238 J. Physiol. (I939) 97, 238-249 6I2.397.2I:6I2.63 FATTY INFILTRATION OF THE LIVER IN PREGNANT EWES By L. C. SNOOK From the Rowett Research Institute, Aberdeen (Received 14 August 1939) THROUGHOUT the world sheep breeders suffer losses among their breeding ewes from a disease variously known as "pregnancy disease", "pregnancy toxemia", and "twin lamb disease". The two most characteristic clinical features of this disease are a marked ketosis and extreme fatty infiltration of the liver. In previous papers [Fraser, Godden, Snook & Thomson, 1938, 1939] it was shown that ketosis can be produced in pregnant ewes by undernutrition, and the possible relation between the experimental condition and pregnancy disease was discussed. The ketosis was associated with marked fatty infiltration of the liver and opportunities were taken to make quantitative and qualitative studies of the fat present in the livers of normal and abnormal experimental animals, and also in livers from field cases of pregnancy disease. It is sometimes stated that a certain degree of fattiness of the liver is a normal physiological associate of pregnancy. Thus Gaiger & Davies [1932] list among the various causes of fatty infiltration-" Physiological -during pregnancy and lactation". Little definite evidence on this point is available, however. The work of Coope & Mottram [1914] has been quoted as showing that fatty infiltration of the liver is a normal concomitant of pregnancy but the increases shown (in cats and rabbits) are relatively small and, in view of recent work, of doubtful significance. Thus they found 3-4 and 5-6 % total fatty acid in the livers of two pregnant rabbits and an average of 2-6 % in controls (7) of mixed sexes. Increases of this order do not warrant use of the term " fatty infiltration ". Best & Ridout [1933] mention the assumption that pregnant rabbits near term tend to have fatty livers, but note that a colleague was unable to confirm this, even when the does were kept on a diet high in fat.
FATTY LIVER IN PREGNANCY 239 No work has been published concerning the chemical nature of the excess fat found in the livers of ewes afflicted with pregnancy disease. Indeed, for reasons mentioned above, it has been suggested that fattiness of the liver might be characteristic of healthy pregnant ewes near term. In the current investigation the opportunity was taken to estimate the amount and nature of the fat present in the livers of normal sheep (wethers, barren ewes, healthy pregnant ewes), apparently healthy but ketonaemic ewes, comatose, ketonaemic ewes, and field cases of pregnancy disease. In all cases concurrent analyses were made on the blood, and for the experimental animals data were available concerning food consumption, bodyweight changes and the composition of the blood over a considerable period prior to slaughter. This supplementary information is of prime importance in interpreting results. Details concerning the feeding and management of the experimental ewes have been reported elsewhere [Fraser et al. 1938]. METHODS The liver was removed from the animal as soon as possible after slaughter or death. Small cubes of tissue for subsequent histological examination were cut from the various lobes and placed in 10 % formaldehyde. The fresh liver was then minced and thoroughly mixed. DETERMINATIONS Moisture. Representative samples of the fresh pulped liver were dried to constant weight at 1000 C. in an electric oven. Fatty substances. 10-14 g. samples of the fresh pulped liver were accurately weighed and then ground in a mortar with anhydrous sodium sulphate. The dry powder so obtained was extracted with chloroform for 4 hr. in a Soxhlet apparatus, the material was then re-ground and extracted for a further 4 hr. The filtered chloroform extract was then made up to 100 c.c. in a measuring flask, aliquots being used, as needed, for the following estimations: (a) Total fat. The chloroform was removed by distillation and the residue dried to constant weight at 1000 C. (b) Total phosphorus. Five c.c. of the chloroform extract were placed in a small Kjeldahl flask, the chloroform was removed by gentle heating and the residue was then digested with 5 c.c. of a mixture of equal parts of concentrated sulphuric and nitric acids until the liquid was colourless and white fumes were given off. On cooling, water was added to the
240 L. C. SNOOK residue, which was then neutralized, and in it phosphorus was estimated by the method of Fiske & Subbarow [1925]. (c) Unsaponifiable residue. Suitable aliquots were taken and the chloroform removed. 25 c.c. of N/2 alcoholic potash were then added to the residue and the mixture boiled until saponification was complete. Water was added and the unsaponifiable residue removed by 6 extractions with ether. The combined ethereal extracts were washed with water until neutral, the ether was removed and the unsaponifiable residue determined by weighing to constant weight after drying in vacuo at 1000 C. (d) Iodine number. Dam's pyridine-sulphate-bromine method was used [Plimmer, 1938]. (e) Glycogen. The glycogen content of the fresh liver was not determined, but as soon as possible after slaughter of the animal small samples of liver tissue were placed in absolute alcohol and preserved until it was convenient to complete the analyses. The liver slices were then drained on blotting paper, and weighed by difference in tubes containing 60 % KOH. Subsequent procedure was similar to that described by Evans, Tsai & Young [1931] but 80 % alcohol was used for precipitation, the sugar in the ultimate hydrolysate being estimated by the Somogyi modification of the Shaffer-Hartmann method [Peters & Van Slyke, 1931]. RESULTS To serve as a guide to the possible effect of pregnancy on the fat content of the liver, a number of small laboratory animals on stock diets were slaughtered just before parturition, the livers being removed and analysed. None of the livers from these small, healthy animals contained unusual amounts of fat, the maximum being 21-6 % total extract in the dry matter of one of the rat livers. The average values are given in Table I. From these data it does not appear that there is any fatty infiltration of the liver near term in the rat, rabbit or guinea-pig. TABLE I. The fatty substances in the livers of pregnant rats, rabbits and guinea-pigs Total fat Number of,a foetuses Moisture Wet liver Dry liver Average: 8 rats 9 72 5 0 17-6 3 rabbits 7 72 4.7 16-7 Guinea-pig A 3 72 2-8 10.0 B 3 71 2-6 9-0 (For comparison: Best, Channon & Ridout [1934] found 5.4 % total extract in the fresh livers of grain-fed rats used as controls.) Samples of liver were obtained from healthy wether hoggets (castrated male sheep) being slaughtered for human consumption. These were
FAT'TY LIVER IN PREGNANCY 241 analysed to give an index of the lipids present in the livers of healthy sheep undisturbed by any sexual activities. It can be seen from Table II that the livers of healthy wether hoggets contain essentially the same percentage of moisture and total chloroform extract as those of pregnant rats and rabbits. No fat could be detected by histological examination following staining of the tissue with Scharlach R. TABLE II. Average: 2 healthy wethers Healthy barren ewe; basal diet (No. 55, Group III, 1939) Healthy fat pregnant ewe killed near term (No. 30, Group I, 1939), 3 lambs Average: 11 non-ketonaemic ewes at term (1-3 lambs) "Healthy" ketonaemic ewe (No. 23, Group IV, 1939), 3 lambs Ketonaemic ewe killed immediately after lambing (No. 4, Group V, 1938), 2 lambs Average: 5 ketonaemic ewes killed when comatose (2-3 lambs) Average: 5 field cases pregnancy disease (2-3 lambs) The moisture, fat and glycogen content of sheep livers Moisture 73 70 74 75 64 57 Wet liver Total fat Dry matter 4-8 18 5*6 19 Glycogen in liver preserved in alcohol 6-0 5-8 22 3.3 5.5 (range 5-7) 16-6 31 60 25 (range 20-32) 66 17 (range 15-20) 22-43 1*2 71 62 Negligible An experimental barren ewe (No. 55, Group III, 1939) was slaughtered for carcass examination after being for 3 months on a basal ration which produced ketosis in pregnant ewes [see Fraser et al. 1939]; the animal had also been fasted for 2 days, a week prior to slaughter. It might have been expected that the fasting and continued low plane of nutrition would have produced fattiness of the liver but it can be seen from Table II that the analytical figures are comparable with those of the wethers. Here again histological examination indicated that the slight fat content was distributed uniformly throughout all the cells. A fat, experimental, pregnant ewe was slaughtered for carcass examination 1 day before the due date of lambing. This animal (No. 30, Group I, 1939) had been well fed throughout gestation and had gained 37 lb. in bodyweight. She appeared to be in the best of health up to the time of slaughter and at autopsy all organs appeared normal. Three 51
242 L. C. SNOOK healthy lambs (total weight 19 lb.) were carried. As can be seen from Table II there was no significant increase in the fat content of the liver. It was interesting to find on histological examination, however, that there was a suggestion of accumulation of fat in the cells adjacent to the portal vessels. As will be described later, it has been found that fatty infiltration of the liver in the ewe appears to progress from the periphery of the lobule towards the centre, and there was therefore some indication here of incipient fatty infiltration. This finding links up with the changes noted in the composition of the blood. Four days prior to slaughter blood drawn from this ewe contained 48 mg. sugar/100 c.c. and no ketone bodies were detected. On the day of slaughter, however, the blood sugar had dropped to 37 mg. and 9 mg. total ketone/100 c.c. blood were present. These blood figures are of interest in connexion with the glycogen content of the liver. It will be noted that the liver of the barren ewe contained 6-0 % glycogen in contrast to the 3-3 % found in the liver of the pregnant ewe receiving a much better ration (daily intakes, 065 and 1-69 lb. starch equivalent respectively). When stained with Best's carmine, sections from the liver of the barren ewe appeared deep red in colour, all cells containing glycogen. In contrast, the staining showed that in the liver of the multiply-pregnant ewe the glycogen was concentrated in the cells around the portal vessels. It seems that the strain of multiple pregnancy tends to exhaust the glycogen reserves of the liver and this may predispose to fatty infiltration near term. However, this ewe was only 1 day antepartum when killed and the analytical data do not indicate fattiness of the liver. During the main investigation, it was possible to obtain liver samples from a number of non-ketonaemic ewes which, for various reasons, had to be slaughtered at, or soon after, parturition. Some of the experimental animals, for example, were excessively fat and had trouble at lambing, being slaughtered for such reasons as excessive haemorrhage or a torn uterus. These ewes cannot be classed as strictly normal, but blood analyses had shown that they were not ketonaemic and it therefore seems that one is justified in considering them as controls for comparison with ketonaemic animals. It will be seen from the table that none of these animals had fatty livers and the data lend support to the suggestion that fatty infiltration of the liver is not a normal concomitant of pregnancy. Histological examination showed that the fat in the livers of nonketonaemic pregnant ewes was evenly distributed throughout the cells. Ketosis results both when ewes are kept on a quantitatively inadequate ration throughout gestation and when overfat, pregnant ewes are
THE JOURNAL OF PHYSIOLOGY, VOL. 97, No. 2 PLATE I Toface,. 242
FATTY LIVER IN PREGNANCY 243 suddenly placed on reduced rations. This ketosis, which may become very pronounced, is preceded, or accompanied, by fatty infiltration of the liver. Thus an experimental ewe (No. 23, Group III, 1939) receiving a quantitatively inadequate diet was slaughtered 14 days prior to due date of parturition. This animal had been markedly hypoglycaemic and ketonaemic for about a month, and on the day of death the blood contained 48 mg. total ketone/100 c.c. As can be seen from the table the liver contained excessive fat (16.6 % in the wet liver) and only 1-2 % glycogen. Microphotograph No. 1 shows the distribution of the fat in the liver of this apparently healthy ewe. It will be seen that the fat is accumulated in the cells around the periphery of the lobules, the tissue around the central vein being normal. This distribution is characteristic of that seen in the livers of all ketonaemic ewes, both experimental and field cases. Many multiply-pregnant ketonaemic, experimental ewes went to term, and although the lambs were very poor, the ewes did not necessarily display any untoward symptoms. This is illustrated by ewe No. 4. A fortnight prior to lambing the blood from this animal contained 24 mg. sugar and 64 mg. total ketone/100 c.c.; the day before lambing the ketone bodies had increased to 83 mg./100 c.c. An uneventful lambing took place and immediately afterwards the beast was slaughtered. The carcass was emaciated but the liver was the only abnormal organ; this was very pale and extremely fatty (71 % total extract in the dry matter). Some idea of the degree of infiltration may be obtained from microphotograph No. 2 which shows that only a few cells around the central vein were normal. As the process of infiltration had probably been proceeding over a period of many weeks, it would appear that this accumulation of fat does not have any serious effect on liver function. A number of ketonaemic experimental ewes lost their appetite and with decreased food intake serious clinical symptoms became apparent. Anorexia was followed by hyper-ketonaemia, lethargy, partial blindness, twitching of the ears, grinding of the teeth, physical weakness and coma. The affected animals remained comatose for several days and most of them were slaughtered when apparently dying. Several actually lambed but did not recover postpartum and died within 2 days. Autopsy revealed emaciated carcasses and marked fatty infiltration of the liver. In Table II are given the average figures for 5 such ewes which were slaughtered antepartum when comatose. In microphotograph No. 3 a section is shown which was prepared from a liver containing 68 % fat in the dry matter. A better index of the degree of the fatty infiltration in
244 L. C. SNOOK the cells of this liver is obtained from the high-power microphotograph No. 4, from which it can be seen that in many of the cells the protoplasm appears to have been almost entirely replaced by fat, the nucleus being pushed right up against the cell wall. This last illustration should be compared with microphotograph No. 5, which shows a section prepared from the liver of a healthy, non-ketonaemic ewe slaughtered at lambing because of a torn uterus. TABLE III. Weight of total fatty substances present in livers of normal and ketonaemic ewes Weight of Weight total fat Number of liver Total fat in the liver of No. lb. % lb. lambs Normal ewes 55. Barren 1-5 5-6 0-08 30. Healthy, fat, pregnant 2-2 5-8 0-13 3 13.,,,,,, 2*4 5-6 013 1 77.,,,,,, 2-5 7*2 0*18 3 Ketonaemic ewes 23. Killed for carcass 19 16-6 0*32 3 35. Killed when dying 2-6 24-0 0-62 3 92.,,,,,, 2*8 26-7 0.75 2 328.,,,,,, 2-6 25-3 0-66 2 58. Died after lambing 2-9 21-1 0-61 2 52. Died after prolonged anorexia 3-2 32-1 1-02 3 Not only does the percentage of fat increase in the liver of an ewe which becomes ketonaemic but the liver tends to be enlarged. A good index of the amount of fat which accumulates in the liver is obtained from Table III. Here the total weight of fat present in the various livers is given. Blood and liver samples were obtained, whenever possible, from field cases of pregnancy disease. The animals were always markedly ketonaemic, and chemical and histological examination of the fatty livers indicated that the changes produced in the field were similar to those found in experimentally-produced ketosis. Complete loss of appetite was also a characteristic feature of those field cases available for treatment. It is interesting to note that the hypoglycaemia and ketonaemia in field cases of pregnancy disease could be corrected by glucose therapy (oral or intravenous), but the animals would not resume eating and eventually died. The livers were still markedly fatty, even where ketonaemia had been eliminated for several days by glucose treatment.
FATTY LIVER IN PREGNANCY 245 POSSIBLE NECROTIC CHANGES In all cases (experimental and field) where ewes died after displaying symptoms characteristic of pregnancy disease, a careful search for possible necrotic changes in the liver was made, using paraffin sections cut at 7,u. Prof. J. S. Young of the Department of Pathology, Aberdeen University, kindly examined a number of representative sections and reported that in no case did it appear that the fatty infiltration was accompanied by degenerative changes in the cell structure. This confirms the work of M'Fadyean [1924] who, as the result of histological examination, stated that the liver changes seen in pregnancy disease in ewes, resulted from extreme physiological infiltration and not a toxic degeneration. CHEMICAL CONSTITUTION OF THE LIVER LIPIDS Various chemical analyses were carried out to test the hypothesis that the fat in the livers of ketonaemic ewes had been transported thereto from the body depots. Phospholipins are considered to play a role in the transport of fat in the body and phosphorus determinations were therefore made on the chloroform extracts in view of the possibility that the fatty infiltration may have been due to an upset in phospholipin metabolism. The results are given in Table IV. TABLE IV. The phosphorus and unsaponifiable extract present in the chloroform extract of sheep livers P as a percentage of r A^ 5 Unsaponifiable Total fat in Non-fat residue in dry matter Total fat dry matter total fat Normal livers Healthy wether 17 2-1 0-43 8-5 18 1-9 0-43 8-5 Ewe 93 21 7-8 85 25 2-0 0-66 8-6 Fatty livers Ewe 35 62 0-42 0-68 3.9 7 56 0-47 0-60 4-3 3 68 0-30 0-65 4 71 0-32 0-71 3-8 81 61 0-44 0-69 4-0 It was not considered likely that the infiltration in the ewe was due to unbalanced sterol metabolism but, to obtain definite evidence on this point, the amount of unsaponifiable residue in various chloroform extracts was determined. The results are included in Table IV. It will be seen that the extract from the fatty livers contains much less phosphorus per unit than the extract from normal livers. This is PH. XCVII. 16
246 L. C. SNOOK what one would expect if the excess fat was normal body fat transported from the depots. A better guide to abnormality is obtained by expressing the phosphorus as a fraction of the non-fatty dry matter of the liver. On this basis of comparison, the phospholipins in the fat-infiltrated livers are of the same order as that found in the liver of the normal pregnant ewe No. 85. Best, Channon & Ridout [1934] found 8 % unsaponifiable extract in the liver fats of normal " control " rats, which corresponds to that found in the liver fat of normal sheep. As one would expect following the infiltration of depot fat, the percentage of unsaponifiable residue in the liver fats decreases as the fat content of the liver increases. It is apparent that the fatty liver produced in the experimental ewes, or found in cases of pregnancy disease, is not of the cholesterol ester type. As a further test regarding the nature of the fat present in the fatty livers, Iodine Numbers of the extracts were determined. No definite correlation was apparent between the I.N. of the extract and the percentage of fat in the liver, but there was an indication that with increased fatty infiltration the I.N. tends to fall until it approximates that of depot fat. Thus the extracts from 5 fatty livers had I.N.s ranging from 77 to 71 as compared with 93 for the extract from the liver of the healthy ewe (No. 30, Group I, 1939) carrying 3 lambs. DIsCUSSION It seems that the ketosis produced in pregnant ewes by undernutrition follows, or is associated with, the mobilization of depot fat and its accumulation in the liver. Such data as are available also indicate that this fatty infiltration is associated with depletion of the liver glycogen. The finding that the liver of the poorly-fed barren ewe contained about double (6 %) the glycogen found in that of the well-fed, pregnant ewe slaughtered 1 day antepartum supports the hypothesis that in the terminal stages of multiple pregnancy, foetal demands prevent the storage of the usual amounts of glycogen. Naturally, if the glycogen reserves are depleted, the animal is less able to withstand undernutrition, and if the food supply is restricted, a need to use depot fat will quickly arise. Thus the liver of the ketonaemic ewe (to all appearances healthy) which was slaughtered for carcass examination contained only 1-2 % glycogen but about twice the normal amount of fat. With continued undernutrition the ketonaemia becomes increasingly severe, and it can be assumed that this is associated with even greater depletion of the glycogen reserves and increased accumulation of fat in the liver.
FATTY LIVER IN PREGNANCY There is reason to doubt whether loss of liver function can be of any significance in precipitating the decline of ketonaemic animals, as various workers have shown that the vital functions can be carried out by only a small fraction of the liver. Thus Bollmann & Mann [1935] excised up to 80 % of the liver from experimental dogs and found that normal functions could be carried out even when regeneration was prevented. Similar retention of function has been seen in animals with cirrhosis sufficient to destroy a large portion of the liver. Bollmann & Mann have shown, however, that it is the detoxicating power of the liver which is most seriously affected when the liver becomes infiltrated with fat, also that livers lacking in glycogen are less able to cope with toxic agents. Bollmann [1938] discusses further work which suggests that the detoxicating power of the liver may be lost before other functions fail. It is therefore possible that lack of liver glycogen may be a factor in reducing ketonaemic ewes to a comatose condition, either directly because of the resultant weakness, or indirectly by a loss of detoxicating ability by the liver cells. There is fairly definite evidence that fatty infiltration in itself will not result in any harmful symptoms. Many severely ketonaemic ewes, which by inference had fatty livers of long standing, went to term and the blood returned to normal within several days postpartum; ewe No. 4, for example, appeared healthy when slaughtered, despite gross fattiness of the liver. Such data as is available does not suggest that fatty infiltration of the liver occurs in normal, well-fed ewes. Many ewes in the field, however, do not receive an optimum diet and it is probable that ketosis and some degree of fatty infiltration of the liver occurs under these conditions. That such ketosis may be pronounced, though unsuspected by the farmer, is indicated by a subsidiary investigation. Through the courtesy of Sir Joseph Barcroft, samples of blood and liver were obtained from experimental ewes being used at Cambridge. These ewes, presumably in apparent good health, were slaughtered at various stages of gestation, and it was found that after about the 100th day blood from some of the ewes contained ketone bodies, as much as 22 mg./100 c.c. being found. Histological examination of the respective liver samples showed that quite marked fatty infiltration had taken place. Blood samples from other ewes in the same flock (one slaughtered as late as the 146th day of gestation) were quite free of ketone bodies and the livers appeared normal histologically. Variable food consumption by ewes being handfed in groups could explain the presence of a certain number of ketonaemic animals in a flock of ewes all carrying the same number of lambs. 16-2 247
248 L. C. SNOOK SUMMARY Chemical and histological studies have been made concerning the degree, nature, and distribution of the fat present in the livers of healthy, pregnant laboratory animals, normal ewes (barren and pregnant), ketonaemic ewes, and field cases of pregnancy disease. It is concluded that Fatty infiltration of the liver is not a normal concomitant of pregnancy in the rat, rabbit, guinea-pig, or sheep. The fatty infiltration of the liver associated with ketosis in the pregnant ewe does not necessarily provoke harmful symptoms; animals may go to term when the infiltration is extreme. In ketonaemic, pregnant ewes the glycogen content of the liver is apparently reduced to negligible amounts. Ketonaemic ewes in apparent good health were found to have livers infiltrated with fat, the nature and degree of the infiltration being indistinguishable from that seen in comatose ewes. The fatty liver produced in the experimental ewes could not be distinguished from that found in field cases of the disease. Fatty infiltration of the liver was not associated with necrotic changes in the liver tissue. I wish to thank Mr W. Godden for advice in the chemical work, and Dr J. T. Irving for guidance in the histological work and also for taking the microphotographs. Part of this study was carried out during the tenure of a Hackett Research Studentship from the University of Western Australia. REFERENCES Best, C. H., Channon, H. J. & Ridout, J. H. [1934]. J. Physiol. 81, 409. Best, C. H. & Ridout, J. H. [1933]. J. Physiol. 78, 415. Bollmann, J. L. [1938]. Ann. intern. Med. 12, 1. Bollmann, J. L. & Mann, F. C. [1935]. Ann. intern. Med. 9, 617. Coope, R. & Mottram, V. H. [1914]. J. Physiol. 49, 23. Evans, L. C., Tsai, C. & Young, F. G. [1931]. J. Physiol. 73, 67. Fiske, C. H. & Subbarow, Y. [1925]. J. biol. Chem. 66, 375. Fraser, A. H. H., Godden, W., Snook, L. C. & Thomson, W. [1938]. J. Physiol. 94, 346. Fraser, A. H. H., Godden, W., Snook, L. C. & Thomson, W. [1939]. J. Physiol. 97, 120. Gaiger, S. H. & Davies, G. 0. [1932]. Veterinary Pathology and Bacteriology. London: Bailliere, Tindall & Cox. M'Fadyean, J. [1924]. J. Comp. Path. Therap. 37, 287. Peters, J. P. & Van Slyke, D. D. [1931]. Quantitative Clinical Chemistry, II. London: Bailli'ere, Tindall & Cox. Plimmer, R. H. A. [1938]. Organic and Biochemistry, 6th ed. London: Longmans Green & Co.
FATTY LIVER IN PREGNANCY 249 EXPLANATION OF PLATE I 1. x 1i. A frozen section stained with Scharlach R and haematoxylin to show distribution of the fat in the liver tissue of a "healthy" ketonaemic ewe (No. 23, Group IV, 1939) with 3 lambs in utero. The liver contained 46 % total fat in the dry matter. 2. x 30. A paraffin section stained with haematoxylin and eosin D showing the extreme fatty infiltration of the liver found in a ketonaemic ewe (No. 4, Group V, 1937-8) slaughtered immediately after lambing (2 lambs). The liver contained 71 % total fat in the dry matter. 3. x 30. A paraffin section stained with haematoxylin and eosin prepared from the liver of a hypoglycaemic, ketonaemic ewe (No. 3, Group VI, 1937-8) killed after 2 days' coma (2 lambs in utero). The liver contained 68 % total fat in the dry matter. 4. x 67. As No. 3 at higher magnification. 5. x 67. A paraffin section stained with haematoxylin and eosin prepared from the liver of a healthy experimental ewe slaughtered at lambing because of a torn uterus. The liver contained 25 % total fat in the dry matter.