Paul Holmes.

REFERENCES I. Hascall, V C. J supramol struct 7 (1977) 101. 2. Hascall, V C & Heinegbrd, D K, Glycoconjugate research red R Jeanloz & J Gregory) Academic Press, New York (1979). 3. Madsen, K & Lohmander. S, Arch biochem biophys 196 (1979) 192.

p. 341.

Effect

of monensin

on proteoglycan

secretion

501

4. Palade, G, Science 189 (1975) 347. 5. Rodtn, L, The biochemistry of glycoproteins and proteoglycans (ed W H Lennarz) p. 267. Plenum. New York (1980). 6. Jansen, H W & Bornstein, P, Biochim biophys acta 362 (1974) 150. 7. Lohmander, S, Moskalewski. S, Madsen, K, Thyberg, J & Friberg, U, Exp cell res 99 (1976) 333. 8. Madsen, K. Moskalewski. S. Thyberg, J & Friberg. U, Experientia 35 (1979) 1572. 9. Moskalewski, S, Thyberg, J, Lohmander. S & Friberg, U, Exp cell res 95 (1975) 440. 10. Thyberg, J, Nilsson, S. Moskalewski, S & Hinek, A, Cytobiologie 15 (1977) 175. I I. Lohmander, S, Madsen, K & Hinek, A, Arch biochem biophys 192 (1979) 148. 12. Madsen, K. Arch biochem biophys 211 (1981) 368. 13. Pressman, B C, Ann rev biochem 45 (1976) 501. 14. Tartakoff. A M, Vassalli, P & DetrBz, M. J exp med 146 (1977) 1332. 15. - J cell biol 79 (1978) 694. 16. - Ibid 83 (1979) 284. 17. Uchida. N, Smilowitz, H & Tanzer, M L. Proc natl acad sci US 76 (1979) 1868. 18. Uchida, N, Smilowitz, H, Ledger, P W & Tanzer, M L, J biol them 255 (1980) 8638. 19. Ledger, P W, Uchida, N & Tanzer, M L, J cell biol 87 (1980) 663. 20. Virtanen. I, Vartio, T, Badley, R A & Lehto, V-P. Nature 298 (1982) 660. 21. KZri%inen. L, Hashimoto, K, Saraste, J. Virtanen, I & Penttinen, K, J cell biol 87 (1980) 783. 22. Tajiri, K, Uchida, N & Tanzer, M L, J biol them 255 (1980) 6036. 23. Kajiwara. T & Tanzer. M L, FEBS lett 134 (1981) 43. 24. Nishimoto, S K, Kajiwara, T & Tanzer, M L. J biol them 257 (1982) 10558. 25. Mitchell. D & Hardingham, T, Biochem j 202 (1982) 249. 26. Ham. R G, Proc natl acad sci US 53 (1965) 288. 27. Madsen, K, Moskalewski, S, von der Mark, K & Friberg, U, Dev biol 96 (1983) 63. 28. Bradford. M M. Anal biochem 72 (1976) 248. 29. Heinegsrd, D, .I biol them 252 (1977) 1980’. 30. Lohmander, S, Antonopoulus, C A & Friberg. U, Biochim biophys acta 304 (1973) 430. 31. Schwartz. N B, Galligani, L, Ho, P L & Dorfman. A. Proc natl acad sci US 71 (1974) 4047. 32. Tartakoff, A M, Trends biochem sci 7 (1982) 174. 33. Thyberg, J, Blomgren. K. Hellgren. D & Hedin, U. Eur j cell biol 27 (1982) 279. Received March 30, 1983 Revised version received June IS. 1983

Exp Cell Ra.7 148 liYX3J