Biol Cell (1995) 83, 163-I 67 0 Elsevier, Paris 163

Original article

Interferon @?-induced abnormalities in adipocytes of suckling mice

Andrea Sbarbati a, FranGoise Leclercq b, Francesco Osculati a, Ion Gresser c

‘Institute of Human Anatomy and Histology, University of Verona, Strada le Grazie 8, 37137 Verona Italy* bL.uboratoire de Chimie Organique-Biologique et Spectroscopique de 1’Unite’ mixte CNRS/RhSne-PoLlenc,’

CLuboratoire d’Oncologie Wale, UPR 274 CNRS, Institut de Recherches sur le Cancer, IFCI CNRS, WllejuiJ France

(Received 26 January 1995; accepted 5 April 1995)

Summary - The modification induced by interferon (lFN) in brown adipose tissue (BAT) was studied by high spatial resolution mag- netic resonance imaging (MRI), histology, and transmission electron microscopy (TEM). In IFN-treated mice, at MRI, the interscapu- lar BAT was slightly enlarged and showed non-homogeneous areas of lipid accumulation. The thickness of the subcutaneous white adi- pose tissue was reduced with respect to control mice. In the liver, MRI showed a lipid accumulation, In lFN-treated mice, by light microscopy, brown adipocytes showed a larger lipid deposit with respect to control mice. At TEM, in BAT, the mitochondria were reduced in number, smaller and the number of crlstae was also significantly reduced with respect to the controls (9.1 * 1.5 vs 20.1 f 1.9, P < 0.01). The inclusions in the mitochondrial matrix were significantly less numerous in m-treated than in control animals (1.9 + 0.7 vs 0.9 f 0.7 for mitochondrial section, P < 0.01). Abnormalities of endoplasmic reticulum described in hepatocytes were not found in brown adipocytes of IFN-treated mice. The present work demonstrates that, in the BAT of sucking mice, IFN-treatment induces morphologic alterations and that brown adipocytes have MRI and TEM features resembling those found in the lipid laden BAT of aged animals.

brown adipose tksue / interferon / magnetic resonance imaging / ultrastructure I mouse

Introduction

Daily administration of potent interferon &jI (IFN) prepara- tions to suckling mice during the first week of life resulted in a syndrome characterized by inhibition of growth, exten- sive liver cell degeneration, and death in the second week of life [4]. When IFN treatment was discontinued after the first week, the mice developed a rapidly progressive glomerulo- nephritis [3]. This IFN treatment resulted in a marked increase in triglycerides and a decrease in the level of some phospholipids in the liver [17]. These biochemical modifi- cations were accompanied by the appearance of abnormal tubular aggregates arising from the endoplasmic reticulum of hepatocytes [7,8]. We have suggested that IFN treatment results in an inhibition of one of the processes that leads to activation of the enzymatic systems responsible for the syn- thesis of species of polyunsaturated phosphatidylcholine and phosphatidylethanolamine in the liver of suckling mice [16]. In view of the marked changes in lipid metabolism induced by IFN in suckling mice, it was of interest to inves- tigate the effect of IFN on the brown adipose tissue (BAT) of suckling mice. BAT is characterized by multivacuolar lipid deposits, whereas the white adipose tissue (WAT) is typically monovacuolar. The prominent activity of BAT is non-shivering thermogenesis. This is effected by unique mitochondria in which oxidation of substrates can be uncoupled from phosphorylation of ADP to ATP [2]. Therefore, BAT mitochondria dissipate energy from the oxidation of fatty acid as heat. The physiological role of

BAT in the adult animal is not clear. One hypothesis is that BAT could be involved in the disposal of excess calories [6]. Deficiency or a reduced metabolic efficiency of BAT could play a role in the pathogenesis of obesity [ 111. In newborn mice and rats, BAT rapidly proceeds to morpho- logical and functional development reaching the peak of differentiation and heat producing capacity at l-2 weeks of age [l, 13 3. In the present article, the morphologic modifi- cations induced by lFN in BAT have been studied by an original approach correlating in vivo evaluations by high spatial resolution magnetic resonance imaging (MRI), his- tology and transmission electron microscopy (TEM).

Material and methods

Mice

Litters of Swiss mice were obtained from the pathogen free colony at the Institut de Recherches Scientitiques sur le Cancer, Villejuif, France.

IFN yreparation

Mouse IFN alp was prepared from suspension cultures of mouse sarcoma C243 cells inoculated with Newcastle disease virus. The methods of production, purification and assay have been previously described [14]. The specific activity of this purified IFN was ca 2 x 107 units/mg protein. Control preparations consisted of supematant from cultures of C243 cells prepared in

exactly be Same way as IFN except that the IFN inducer virus was omitted.

Experimental plan

half of the &e in each litter were marked ad injected subcu- taneously daily from birth to sixth day-with 0.05 ml of purified mouse 1~ having a titer of 3.2 x 106 or with the control preparation.

Magnetic resonance examination

MRI was p&om.ed on the sixth day using a magnetic resonance spectrometer (Bruker 300 MSL) operating at 7.05 T&a equipped with a mini-imaging system. The.imaging sequence employed was a spin who multi&e: the best. images were obtained with a rep&ion time of 0.8 s and an e&-time of 0.026 s and averaged over eight scans for 256 phase encoding steps. The fidd of view was 3 x 3 cm and the final voxei dimension was 120 X 120 X 3000 pm.

Ultrastructural examination

Five IFN-treated and five control mice were killed by dislocation of cervical vertebrae. The mice were autopsied, and the inter- scapular BAT was removed. Fragments of -BAT for TEM were fixed in 2.5% glutaraldehyde in Sorensen buffer for 2 h, post- fixed in 1% osmium tetroxide plus 1.5% sod&ii ferrocyanide in Sorensen buffer for 1 h, dehydrated in graded ethanols, embed- ded in Epon-araldite and cut with a Ultracut E ultramicrototie (Reichert). Semithin sections were stained with toiuidiine blue. Ultrathin sections were stained with lead citrate and uranyl ace- tate and observed under a EM10 electron microscope (Zeiss). Mnrphometric examination w&s performed in thr?e mice for each group, at 20000 x in 30 cells for each animal selected from three different tissue blocks according to a standardized randomized protocol. Results were expressed as mean it standard deviation. Statistical comparison of the data was performed with the Student’s t-test.

ReSUltS

Magnetic resonance examination

in control mice, MRI did not show alterations in any organ. In particular, the BAT displayed the low intensity signal (fig 1) typical of suckling rodents [9, 121. In contrast, the interscapular BAT of m-treated mice which was slightly enlarged showed non-homogeneous areas of lipid accumu- lation recognizable for an intensity of the sigql higher than in the controls (fig 2). The thickness of the subcutaneous WAT was reduced with respect to the controls (figs 3,4). In the liver, MRI showed a lipid accumulation, most marked in the left lobe (figs 5,6).

Histology and TEM

The BAT of control mice displayed a normal morphology, fully consistent with previous studies [la]. The cytoplasm contained numerous large lipid droplets and-the remaininig cytoplasm was almost completely filled wi$h large mito- chondria whose cristae were tightly packed and parallel. 1n the mitochondrial matrix, round electron dense inclusions of W-90 nm were present. Glycogen particles were sparse.

In IFN-treated mice, by light microscopy, adipwytes showed a larger lipid deposit compared &th control mice (figs 7, 8). By TEM, the mitochotidria were reduced in number, smaller and usually with a more irregular shape than in control mice (figs 9-12). T& number of &siae in

each mitochondrial section was also significantly reduc:ec (9.1 + 1.5 vs 20.1 -t 1.9, P < 0.01). The disposition of th cristae was less regular, the typical parallel pattern was fre quently lost and tubular cristae were not infre@ent. Mite chondria with a tubuloreticufar pattern of ihe cfistae wcn rarely visible. The electron dense inclusions in the mite chond&.-matrix :were significantly less numerous in @I+4 treated than in control animals (1.9 f 0.7 VS 0.9 k .0.7 fip mitochondrial~section, P < 0.01). Abnormalities of end<> pIas& reticulum described in hepatdcytes of IFN-treatec mice 17.81 Were not found in brown adipocytes.

Conclusions

In newborn animals, almost al! the adipose. tissue is browr; In the adult animal, BAT accuinulates lipid in an unilocula form and reduces the- number of the mitochondrial cristae This process leads to the formation of a WAT-like tissue and typical BAT represents only a small percentage of tl% total adipose tissue in adult mammals. The factors involver in this process are--poorly understdod. It can be induced b; exposure to a warm environment or by a denervation. Thl present study demonstrates that a new model is now avail able to study -the process df transformation of the EAT into a WAT-like tissue, the IFN-treated-newborn monse. In thi: model the BAT transformation occurs in about 1 week therefore--being more rapid than in physiologic involitior and similar to those in warm-exposed or denervated-ani mals. IFN action is probably due to alterations of the- fipic metabolism that have been described in ~biocheniical stt%Se! and which causes an accu.mulation of neutral @id atid I decrease iti phospholipids ip the liver [ 16, 173. In~BGT, thhe reduction of cristae in mitochondria may be related to the inhibition of the enzymatic systems responsible for the ;syn- thesis of polyunsaturated phosphatidylcholine and phospha- tidylethanolamine described in the liver of suckling mice 1X6]. Mitochondrial changes have not been described in organs other than BAT. This finding can be due to the idio- syncratic biochemistry of BAT mitochondria in which oxi- dation of substrates can be uncoupl&l from phosphoryia$ion of ADP to ATP [Z].

fn our work, the use of a paradigm including in viw evaluation by MRI has allowed us to show -that the deposi- tion of neutral lipid in liver and BAT is paralleied bye a reduction of the s&cutaneous WAT, a previously unrecog- nized effect. This result show$ the utility of contemporary ultrastructural and MRr studids which make it posSible to obtain information on the cytologic tra&foFation in BAT adipocytes in the living animal. The basic method for co@&- lating.the MRI signal with the different cel!$ztr patterns of’ the adipocyt& has been published previously [lo, 12]. if4 and 13C magnetic resonance spectroscopy is also a useful tool for investigation of WAT and BAT chemical composi- tion U51. Now this work demonstrates that the method can be useful in the follow-up of experimental research on l3kl transformation.

In conclusion, the present study demonstrates that IFN induces neutral lipid accumulation and mitochondrial reduction in BAT with features resembling those of the &- ma1 involution of the tissue. This process is paralleled -by a reduction of the subcutaneous WAT and an accumul@ion of lipid in the liver.

In our opinion, the IFN+eated newborn mouse -could represent a model for testing the actidn of treatr&nts on BAT transformation and it shouid be inves@&d v&e&r an endogenous liberation of substances with art IF&J-1.&e

165

Figs l-8. MRI, axial section of the thorax. Control mouse. The triangular medial body of the interscapular BAT and its lateral process emit a very low intensity signal (asterisks). s, spinal medulla. 2. MRI, axial section of the thorax. IFN-treated mouse. BAT shows areas emitting a higher intensity signal (asterisks). 3. MRI, axial section of the cervical region. Control mouse. The subcutaneous WAT (asterisks) is visible. t, trachea. 4. MRI, axial section of the cervical region. EN-treated mouse. The subcutaneous WAT (asterisks) is reduced with respect to the control mouse. 5. MRI, axial section of the abdomen. Control mouse. In the liver the intensity of the signal is low (square). 6. MRI, axial sec- tion of the abdomen. IFN-treated mouse. A lipid accumulation in the liver (square) is visible. 7. Light microscopy of the BAT. Control mouse. 8. Light microscopy of the BAT. IFN-treated mouse showing a more intense lipid content with respect to the control mouse.

166 A Sbarbati et UI

Figs 9-12. 9. TEM, control mouse. Normal brown adipocytes are visible. 4000 x. 10. TEM, control mouse. Enlargement of the typrcal BAT mitochondtia with parallel crisfae and dense inclusions. 25000 X. 11. TEM, IFN-treated mouse. The brown adipoeytes show a reduced number of mitochondria. 3500 x. 12. TEM, IFN-treated mouse. At higher enlargement, the BAT mitochondria show a reduced number of ctistae. 10000 X.

action on lipid metabolism represents a step in the physio- logic process of BAT transformation in a WAT-like tissue. Further study is necessary to demonstrate the eventual bio-

chemical or functional counterparts of the treatment with IFN in BAT. lFN is widely used in the therapy of several diseases and BAT is present in man in all ages 151. The pas-

Interferon and brown adipose tissue 167

sibility that IFN-treatment induces modifications in human brown adipocytes could lx tested.

Acknowledgments

We gratefully acknowledge the technical assistance of Mrs Chantal Maury. This work was in part supported by the Association pour la Recherche sur le Cancer.

References

1 Cameron IL (1975) Age-dependent changes in the morphol- ogy of brown adipose tissue in mice. Texas Rep Biol Med 33,391-396

2 Cannon B, Nedergaard J (1984) The biochemistry of an inefficient tissue: brown adipose tissue. Essays Biochem 20, 110-165

3 Gresser I, Morel-Maroger L, Maury C, Tovey M, Pontillon F (1976) Progressive glomerulonephritis in mice treated with interferon preparations at birth. Nature 263,42W22

4 Gresser I, Tovey M, Maury C, Chouroulinkov I (1975) Lethality of interferon preparations for new-born mice. Nature 258,76-78

5 Heaton JM (1972) The distribution of brown adipose tissue in the human. JAnat 112,35-39

6 Himms-Hagen J (1984) Thermogenesis in brown adipose tissue as an energy buffer. N Engl JMed 311, 1549-1558

7 Moss J, Woodrow DF, Gresser I (1984) Cytochemistry of the tubular aggregates found in hepatocytes of interferon- treated suckling mice. Histochemical J 17,33-41

8 Moss J, Woodrow DF, Sloper JC, Riviere Y, Guillon JC, Gresser I (1982) Interferon as a cause of endoplasmic reticu-

9

10

11

12

13

14

15

16

17

lum abnormalities within hepatocytes in newborn mice. Br J Exp Path01 63,43-49 Osculati F, Leclercq F, Sbarbati A, Zancanaro C, Cinti S, Antonakis K (1989) Morphological identification of brown adipose tissue by magnetic resonance imaging in the rat. Eur J Radio1 9, 112-l 14 Osculati F, Sbarbati A, Leclercq F, Zancanaro C, Accordini C, Antonakis K, Boicelli A, Cinti S (1991) The correlation between magnetic resonance imaging and ultrastructural pat- terns of brown adipose tissue. J Submicrosc Cytol Path01 23, 167-174 Rotbwell NJ, Stock MJ (1979) A role for brown adipose tis- sue in diet-induced thermogenesis. Nature 281,84-88 Sbarbati A, Baldassarri A, Zancanaro C, Boicelli A, Osculati F (1991) In vivo morphometry and functional morphology of brown adipose tissue by magnetic resonance imaging. Anat Ret 23 1,293-297 Sbarbati A, Morroni M, Zancanaro C, Cinti S (1991) Rat interscapular brown adipose tissue at different ages: a mor- phometric study. Znt J Obesity 15,581-588 Tovey MG, Begon-Lotus J, Gresser I (1984) A method for the large scale production of potent interferon preparations. Proc Sot Exp Biol Med 146,809-g 15 Zancanaro C, Nano R, Marchioro C, Sbarbati A, Boicelli A, Osculati F (1994) Magnetic resonance spectroscopy investi- gations of brown adipose tissue and isolated brown adipocy- tes. J Lip Res 35,2191-2199 Zwingelstein G, Brichon G, Meister R, Maury C, Gresser I (1987) Interferon tip induces changes in the metabolism of polyenoic phopholipids and diacylglycerols in the liver of suckling mice. Lipids 22,736-743 Zwingelstein G, Meister R, Abdul M&k N, Maury C, Gres- ser I (1985) Interferon alters the composition and meta- bolism of lipids in the liver of suckling mice. J Interferon Res 5,315-325