doi:10.1006/cyto.1999.0611, available online at http://www.idealibrary.com on

SHORT COMMUNICATION

IN VIVO SUBCELLULAR TARGETCOMPARTMENTS OF INTERFERON-� AND

INTERFERON-� RECEPTOR (�- AND �-CHAINS)IN MOUSE LIVER

Anne Lambert,1 Rabia Sadir,1 Christine Brisson,2 Gerard Morel1

IFN-� displays several effects on different tissues via its specific cell surface receptor (IFN-�R).In order to identify the target compartments of IFN-� and IFN-�R (� and �-chains), we used aquantitative immunogold approach. In physiological conditions, IFN-� and IFN-�R immuno-reactivities were detected in the plasma membrane, in the endoplasmic reticulum area, in themitochondria and in the nucleus. After a single IFN-� injection, we observed, in a quantitativemanner, an increase of signal density without modification of the subcellular distribution ofIFN-� and IFN-�R subunits.

� 2000 Academic Press

IFN-� mediates its pleiotropic functions through aspecific receptor (IFN-�R), which is widely distributedin almost all tissues.1 The IFN-�R is composed of twochains designated IFN-�R� and IFN-�R�, respect-ively.2,3 In addition, IFN-� binds also to heparansulfate (HS). It has been demonstrated that HS con-centrates IFN-� in restricted areas within the liver,mainly in the Disse’s spaces, at the surface of hepato-cytes and endothelial cells.4 IFN-� is then supposed tointeract with its high affinity receptor, a process whichleads to signal transduction.5

In order to show the in vivo internalization andthe subcellular distribution of IFN-� and IFN-�R, weused quantitative ultrastructural immunocytology, inphysiological conditions and after murine IFN-� injec-tion. We found that the cytokine and its receptor areinternalized and distributed in the plasma membrane,mitochondria, endoplasmic reticulum and nucleus.

CYTOKINE, Vol. 12, No. 6 (June), 2000: pp 715–719

RESULTS

Subcellular targets of IFN-�, IFN-�R� andIFN-�R�

In physiological conditions, i.e. the non-injectedmouse, only endogenous IFN-� (Fig. 1A, C, E) wasdetected as IFN-�R� and IFN-�R� in some subcellularcompartments. These three immunoreactivities werevisualized at the plasma membrane level, mainly inthe Disse’s space (Fig. 1A) in the cytoplasm, in therough endoplasmic reticulum compartment area, inmitochondria (Fig. 1C), and in the nucleus (Fig. 1E).

After IFN-� stimulation, i.e. 2 h after IFN-� injec-tion, no modification of signal localization was notedfor the three molecules (e.g. Fig. 1B, D, E), however weobserved a modification of signal density.

From the 1CNRS UMR 5578, Universite Claude Bernard-Lyon 1,Villeurbanne, France; 2INSERM U 433, School of MedicineLaennec, Lyon, France

Correspondence to: Dr Gerard Morel, CNRS UMR 5578, Bat. 404,3eme etage, Universite Claude Bernard-Lyon 1, 43, Blvd du 11Novembre 1918, 69622 Villeurbanne cedex, France; E-mail:gerard.morel@univ-lyon1.fr

Received 8 February 1999; received in revised form 26 July 1999;accepted for publication 25 August 1999

� 2000 Academic Press1043–4666/00/060715+05 $35.00/0

KEY WORDS: electron microscopy/endogenous IFN-�/immunogold quantification/subcellular target

Quantification of dataThe quantification of endogenous IFN-� and

IFN-�R immunoreactivities in these different cell com-partments are summarized in Figures 2, 3 and 4. Allthese data were significantly higher than the back-ground (0.8 gold particles/100 �m2). No significantdifference was shown for IFN-� immunoreactivitybetween the densities (number of gold particles/100 �m2) observed in endoplasmic reticulum area,mitochondria, and in the nucleus (111�23, 90�29,and 131�7, respectively). On the other hand, thedensity of IFN-�R� signal was significantly different

715

716 / Lambert et al. CYTOKINE, Vol. 12, No. 6 (June, 2000: 715–719)

Figure 1. Subcellular target compartments of IFN-�.

Before (A, C, E) and after IFN-� administration (B, D, F). The immunoreactivity (10 nm gold particles) is observed at the plasma membranelevel (A, B), in the mitochondrial and endoplasmic reticulum areas (C, D), and in the nucleus (E, F). *Disse’s space. Scale bar=300 nm.

between the same compartments (P<0.05) (83�26,206�69, and 34�4, respectively). We noted thatIFN-�R� was mainly detected in mitochondria. ForIFN-�R�, no significant difference was observed inendoplasmic reticulum and in mitochondria but asignificantly lower signal was detected in the nucleus(119�26, 113�24, and 36�5, respectively).

After IFN-� injection, the ultrastructural quanti-fication showed that all data obtained were signifi-cantly different from those without injection (P<0.05)(Figs 2, 3 and 4). At the plasma membrane level thenumber of gold particles/100 �m was increased morethan 3, 4 and 2 times for IFN-�, IFN-�R� andIFN-�R�, respectively. In the endoplasmic reticulum

Subcellular target compartments of IFN-� and its receptor / 717

area, the signal density (gold particles/100 �m2) wasincreased 3, 2.4 and 1.4 times for IFN-�, IFN-�R� andIFN-�R�, respectively. In the mitochondria theseincreases were 2.7, 1.2 and 1.7 for IFN-�, IFN-�R�and IFN-�R�, respectively. In nuclei, the signal densityincreased 2.2, 3 and 1.8 times for IFN-�, IFN-�R� andIFN-�R�, respectively.

0

500

Gol

d pa

rtic

les

nu

mbe

r/10

0 µm

2

Nu

cleu

s

0

500

450

400

350

300

250

200

150

100

50

Gol

d pa

rtic

les

nu

mbe

r/10

0 µm25

20E

ndo

plas

mic

ret

icu

lum

Mit

och

ondr

ia

Con

trol

Pla

sma

mem

bran

e

Con

trol

5

15

10

Figure 2. Quantification of the IFN-� immunoreactivity.

The quantification of data was performed for the different compartments before and after IFN-� injection. The increase of signal density is higherin the endoplasmic reticulum area. Non-injected liver (open bars) and liver injected with IFN-� (shaded bars).

DISCUSSION

In order to exercise its biological activities, IFN-�binds to its specific cell surface receptor, the IFN-�R�.These results have shown important peripheral label-ling at the plasma membrane level, probably indicatinga ligand-receptor complex. It has previously beenshown that hepatocytes were target cells for IFN-�.6

This immunoreactivity at the plasma membrane levelmay reflect another complex. Indeed, IFN-� presents ahigh affinity for HS. This glycosaminoglycan is foundat the cell surface of some tissues, notably in the liver.7

We have demonstrated recently that, in contrast to

growth factors, interactions of IFN-� with IFN-�R�and with HS are mutually exclusive.8 Since HS ispresent in larger amounts than the receptor at the cellsurface,9 IFN-� should bind preferentially to HS,which could constitute a storage structure forIFN-�.

The immunoreactivity observed in the cytoplasmmay reflect endocytosis and transport of IFN-� andIFN-�R in vesicles, probably in caveolae. In a recentstudy, it has been shown by electron microscopy thatanother cytokine, interleukin-8 was internalized andtranscytosed via caveolae.10 Immunodetection showingreactivity in or near the citernae of the endoplasmicreticulum could suggest synthesis of IFN-�R byhepatocytes.

The presence of IFN-� in the nucleus has beenpreviously demonstrated.11 Other cytokines and hor-mones such as growth hormone and prolactin, havebeen also detected in the nucleus.12 IFN-� has beenshown to contain a functional nuclear localizationsequence.13 On the other hand, we observed thepresence of the two subunits of IFN-�R in the nucleus

described.

718 / Lambert et al. CYTOKINE, Vol. 12, No. 6 (June, 2000: 715–719)

in a quantitative manner. It is probable that a partof IFN-� and its receptor enters the nucleus in acomplexed form.14

In conclusion, the current study demonstrates, inpart, the in vivo internalization and nuclear trans-location of endogenous IFN-� and IFN-�R (�- and�-chains) and their subcellular distribution. Moreover,the injection of IFN-� shows an increase of the signaldensity of IFN-�, but also of IFN-�R, suggesting eitheran upregulation or a mobilization of the receptor.These data support the growing evidence that hepato-cytes are actively involved as target cells in the cytokinenetwork during cellular response to IFN-�.

MATERIALS AND METHODS

The mice were killed 2 h after intraperitoneal injectionof 5 �g of recombinant IFN-� (kind gift from RoussellUclaf) or 150 mM NaCl alone. Livers from non-treated orinjected mice were fixed in 2% paraformaldehyde–0.05%

glutaraldehyde for 24 h. Specimens were embedded inhydrophilic LR white resin.

Immunocytological reactionSections were incubated for 1 h with either polyclonal

IgG raised against IFN-� (1000 U/ml) (Biosource, Camarillo,CA, USA), or polyclonal IgG raised against IFN-�R� (10 �g/ml), or polyclonal IgG raised against IFN-�R� (20 �g/ml)(Santa Cruz Biotechnology, California, USA). Antigen–antibody complexes were revealed with anti-rabbit IgG con-jugated with 10 nm gold particles (Biocell, Cardiff, UK). Theultrastructural quantification was determined as previously

15

0

500

Gol

d pa

rtic

les

nu

mbe

r/10

0 µm

2

Nu

cleu

s

0

500

450

400

350

300

250

200

150

100

50

Gol

d pa

rtic

les

nu

mbe

r/10

0 µm15

10

En

dopl

asm

ic r

etic

ulu

m

Mit

och

ondr

ia

Con

trol

Pla

sma

mem

bran

e

Con

trol

5

Figure 3. Quantification of the IFN-�R� immunoreactivity.

The increase of signal density is higher 2 h after IFN-� injection at the plasma membrane and in the nuclear compartment, compared withphysiological conditions. Non-injected liver (open bars) and liver injected with IFN-� (shaded bars).

Acknowledgements

We are grateful for the support of Dr H. Lortat-Jacob. This work was supported by the ‘‘CentreNational de la Recherche Scientifique’’, the ‘‘Associ-ation pour la Recherche sur le Cancer’’ (No. 1209), andthe ‘‘Fondation pour la Recherche Medicale’’.

Subcellular target compartments of IFN-� and its receptor / 719

0

500

Gol

d pa

rtic

les

nu

mbe

r/10

0 µm

2

Nu

cleu

s

0

500

450

400

350

300

250

200

150

100

50

Gol

d pa

rtic

les

nu

mbe

r/10

0 µm15

10

En

dopl

asm

ic r

etic

ulu

m

Mit

och

ondr

ia

Con

trol

Pla

sma

mem

bran

e

Con

trol

5

Figure 4. Quantification of the IFN-�R� immunoreactivity.

A similar increase of signal density is visualized in all compartments 2 h after IFN-� injection. Non-injected liver (open bars) and liver injectedwith IFN-� (shaded bars).

REFERENCES

1. Valente G, OzmenL , Novelli F, Geuna M, Palestro G,Forni G, Garotta G (1992) Distribution of interferon-� receptor inhuman tissues. Eur J Immunol 22:2403–2412.

2. Hemmi S, Bohni R, Stark G, Di Marco F, Aguet M (1994)A novel member of the interferon receptor family complementsfunctionality of the murine interferon gamma receptor in humancells. Cell 76:803–810.

3. Aguet M, Dembic Z, Merlin G (1988) Molecular cloningand expression of the human interferon-� receptor. Cell 55:273–280.

4. Lortat-Jacob H, Brisson C, Guerret S, Morel G (1996)Non-receptor-mediated tissue localization of human interferon-�:role of heparan sulfate/heparan-like molecules. Cytokine 8:557–566.

5. Heim MH (1999) The jak-STAT pathway: cytokine signal-ling from the receptor to the nucleus. J Recept Signal Transduc Res19:75–120.

6. Masuyama J, Minato N, Kano S (1986) Mechanisms oflymphocyte adhesion to human vascular endothelial cells in culture.J Clin Invest 77:1596–1605.

7. Stow JL, Kjellen L, Unger E, Hook M, Farquhar MG(1985) Heparan sulfate proteoglycans are concentrated on the sinu-soidal plasmalemmal domain and in intracellular organelles ofhepatocytes. J Cell Biol 100:975–980.

8. Sadir R, Forest E, Lortat-Jacob H (1998) The heparansulfate binding sequence of interferon-� increased the on rate of theinterferon-�-interferon-� receptor complex formation. J Biol Chem273:10919–10925.

9. Yanagishita M, Hascall VC (1992) Cell surface proteo-glycans. J Biol Chem 267:9451–9454.

10. Middleton J, Neil S, Wintle J, Clark-Lewis I, Moore H,Lam C, Manfred A, Hub E, Rot A (1997) Transcytosis and surfacepresentation of Il-8 by venular endothelial cells. Cell 91:385–395.

11. Bader T, Wietzerbin J (1994) Nuclear accumulation ofinterferon-�. Proc Natl Acad Sci USA 91:11831–11835.

12. Morel G (1994) Internalization and nuclear localization ofpeptide hormones. Biochem Pharmacol 47:63–76.

13. Subramaniam PS, Mujtaba MG, Paddy MR, Johnson HM(1999) The carboxyl terminus of interferon-gamma contains a func-tional polybasic nuclear localization sequence. J Biol Chem 274:403–407.

14. Mertani HC, Morel G, Lobie PE (1999) Cytoplasmic andnuclear cytokine receptor complexes. Vit Hormones 57:79–121.

15. Mertani HC, Waters MJ, Jambou R, Gossard F, Morel G(1994) Growth hormone receptor binding protein in rat pituitary.Neuroendocrinology 59:483–494.