Absorption of Intact Protein by Colonic Epithelial Cells of the Rat Bonnie S. Worthington, PhD, and Cyril Enwonwu, ScD

Colonic absorption of intact protein was examined in adult rats using histological and ultrastructural procedures. Horseradish peroxidase was introduced into ligated colonic loops and retained therein for 5, 10, 20, or 30 minutes prior to excision of the loops, and their processing for microscopy. Morphological evaluation revealed evidence of. peroxidase absorption via pinocytosis. Tracer particles were observed adherent to the mucosal border, in apical pinocytotic vesicles, in vesicles adjacent to and fusing with lateral and basal cell membranes, in extracellular spaces throughout the mucosa, in lymphatic channels of the submucosa, and occasionally in blood capillaries of mucosal and submucosal regions. The significance of these findings is discussed in light of the frequent presence of dietary and/or microbial macromolecules in the luminal milieu of the large intestine. It is suggested that pinocytotic uptake and subsequent vesicular transport and exocytosis of intact protein may occur in the colon of some species, and that such a phenomenon may be responsible for penetration of the mucosal barrier by macromolecular antigens or toxins. Morphological and physiological characteristics of the colon have been well described (17), but its frequent involvement by neoplastic and inflammatory processes has provoked the continued interest of numerous investigators (8-16). Aside from its role in the elaboration of mucus, the chief function of the colon is absorption of water and sodium from the incoming chyme, and secretion of potassium and bicarbonate into the residuum (1, 17-21). The digestive role of the colon is believed to be minimal since foodstuffs entering the gastrointestinal tract via the mouth are largely degraded and absorbed long before leaving the small bowel. A nutritional enema is of doubtful value because

glucose and amino acids contained in such preparations are not absorbed in measurable amounts by the colonic mucosa (1). Trace nutrients synthesized by colonic flora are partially absorbed (1); free fatty acids (1, 22-24) and bile acids (1) likewise enter the mucosa in significant amounts. Ultrastructural features of the colonic absorptive cell are indicative of its transport capabilities (2, 4-7). In the late 1800's, Metchnikoff proposed that a large number of the ills of man are due to absorption of microbial products from the large intestine (25). T h e validity of this statement has been repeatedly challenged but bacterial fermentations are known to explain the release and absorption of intestinal gases and other From the Departments of Nutrition and Oral Biology, small toxic compounds (1, 26-28). In recent University of Washington, Seattle, Washington. years, researchers have theorized that the high Supported in part by the University of Washington Graduate School Research Fund and the Nutrition Founda- incidence of carcinoma of the colon in developed countries is related directly to the low-residue tion, Inc. Address for reprint requests: Dr. Bonnie Worthington, nature of the typical diet (14-16). Such a diet is Child Development and Mental Retardation Center, Nutrition Section, Wj-10, University of Washington, Seattle, known to promote minimal intestinal peristaltic activity with increased opportunity for microWashington 98195.


Digestive Diseases, Vol. 20, No. 8 (August 1975)


Fig 1. Light micrograph of a toluidine blue-stained resin section from a colonic loop showing a simple epithelium with smooth surface (S) and deep crypts (C). Numerous goblet cells (arrows) are interspersed between surface and crypt absorptive cells. ( • 150)

bial production and retention of carcinogenic compounds (14-16). While small toxic or carcinogenic substances may be absorbed by simple diffusion or active transport, these processes alone could not explain the movement of larger molecules across epithelial membranes, Absorptive cells of the jejunum and ileum are known to engulf intraluminal macromolecules by pinocytosis (2935), but a comparable event has not been demonstrated in more distal regions of the gut. The present work, therefore, tests the hypothesis that colonic epithelial cells absorb intact protein by pinocytosis. A well-known cytochemical marker, horseradish peroxidase (MW: 40,000), was used in laboratory rats to evaluate the degree of uptake and transport, using light and electron microscopic procedures. MATERIALS AND METHODS Animals and Diets Male Sprague-Dawley rats (Simonson Laboratories, Gilroy, California) weighing 100 to 150 grams were used in the study. T h e animals were housed individually in an air-

Digestive Diseases, Vol. 20, No. 8 (August 1975)

conditioned room with lighting regulated to provide equal hours of light and dark. During the first week in our laboratory, the rats were fed a commercial ration (Purina Laboratory Chow, Ralston Purina Co., St. Louis, Missouri) ad libitum and at the end of this period, the rats whose body weights had not increased at a normal rate were eliminated from the study. The remaining ten animals were maintained on the commercial ration and experimental procedures were conducted over the course of the subsequent four-week period.

Experimental Format Colonic absorption of intact protein was evaluated using ligated intestinal loop procedures previously described (3436). Briefly, each animal was anesthesized intraperitoneally with sodium pentobarbital (4 rag/100 g body weight), a laparotomy was performed to expose the large intestine, and thread ligatures approximately 2-4 cm apart were placed around the circumference of the large bowel at a location approximately 4 cm distal to the ileocecal valve. After flushing the lumen several times with saline solution, Sigma type VI horseradish peroxidase (5 m g / m l normal saline solution) was introduced into the ligated segment of the colon. Each loop preparation was retained in vivo for 5, 10, 20, or 30 minutes, after which time the central portion of each loop was excised and processed for light and electron microscopy. Controls consisted of segments of colon which (1)



Fig 2. Light micrograph from an unstained preparation of a control loop showing endogenous peroxidase activity in erythrocytes (E) of the mucosa and occasionally around goblet cells in the epithelium (arrows). (X 200)

had received no peroxidase or (2) had received peroxidase, but were incubated in solutions from which hydrogen peroxide or diaminobenzidine had been omitted. All loop preparations were completed between nine and ten o'clock in the morning. This time was recognized as the approximate midpoint of a "liquid" feces period of the day.

Tissue Processing Excised colonic tissue was immersed in a cold solution of 3% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.4, diced into 1 mm cubes, and placed in fresh fixative for 1.5 hours at 0~ C. The small tissue cubes were rinsed in 0.1 M cacodylate buffer, pH 7.4, for about 18 hours and then sectioned into slices of 40-75 #m thickness, using a Smith-Farquhar tissue chopper (Ivan Sorvall, Inc., Norwalk, Connecticut). Slices were then incubated for 30 minutes at room temperature in a solution containing 5 mg of diaminobenzidine tetrahydrochloride dissolved in 10 ml of tris-HC1 buffer, pH 7.6, after which they were transferred to a fresh solution of diaminobenzidine containing 0.1 ml of 1% hydrogen peroxide and incubated for a further 15 minutes. The slices were then washed three times in 0.1 M cacodylate buffer and postfixed for one hour in 1% osmium tetroxide in cacodylate buffer, pH 7.4, at 0~ C. Tissue slices


were then stained en bloc with 1.5% uranyl acetate for 1.5 hours, dehydrated in ethanol and propylene oxide, and embedded in Epon 812 (37). Thick and thin sections were cut on a Sorvall MT2-B ultramicrotome (Ivan Sorvall, Inc., Norwalk, Connecticut) with glass and diamond knives, respectively. Thin sections were examined in an AEI type 801 electron microscope (AEI Scientific Apparatus Division, Harlow, Essex, England).

RESULTS Control Loops Light and electron microscopic evaluation of each control sample revealed a smooth surface with long crypts (Figure 1). The extent of the mucosa was covered by a simple columnar epithelium containing numerous goblet cells interspersed between surface and crypt absorptive cells. Each surface absorptive cell displayed an apical microvillus border, a basal nucleus, and a s u p r a n u c l e a r cytoplasm containing mitochondria, free and membrane-bound ribosomes, Digestive Diseases, Voi. 20, No. 8 (August 1975)


Fig 3. Light micrograph of an unstained section of colonic mucosa from a peroxidase-exposed loop showing adherence of peroxidase reaction product to surface epithelial cells and its presence in intercellular spaces of the epithelium and lamina propria (arrows). Erythrocyte (E). (• 220)

lysosome-like bodies, and saccules and vesicles of the Golgi complex. Histochemically processed tissue samples from loop preparations not exposed to intraluminal peroxidase demonstrated endogenous enzyme activity in red and white blood cells, and occasionally around goblet cells in the epithelium (Figure 2). Ultrastructural observations of surface absorptive cells demonstrated occasional endogenous enzyme activity in lysosome-like bodies in the perinuclear area, but no such reactivity was apparent in apical vesicles or in other intracellular locations. Neither exposure to horseradish peroxidase nor incubation in diaminobenzidine and/or hydrogen peroxide media produced noticeable alterations in histochemical, histological, or ultrastructural characteristics of the tissue. E x p e r i m e n t a l Loops

Histological examination of peroxidase-exposed colonic tissue revealed evidence of perDigestive Diseases, Vol. 20, No. 8 (August 1975)

oxidase absorption by surface absorptive cells (Figure 3). While some areas of the mucosa demonstrated only limited uptake of intraluminal tracer particles, other areas revealed substantial penetration by peroxidase material. The degree of mucosal penetration did not appear to relate to the degree of integrity of the colonic tissue because all specimens appeared intact and normal by established standards (2, 4-6). Ultrastructural evaluation of peroxidase-exposed colonic samples demonstrated the presence of peroxidase reaction product in the epithelial and connective tissue compartments of the mucosa as well as in lymphatic capillaries of the submucosa. Peroxidase reaction product was observed adherent to the mucosal border and in apical pinocytotic vesicles of surface absorptive cells (Figure 4). Tracer-containing vesicles were also found in close apposition to lateral and basal intercellular membranes (Figure 5) and occasionally were seen fusing with these 753




(1.~ o

g~E "~







O'I O I ~ .E~~


~ ~ r- ~ , ' 1 O ~ ., ~






~._= ~ g~ ea

~_o= ~


~ ~~ ex~


Digestive Diseases, Vol. 20, No. 8 (August 1975)






~E ~:E O


- -






X i


~'~ ~--o

~8 E~ 0





._~ ~ E'~. e-






~ 5 d


Digestive Diseases, Vol. 20, No. 8 (August 1975)




O "O


Q- 0

~ 0

Q ~






"~E O-r-




o.~. e-

.9 ~

E _e- ~"~ r ~'~=


Digestive Diseases, Vol. 20, No. 8 (August 1975)


cO o

O) tO ru) Q. O

_o ._o o 0 0 "O 0 X




E~ 0







._~ r~ E~

.-~ s

Digestive Diseases, Vol. 20, No. 8 (August 1975)



ogenous source. The distribution of exogenous tracer material in a given specimen related directly to the duration of in vivo loop maintenance. Tissue from short-term loops (5-10 min) displayed substantially more macromolecular material in the lumen and in apical vesicles than elsewhere in the mucosa or submucosa; tissue from long-term loops (20-30 min) demonstrated greater penetration of reaction product into deeper regions of the colonic surface.


Fig 8, Electron micrograph of an unstained section from a horseradish peroxidase-exposed colonic loop showing adherence of reaction product to the microvitlus border (large arrow), but its lack of penetration into the apical junctional complex (JC). (• 70,000)

membranes, permitting direct communication between the intracellular and extracellular compartments of the epithelium (Figure 5). Intercellular spaces of the mucosa contained peroxidase reaction product (Figure 6) and the same material was occasionally apparent in lymphatic vessels of the submucosa (Figure 7) but on no occasion was the tracer apparent within apical junctional complexes of the epithelium (Figure 8). Endogenous peroxidase activity was observed throughout the cytoplasm of red blood cells and in intracytoplasmic digestive bodies of cells from the white blood cell series (Figure 9). It was not established whether or not any of the peroxidase reaction product in the white blood cells was derived from the ex758

It is generally believed that the normal adult intestine maintains a protective barrier against absorption of macromolecules. Recent investigations, however, have clearly shown that intact proteins (29-31, 34-36, 38) and other large particles (32, 33, 39-41) can penetrate the epithelial lining of the small bowel. The degree of the macromolecular penetration appears to be small, and adverse changes in the health of the host appear to be minimal. It is possible, nevertheless, that such a phenomenon may contribute to the pathophysiology of various disease states, especially those involving allergic reactions to food proteins, or toxic reactions to bacterial and parasitic products. The capacity of the large intestine to absorb macromolecules has not been examined by morphological methods. Since the luminal contents in this region of the gut contain substantial amounts of microbial toxins, it seems important to determine if such compounds can enter the adjacent mucosal tissue. The results of this study clearly demonstrate that surface epithelial cells of the colonic mucosa absorb horseradish peroxidase via pinocytosis. In addition, the absorbed protein is conveyed within vesicles to lateral and basal intercellular membranes where membrane fusion and exocytosis occurs. From the extracellular spaces of the epithelium, the macromolecular material crosses the basal lamina and enters intercellular spaces of the lamina Digestive Diseases, Vol. 20, No. 8 (August 1975)


Fig 9. Electron micrograph of an unstained section from a horseradish peroxidase-exposed colonic loop showing endogenous peroxidase reaction product in the cytoplasm of an erythrocyte (E) and in phagolysosome-like bodies (arrows) in the cytoplasm of a neutrophil (N). (X 20,000)

propria. Eventually it is identified in lymphatic channels of the submucosa and occasionally it is seen in blood capillaries of the subepithelial compartments. The colonic absorptive cell, therefore, is similar to absorptive cells of the small intestine in its capacity to absorb macromolecules by pinocytosis and transport them to subepithelial vessels in the mucosa and submucosa. This process very likely is active in the movement of large microbial byproducts from the intestinal lumen into the circulatory system of the host. On occasion, substantial peroxidase reaction product was observed between epithelial cells, which themselves displayed rather limited numbers of peroxidase-containing vesicles. Since apical junctional complexes were intact and diDigestive Diseases, Vol. 20, No. 8 (August 1975)

rect intercellular migration of peroxidase thus prevented, it is possible that the accumulation of intercellular tracer might be explained by a relatively rapid intracellular vesicular transport process with subsequent retardation of downward intercellular flow. Retarded migration of macromolecular particles between epithelial cells might be explained by a partial inhibition of macromolecular passage through the basal lamina (42); alternatively, it is possible that the dynamics of fluid movement into and from the colonic mucosa are periodically conducive to t e m p o r a r y stagnation. W h a t e v e r the explanation might be, morphological data strongly suggest that occasional retention of absorbed material between epithelial cells does not promote obvious changes in mucosal archi759


% D 8

f3MC 84184 ,,,,



Fig 10. Diagrammatic representation of a surface absorptive cell from rat colon showing progress o4 exogenous horseradish peroxidase tracer material (HP) from the intestinal lumen into apicaJ vesicles (V), subsequent movement into intercellular spaces (ICS) via vesicular exocytosis, and eventual presence in blood vessels (BV) and lymphatic channels (L) of underlying connective tissue compartments. RER, rough endoplasmic reticulum, Ly, lysosome-like body, N, nucleus, M, mitochondrium, BL, basal lamina, and D, desmosome.


Digestive Diseases, Vol. 20, No. B (August 1975)


tecture, but probably represents a normal phenomenon characteristic of mucosae with active absorptive and secretory functions. Colonic tissue from animals not exposed to intraluminal peroxidase demonstrated significant amounts of endogenous peroxidase activity only in erythrocytes, eosinophils, macrophages and polymorphonuclear leukocytes. Very little enzyme activity was found within absorptive cells in the surface or in the crypts and a limited amount was apparent in or around mucus-secreting cells in either region. These findings are in contradiction to those of Venkatachalam et al (43) who found prominent brown staining of epithelial cells in the lower portions of the crypts. By electron microscopy they observed that these crypt cells contained an electron dense reaction product associated with free ribosomes and in cisternae of the endoplasmic reticulum. Since some Golgi saccules and vesicles also displayed a prominent stain, these authors suggest that the crypt cells synthesize peroxidase in the endoplasmic reticulum and subsequently package and secrete it via the Golgi apparatus. They further recommend that once within the intestinal lumen, the peroxidase participates in a bactericidal mechanism which protects the colonic mucosa from invasion by microbial inhabitants. No explanations can be offered as yet for the lack of agreement between the present observations and those of Venkatachalam et al (43). The methods of tissue preparation were not markedly different in the two reports and the control experiments recorded by Venkatachalain et al (43) strongly suggest that the reaction they observed was enzymatic in nature and that the enzyme was probably peroxidase. One is left, then, with the possibility that the region of the colon examined by Venkatachalem et al (43) was different from that studied in this experiment; alternatively, the difference in strain and/or age of the laboratory rats examined might also be responsible for the contradictory results. The strain and age factors have been previously demonstrated to influence Digestive Diseases, Vol. 20, No. 8 (August 1975)

intestinal enzyme development and maturation (44). The observations reported in this study have shown that colonic epithelial cells of adult Sprague-Dawley rats can absorb intact protein from the intestinal lumen. The mechanism involves a pinocytotic process followed by vesicular transport and exocytosis at lateral and basal cell membranes (Figure 10). In addition, endogenous peroxidase production by surface and crypt cells is minimal in adult rats of the Sprague-Dawley strain. Consequently, this enzyme is likely insignificant as an intraluminal bactericidal agent under the circumstances here described. The significance of these findings is presently unknown. It is possible, however, that the observed transport phenomenon might contribute to the establishment of pathological changes in the colonic mucosa or to the manifestation of systemic disorders associated with circulating toxic and/or antigenic compounds. ACKNOWLEDGMENTS

The authors wish to thank Dr. Edwin S. Boatman, Ms. Janet GouLd, and Ms. DeLoris Highsmith for their assistance in the preparation of this manuscript. REFERENCES

1. Davenport HW: Physiology of the Digestive Tract. (3rd ed). New York, Yearbook Medical Publishers, Inc., 1971, pp. 211-217 2. Donnellan WL: The structure of the colonic mueosa. The epithelium and subepithelial reticulohistiocytic complex. Gastroenterology 29:496514, 1965 3. Bloom W, Fawcett D: A Textbook of Histology, (9th ed). Philadelphia, WB Saunders, 1968, pp. 568-571 4. Kaye GI, Lane N, Pascal RR: Colonic pericryptal fibroblast sheath: replication, migration and cytodifferentiation of a mesenchymal cell system in adult tissue. Gastroenterology 54:852865, 1968 5. Pascal RR, Kaye GI, Lane N: Colonic pericryptal fibroblast sheath: replication, migration and cytodifferentiation of a mesenchymal cell system in adult tissue. I. Autoradiographic stud761










14_ 15.






ies on normal rabbit colon. Gastroenterology 54:835-851, 1968 Pittman FE, Pittman JC: An electron microscopic study of the epithelium of normal human sigmoid colonic mucosa. Gut 7:644-661, 1968 Hampton JC: An electron microscopicstudy of mouse colon. Dis Colon Rectum 3:423-440, 1960 Takeuchi A, Zeller J: Scanning electron microscopic observations on the surface of the normal and spirochete-infested colonic mucosa of the rhesus monkey. J Ultrastruct Res 40:313-324, 1972 Shorter RG, Cardoza M, Huizenza KA, ReMine SG, Spencer R J: Further studies of in vitro cytotoxicity in lymphocytes for colonic epithelial cells. Gastroenterology 57:30-35, 1969 Mekhjian HS, Phillips SF: Perfusion of the canine colon with unconjugated bile acids. Effect on water and electrolyte transport, morphology and bile acid absorption. Gastroenterology 59:120-129, 1970 Hodgson J: Effect of methylcellulose on rectal and colonic pressures in treatment of diverticular disease. Br Med J 2:729-731, 1972 Hill M J, Drasar BS, Aries V, Crowther JS, Hawksworth G, Williams, REO: Bacteria and aetiology of cancer of large bowel. Lancet 1:95100, 1971 Gold P: Circulating antibodies against carcinoembryonic antigens of the human digestive system. Cancer 20:1663-1667, 1967 Burkitt DP: Epidemiology of cancer of the colon and rectum. Cancer 28:3-13, 1971 Aries V, Crowther JS, Drasar BS, Hill M J, Williams REO: Bacteria and the etiology of cancer of the large bowel. Gut 10:334-335 Berg JW, Howell MA, Silverman SJ: Dietary hypotheses and diet-related research in the etiology of colon cancer. Health Serv Rep 88:195924, 1973 Devroede G J, Phillips SF: Conservation of sodium, chloride, and water by the human colon. Gastroenterology 56:101-109, 1969 Curran PF, Schwartz GF: Na, C1, and water transport by rat colon. J Gen Physiol 43:555571, 1960 Cooperstein IL, Brockman SK: The electrical potential difference generated by the large intestine: its relation to electrolyte and water transfer. J Clin Invest 38:435-442, 1959

20. D'Agostino AW, Leadbetter WF, Schwartz WB: Alterations in the ionic composition of isotonic saline solutions instilled into the colon. J Clin Invest 32:444-448, 1953 21. Ammon HV, Phillips SF: Inhibition of colonic water and electrolyte absorption by fatty acids in man. Gastroenterology 63:744-749, 1973 22. Valdivieso VD, Schwabe AD: Absorption of medium-chained lipids from the rat cecum. Am J Dig Dis 11:474-479, 1966 23. Pihl EG, Glotzer D J, Patterson JF: Absorption of medium-chain fatty acids from the dog colon. J Appl Physiol 21:1059-1062, 1966 24. Levitan R, Patterson JF: Sodium hexonoate absorption from the human small bowel and colon. Gastroenterology 50:856, 1966 25. Ham AW: Histology (Sth ed). Philadelphia, J.B. Lippincott, 1966, pp. 697-800 26. Finkelstein JD: Malabsorption. Med Clin. North Am 52:1339-1354, 1968 27. Floch MH: Recent contributions in intestinal absorption and malabsorption. A review. Am J Clin Nutr. 22:327-351, 1969 28. Dack GM, Hoskins D: Absorption of botutinum toxin from the colon of Macaca mulatta. J Infect Dis 71:260-264, 1942 29. Walker WA, Cornell R, Davenport LM, Isselbacher K J: Macromolecular absorption. Mechanism of horseradish peroxidase uptake and transport in adult and neonatal rat intestine. J Cell Biol 54:195-205, 1972a 30. Walker WA, Isselbacher K J, Bloch K J: Intestinal uptake of macromolecules: Effect of oral immunization. Science 177:608-610, 1972b 31. Warsaw AL, Walker WA, Cornell R, Isselbacher K J: Small intestinal permeability to maeromolecules. Transmission of horseradish peroxidase into mesenteric lymph and portal blood. Lab Invest 25:675-684, 1971 32. Worthington B, Graney DO: Uptake of adenovirus by intestinal absorptive cells of the suckling rat. I. The neonatal ileum. Anat Ree 175:37-62, 1973a 33. Worthington B, Graney DO: Uptake of adenovirus by intestinal absorptive cells of the suckling rat. II. The neonatal jejunum. Anat Rec 175:63-76, 1973b 34. Worthington BS, Boatman ES, Kenny GE: Intestinal absorption of intact proteins in normal and protein-deficient rats. Am J Clin Nutr 27:276-286, 1974 Digestive Diseases, Vol. 20, No. 8 (August 1975)


35. Worthington BS, and Boatman ES: The influence of protein malnutrition of ileal permeability to macromolecules in the rat. Am J Dig Dis 19:43-55, 1974 36. Cornell R, Walker WA, Isselbacher K J: Small intestinal absorption of horseradish peroxidase. A cytochemicat study. Lab Invest 25:42-48, 1972 37. Luft J: Improvements in epoxy resin embedding methods. J Biophys Biochem Cytol 9:409-414, 1961 38. Ravin HA, Rowley D, Jenkins C, Fine J: On the absorption of bacterial endotoxin from the gastrointestinal tract of the normal and shocked animal. J Exp Med 112:783-792, 1960 39. Sanders E, Ashworth CT: A study of particulate intestinal absorption and hepatocellular uptake. Use of polystyrene latex particles. Exp. Cell Res 22:137-145, 1961

Digestive Diseases, Voi. 20, No. 8 (August 1975)

40. Schatten WE: The role of intestinal bacteria in liver necrosis following experimental excision of the hepatic arterial supply. Surgery 36:256-269, 1954 41. Schatten WE, Des Pres JD, Holden WD: A bacteriologic study of portal vein blood in man. Arch Surg 71:404-407, 1955 42. Graham RC, Karnovsky M J: Glomerular permeability. Ultrastructural cytochemical studies using peroxidases as protein tracers. J. Exp. Med. 124:1123-1133, 1966 43. Venkatachalam, MA, Soltani MH, Fahimi HD: Fine structural localization of peroxidase activity in the epithelium of large intestine of rat. J Cell Biol 46:168-173, 1970 44. Koldovsky OP, Hahn OP, Faltova E, Vacek Z: The functional development of the intestinal tract in rats. Development of Homeostasis P. Hahn (ed). New York, Academic Press, 1963, pp. 155-165


Absorption of intact protein by colonic epithelial cells of the rat.

Colonic absorption of intact protein was examined in adult rats using histological and ultrastructural procedures. Horseradish peroxidase was introduc...
3MB Sizes 0 Downloads 0 Views