Induction of the pro-myelocytic leukaemia gene by type I and type II interferons.

The physiological role of the pro-myelocytic leukaemia (PML) gene product is poorly defined. Among other functions, PML is involved in haematopoietic differentiation and in control of cell growth and tumorigenesis. We investigated the regulation of human PML expression by interferons (IFNs) and IL-1 in various human haematopoietic lines (U937, THP1, HL60, NB4), in human diploid fibroblasts and in human peripheral blood leukocytes. Cytokine-induced modulation of PML expression was assessed by Northern blot analyses, flow cytometry studies and in situ immunolabelling. Our data show that IFNs and IL-1 upregulate PML transcript and protein expression in a time and dose-dependent manner. In situ immunolabelling revealed that upregulation of protein expression by IFN-alpha is a consequence of a marked increase in both the number and the intensity of the staining of so-called PML nuclear bodies. Our data suggest that stimulation of PML expression by interferons and IL-1 may account for upregulation of PML proteins observed in inflammatory tissues and in proliferative states.


Introduction
The PML (pro-myelocytic leukae mia) ge ne w as ide ntifie d originally as a fusion partner of the RARa gene in the chromosomal translocation t(15;17) specific for acute promyelocytic le ukae mia (APL). [1][2][3][4] Seve ral PML-cDNAs have be en isolate d, re sulting from alternativ e RNA splicing in the 39 coding re gion of a single gene . 5 The e ncoded proteins have a pre dicte d molec ular w eight ranging from 47 to 98 kDa. PML prote ins be long to a hete rogene ous family of DNA-binding prote ins characte rized by the C 3 HC 4 zinc binding ring finger. PML prote in ex pre ssion fe ature s a typic al spe ckled nucle ar pattern w hich is the conse que nce of the localization of the prote in in nuclear c ompartme nts name d nuclear bodie s. 2 The func tion of the PML nuclear bodies is still unknow n. Rec ently it has be e n show n that PML participates in re gulation of haematopoie tic differe ntiation and in c ontrol of ce ll grow th and tumorige ne sis. 6 In addition, there is e vide nce that PML protein is aberrantly ex pre sse d in a varie ty of p athological c onditions including malignanc ie s. 7,8 The le ve l of PML ex pre ssion is upre gulate d in inflammatory tissues and in be nign or malignant prolife rative state s. 7 -9 Cytokine s, espe cially inte rferons (IFNs), are likely to re pre sent some of the me diators that may be involve d in overex pre ssion of PML under the se c irc umstance s. Interferons are polypeptide s that mediate pleiotropic e ffec ts on se nsitive ce lls (re view ed in Sen and Leng yel 1 0 ). Major activities of IFNs include immunomodulating ac tiv-itie s, antiviral effects and inhibition of c ellular grow th. Tw o families of IFNs can be distinguishe d: type I IFNs (IFN-a /b and IFN-v ) and type II IFN (g ). Type I IFNs are enc oded by a family of over 20 gene s, w here as type II IFN is structurally unrelated and enc ode d by a single ge ne . 1 0 An important func tion of IFNs is re gulation of ce llular grow th of normal and malignant ce lls. The molec ular me chanisms involve d in grow th inhibition are poorly unde rstood.
Here , w e investigate d modulation of PML ex p re ssion by inte rferons, TNF-a and IL-1. We pre se nt e vide nce that PML gene ex pre ssion is enhance d by type I and type II inte rferons and by IL-1. We analyse d cytokine induc ed up-re gulation of PML transcript and prote in ex pre ssion in human haematopoie tic ce ll lines, in human diploid fibroblasts and in human peripheral blood leukocyte s. Modulation of PML prote in ex pre ssion by IFN-a and IL-1 w as asse sse d w ith immunofluore scenc e flow c ytometry. Immunofluoresce nce in s itu analysis of PML nuclear bodie s re vealed that IFN tre atment re sulted in up-re gulation of both the inte nsity and the number of the se nuclear struc tures.

Northern blot analysis
Total RNA w as is olated by the me thod of Chomc zynski and Sac chi 12 and Northe rn blot analysis w as performe d essentially as desc ribed. 11 RNA (20 m g) w as subje cte d to e lectrophore sis on a 1% formaldehyde agarose gel and transfe rre d to nylon me mbrane s (Hybond N, Ame rsham Buchler, Braunschw e ig, Ge rmany). All hybridizations w e re c arried out at 65°C for 16 h in a solution c ontainin g 1% BSA, 7% SDS, 1 mM EDTA, and 0.5 M sodium p hosphate buffer (pH 7.2). Filte rs w e re hybridized w ith PML cDNA inse rt radiolabe lle d by the random p riming method. A partial cDNA p robe for PML w as synthesized by RT-PCR from total RNA of THP1 ce lls using prime rs S1: 59 -ACCAGTCGGTGCGTGAGTT-39 and AS1: 59 -TGGATCTCTGCGCTGATGTC-39 , c orre sponding to nucleotide s 525-544 and 715-734, re spec tively of the published se que nce . 13 This part of the PML se que nce is 100% homologous for all PML ge ne s. The synthesized cDNA w as controlle d by DNA-se que ncing and w as subclone d into pCRTM II using a c omme rcial cloning kit (Invitrogen, Leek, Netherlands). For quantitativ e analysis of mRNA le vels de nsitome tric scanning of autoradiographs w as applied by using the Quanti-sc an® program. PML-spec ific transcript le vels w ere normalized to b -ac tin mRNA ex pre ssion.

Immunofluorescence analysis
Quantitativ e immunofluore sc enc e studies of PML ex pre ssion w e re p erformed by flow c ytometry (EPICS, Coulter, Ge rmany) as described 11 by using saturatin g conc entrations of a PML spe cific mAb. 1 4 Normal mouse IgG (2 m g/ml) (Coulter, Krefe ld, Ge r-many) w as used as background control, and FITClabelled goat-anti-mouse IgG (Coulte r, Krefe ld, Ge rmany) as sec ond re age nt. A total of 50 000 ce lls w as analysed for each dete rmination and spec ific fluore sce nce intensity w as calculated on the basis of background fluore sc enc e w ith control antib odie s by using the EPICS program. Mean spe cific fluore sc enc e inte nsity is given as channe l numbe r on a log sc ale from 1 to 1024 or in pe rc entage of the re spe ctive untreate d c ontrol.

Indirect immunofluorescence microscopy
Ce lls w ere air drye d ove rnight, fix ed for 5 min w ith aceton follow ed by inc ubation w ith PG-M3 antibody 1 4 as undilute d hybridoma supe rnatant or w ith normal mouse IgG (Coulter, Krefeld, Ge rmany) at a conc entration of 2 m g/ml for 5 min. Ce lls w ere then w ashe d w ith PBS and inc ubated w ith a FITC labelled goat-antimouse antibody (Coulte r, Kre feld, Ge rmany) (5 m g/ml). Ce lls w ere analysed w ith a fluore sc enc e mic rosc ope (Orthoplan, Leitz, We tzlar, Ge rmany).

PML mRNA levels are upregulated by type I and type II interferons
To inve stigate w he ther IFN affec ts PML gene ex pre ssion, w e pe rforme d Northern blot analyse s of total ce llular RNA ex trac te d from various hae matopoie tic ce ll lines cultured w ith and w ithout IFNs. In all ce ll type s te ste d inc luding the myeloid c ell line s HL60 and THP1, human diploid fibroblasts and human pe ripheral blood mononucle ar c ells a marked incre ase in the thre e major PML transc ript levels w as obse rved by tre atment w ith both type I (a , b ) and typ e II (g ) inte rferons (Figs 1-4). Dose re sponse ex periments show ed that upre gulation of PML mRNA le vels w as alre ady evide nt upon incubation w ith 30 IU/ml IFN-a for 6 h ( Fig. 1). At IFNa dose s of 30 IU/ml and 3000 IU/ ml, quantitativ e analysis employing de nsitome tric sc anning re veale d a 3.2-fold and a 5.3-fold induction of PML mRNA le vels, re spec tively (Fig. 1). As a c ontrol, PML transc rip t le ve ls w ere monitore d upon inc ubation for 6 h w ith TNF-a and LPS and w e re found to be unchanged (Fig. 1). In ex pe riments ( Fig. 2 and data not show n) investigating the kinetics of IFN-a induc e d PML mRNA upre gulation in PBMNC, max imum enhance ment of PML transcript ex pre ssion w as alre ady observe d upon inc ubation for 3 h (3000 IU/ml IFN-a ). The re after PML transc ript le ve ls decline d and w ere barely visible after 24 h de spite c ontinuous inc ubation w ith 3000 IU/ml IFN-a (Fig. 2). In contrast to IFN-a , IFN-b (3000 IU/ml) induce d PML transcript induc tion incre ased gradually ove r time and max imum enhance ment w as re ache d at 24 h (Fig. 2). The time course of IFN-g induced PML transc ript enhanc eme nt during a 6-h-inc ubation period w as similar to IFN-a (Fig. 3). Pre vious studie s indic ated that transcriptional re gulation of the PML gene is involved in up-re gulation of PML mRNA by IFNs. 1 5,16 To c onfirm that IFNinduc ed modulation of PML mRNA is indepe nde nt from de n o vo protein synthe sis, w e pe rforme d PML g e ne e xpre s s io n is u pre g u la te d by in te rfe ro n s a n d IL-1 Human diploid foreskin fibroblasts were treated with IL-1a (100 IU/ml), with IFN-a (1000 IU/ml) or with a combination of IFN-a (1000IU/ml) and IL-1a (100IU/ml) for 6h. Cells were grown in the absence or presence of an inhibitor of protein tyrosine kinases (genistein 50 m g/ml). Northern blotting was performed as described above. For quantitative analysis, PML-specific transcript levels were analysed by densitometric scanning and normalized to b -actin mRNA expression as described in Materials and Methods.

Mediators of Inflammation
Northern blot analyse s w ith and w ithout an inhibitor of protein synthesis (CHX). These ex pe riments demonstrate d that CHX had no e ffec t on IFN-induce d upre gulation of PML mRNA (data not show n).
The influenc e of IL-1a on PML transc ript ex pre ssion w as investigate d in human foreskin fibroblasts (Fig. 4). Tre atme nt w ith IL-1a at a dose of 100 IU/ml for 6 h re sulted in a 3.8-fold upre gulation of PMLspe cific mRNA le ve ls as analysed by de nsitome tric sc anning. A more pronounce d e nhance ment of PML transc ript levels (7.1-fold induction) w as observe d w hen ce lls w ere inc ubate d w ith IFN-a . IL-1 in combination w ith IFN-a stimulated PML mRNA ex pre ssion to more than tw ic e the level obtained w ith IFN-a alone (15.3 fold induction re lative to me dium control), as estimated by densitometric sc anning. Coinc ubation w ith ge niste in, 17 an inhibitor of tyrosine kinases, significantly inhibite d the IFN-induce d inc re as e in the le vel of PML transc ripts, w he re as no effect of ge niste in on IL-1 induce d PML mRNA ex pre ssion w as noted (Fig. 4). Similar re sults w ere obtain ed using the amnion fibroblast ce ll line WISH (data not show n).

Influence of IFN-a on PML protein expression
In an atte mpt to investigate modulation of PLM prote in ex pre ssion, w e pe rforme d flow c ytometry analyses. IFN-a treatme nt of the thre e mye loid ce ll lines HL60, NB4 and U937 re sulte d in upre gulation of PML prote in ex pre ssion. For ex ample , in HL60 cells, upre gulation of PML ex pre ssion w as dete ctable upon inc ubation w ith IFN-a for 18 h (Fig. 5). The time course of IFN-induc ed e nhanc eme nt of PML prote in ex pre ssion w as monitore d in U937 c ells (Fig. 6). Induction of PML w as not de te c table upon inc ubation w ith IFN-a (1000 IU/ml) for only 6 h or 12 h but starte d there after and re ached a plate au at 20 h ( Fig. 6 and data not show n). Dose -re sponse ex perime nts in U937 cells re vealed that incubation w ith 10 U/ml IFNa for 20 h already re sulte d in slight upre gulation of PML prote in ex pre ssion to 113% of the untreate d control (Fig. 7). Max imum upre gulation of PML ex pre ssion to 160% of that of untre ate d cells w as achie ved by inc ubation w ith dose s gre ater than 500 IU/ml IFN-a (Fig. 7). The se data indicate that half- max imum IFN-a re sponse is re ached at 50 U/ml IFN-a w hich c orresponds to the IFN-a conc entration ne eded to achie ve a half-max imum re sponse in other biological ac tivitie s of IFN-a . 1 0 When human foreskin fibroblasts w ere grow n in the absenc e or or in the pre se nce of saturatin g conc entrations of IL-1a , a ne arly tw o-fold inc re ase of PML prote in ex pre ssion w as de te c te d by flow cytome try analysis (Table 1). A similar re sult w as obtaine d upon inc ubation w ith IFN-a (100 IU/ml). The combination of IL-1 and IFN-a stimulated PML protein ex pre ssion to about three times the level observe d in untreate d c ontrols (Table 1).

Analysis of PML nuclear bodies
Sinc e it has be en show n that PML prote ins are locate d w ithin subnuclear structure s from ac tive sites of transc ription and splicing, the so-calle d PML nuclear bodie s w e re monitore d during IFN-a treatme nt using indire ct immunofluore sc enc e staining. PML-specific immunofluore sc enc e varied in inte nsity be tw ee n the ce ll lines te ste d. In all cell lines te ste d (NB4, HL60, U937) and in pe riphe ral blood leukoc ytes the constitutiv e leve ls of e ndogenous prote ins w ere de te c table (Figs 8 and 9). Pe ripheral blood le ukocytes re vealed spe cific immunolabelling of PML nuclear bodie s in ne utrophils and monoc ytes (Fig. 8). When HL60 and U937 ce lls w e re inc ubated for 15 h w ith and w ithout IFN-a (1000 U/ml) a marke d enhanc eme nt of the immunolabe lling w as observe d in treate d ce lls. Both the inte nsity of the staining and the number of the labe lle d PML nucle ar bodie s w as inc re as ed signific antly ( Fig. 9 and data not show n).

Discussion
Our data c onfirm and ex te nd pre vious studies demonstrating that c ytokines are involve d in modulation of PML ge ne ex pre ssion. 15 ,16 We show that PML transc ripts rapidly acc umulate upon treatme nt w ith inte rferons in a varie ty of haematopoie tic ce ll line s and in primary diploid ce lls. How e ve r, our ex perime nts demonstrate a marked diffe re nc e in the time course of PML-mRNA induction be tw e en IFN-a and IFN-b . Induction of PML in re sponse to IFN w as show n to be inde pendent from de n o vo prote in synthesis and ac tinomyc in sensitive indicating that PML mRNA le ve ls are transcriptionally re gulated by IFNs. 16 Re ce ntly, evidenc e has be e n provided that the PML ge ne is transcriptionally upre gulated by IFNs through activation and binding of STATs to an ISRE and a GAS eleme nt in the untranslate d first ex on. 1 8 IFNs activate STATS through tyrosine phosphorylation by me mbers of the JAK tyrosine kinase family. 1 9 Our re sults demonstrating inhibition of IFN-induc ed upregulation of PML transcripts by the tyrosine kinase inhibitor geniste in are c onsiste nt w ith the se findings. Flow cytome try analyse s using a spe cific monoclonal anti-PML antibody re c ognizing the aminote rminal domain clearly demonstrate that IFNs induce PML Fibroblasts were incubated with and without the indicated cytokines at the indicated dose for 21.5 h. Induction of PML protein expression was analysed by flow cytometry using a specific anti-PML monoclonal mouse antibody and normal mouse IgG as control as described above. Given is the mean log specific fluorescence intensity as described in Materials and Methods.
FIG. 8. In situ immunofluorescence analysis of PML proteins in human peripheral blood leukocytes. Indirect immunofluorescence analysis of a peripheral blood smear was performed using the specific mouse monoclonal antibody PG-M3 (B) and normal mouse IgG (A) as control as described in Materials and Methods. Specific immunolabelling can be detected in neutrophils and monocytes which were identified by morphologic criteria.
prote in in a dose depe ndent manner. Re ce ntly, it has be e n show n that PML p rote in ex pre ssion has a spe ckled nuclear pattern. 1 5 This is due to localization of the prote in in p oorly de fine d nuclear organelles named nucle ar bodie s. The func tions of the nuclear bodie s are unknow n ex ce pt for the c oiled bodie s w hich act as storage compartments for splicing fac tors. 2 0 To analyse ex pre ssion of PML p rote ins w ithin nuclear bodies w e pe rformed indire ct in s itu immunofluore sc enc e studie s. In s itu immunolabelling of PML prote in ex pre ssion re ve aled an IFNinduc ed incre ase in numbe r and intensity of so-calle d PML nuclear bodie s. To the be st of our know ledge induc tion of PML by IL-1 has not be e n de scribe d so far. Here w e show that IL-1a rapidly induc es PML mRNA ex pre ssion in fibroblasts. Since IL-1 is a potent induce r of IFN-a and IFN-b 2 1 it could be argued that the obse rved induction of PML mRNA ex pre ssion is a conse que nce of IL-1 induc ed autocrine produc tion of inte rferon-b . We addressed this is sue by e mploying the tyrosine kinase inhibitor geniste in. Ge niste in w as show n to strongly inhibit IFN signalling. 1 7 Our ex p eriments clearly show that in fibroblasts geniste in is not able to suppre ss the IL-1 induc ed upre gulation of PML transc ripts, w here as IFN-induced e nhance ment of PML-spec ific transcripts w as strongly inhibite d. Thus, indire ct re gulation of PML by IL-1 induc ed autoc rine production of IFN-b app ears unlike ly. Flow c ytometry analysis re veale d that IL-1 and IFN-a ac t synergistic ally in induction of PML proteins. This re gulation c ould be involved in the observe d upre gulation of PML p rote ins in inflammation. In c ontrast to normal tissue s, in inflammatory tissue s the level of PML prote ins is upre gulate d. 7 ,8 Local produc tion of IL-1 has be en demonstrate d in inflammatory and autoimmune disease and this may c ontribute to tissue damage and re pair. 21 ,22 In fibroblasts, IL-1 mediates tissue re p air after inflammation and injury by re gulating ke y prote ins ne ce ssary for formation of the ex tracellular matrix . 21 ,2 2 The data pre sente d here sugge st that IL-1 induc ed PML ex p re ssion may re pre sent the molec ular basis for the observed upre gulation of PML prote ins during inflammation. IFNs, as part of the cytokine ne tw ork are indire ctly induced by re gulators such as tumour ne c rosis factor during immune re sponses and inflammation. This may contribute to upre gulation of PML in inflammatory tissue s by syne rgistic action of IFN-a and IL-1 in enhancing PML prote in ex pre ssion in fibroblasts .