el B/og/~,s/mAria, 1135(1992t171-179 1992Eis~'ier Science Publishers B.V. All fights reser~'ed0167-4889/92/$05.00

~zi~

171

In vivo and in vitro induction of 'tissue' transglutaminase in rat

hepatocytes by retinoic acid M . P i a c e n t i n / ~ , M . P . C e r h b, L. D i n i ~, M . D i R a n ", L. P i r e d d a a, V . Tl~omazy d P.J.A, Davies ° and L. Fesus ~ a D e p a r l ~ o [ B i o ~ . Unit'e~ity o f Rome " T ¢ Vergata'. Rome (Italy), h D¢imrtraen t o f Science ttrRf Biotnedical Tedlnolo, g~ and Biometry. Unk'ersdt~"o f L 'A~ila. L "Aquila (Italy) ~ Departments of Bi~hemistry and a Palhologr" L'nil'er~i(y Medical School o [ De,br~cet~ Debrecen (Hltngoey) and • Department of Pharmacology; Unil'e~lty o f Texas, Medical School at Hou~te~l, Ilouston, T X (USAi

(Received 12December 1991)

Keywords: AlXlgt~sis;Di- and lmlyamine-/-glutamylderivative:Cellproliferation;Retinoi¢acid ~ssce transglutaminase (tTG) expression was found to be induced in rat liver following in vivo mtinnic acid (RA) treatment (Piaoemini ct al. (I988) B/ocbem. J. 253, 33-38). Here we show that the increased enzyme expression in rat liver is at least partially the result of the action of ~ in parenehymal cells. In fact, (a) when hepat.lcytes are isolated from RA-4reated animals their transglmaminase protein content is much higher 1than in similarly isolated cGi:,rol cells; (b) higher tTG protein level is also found by imatunoclectronmicroscopy in the hepatoeytes of the ILA-treated rats as compared with the ve~ low amount detected in the controls; (c) RA induces t'I'G in hepatoq,les under culture conditions as well. One of the functions of tTG is m form a protein polymer in dying apoptotic cells by e(y-glulamgl)lysineand, specifically ~.-glutamylpolyatnine cross-links (Fe.~'rg et at. (1989) F-EBS LeH. 245, 150-154). Noteworthy, after in vivo and in vitro RA-treatment we could not determine any increase (there was even a slight decrease) in the number of die cross-linked apoptotic envelopes. In keeping vdth this is the significant reduction of protein bound ~,-glutamylpolyaminedetected in hepatceytes exposed to RA in culture. These findings suggest that the ILat-induced tTG in parenchJmal cells is an inactive form.

Introdudioa Transglutaminases (EC 2.3.2.13) are a group of Ca2+-depandant e r t ~ ~talyzing crqss-linking reactions among polypeptide chains through the establishmerit of d?.-glutamyl)lysine, N,N-bis(-y-glutamyl)polyamine linkages and other 7-glutamylamine derivatives [1-5]. The formation of covalent bonds among polypeptide chains results it, protein polymers highly insoluble in detergents and chantropic agents [1-3,5-6]. Three biocbemieally and immunologicafly distinct transglutaminase gen¢ products, namely blood coagulation factor XIII, tissue and karatinocyte transglotaminases, have so far been characterized [7-9] and the existence of at least two others has been suggested [10,11]. As far as tbe biological role of transgintami noses is concerned it seems that the various T G forms are involved in apparently different phenomena such .ed~bfeviations: t'l~. tissue ~r~lrsglatamirlase;~

retino/cacid.

Correslzondenc¢: M. piacentini, Deparrmenl of Biology,Unh~ity of Rome 'Tot Vergata',Via F- Cameval~ 00173Rome, Italy.

as blood coagulation, wound healing, clotting of seminal fluid, terminal differentiation of kcratinceytas and cell death by apoptosis, however, all these events deal with general protection of cell and tissue integrity [2,6,12-141. 'Tissue' (liver) transgintaminase (tTG) was first deserved in and ~ater purified from liver homogenate [15,16[. in mammalian liver this ~vinble form of the enzyme accounts for over 90% of the total transglutaminase activity [16,17]. However, immanohistochemical studies indicate that tTG in liver is expressed in endothelial (cloning of human tTG was achieved from an endothelial cell eDNA library [18]) and smooth muscle cells and in those rare scattered hepatocytcs undergoing physiological cell death (apoptosis) [14,19-21]. In fact, it has been recently demonstrated that in mammals the tTO gene is among those specifically induced during the apoptotic program [6,19-21]. The aclivatian of t'l'G in apoptotic cells results in the assembly of a highly cross-linked protein seaffuld which prevents the release of harmful intraceflular components from the dying ceils ]6,22], thus avoiding inflammation in the surrounding tissues [23,24]. The cross-linked protein

172 enve.h3pe presant in mature apoptotie cells is stabilized by polyamine-derived and e(y-glutamyl)lysine crosslinks [6]. The findings that tTG-catalyzed cross-finks are normally detectable in liver and in isolated hepatocytes [6,25,26], as weft as that neosynthesis of tTG is induced by retinoic acid (RA) in rat liver [27] raise the following questions: (i) is RA-indueed tTG expressed in hepatocytes a n d / o r in other liver cells?; (ii) is the RA-induced tTG activity related to apoptosis in liver evils? In an attempt to address these points we have studied tTG expression in normal hepa~Jcytes and upon RA-treatment both in rive and in vitro. Materials end Meih0ds

Chemicals [l,~n)-3H]putrcscine dibydiochloride (26.3 C i / retool) was purchased from Amersham (Bucks); Instagal II was from Packard (Zurich). N,N'-dimethy[casein, bovine serum albumin, di- and polyamine bydrochlorides were from Fluka (Buehs). All-trans retinnie told was from Sigma (St. Louis, MO). Cell culture media were from Flow (Irvine, VA). OnmiTags Avidin-Biofin immuuoperoxidase staining system was from Immtmon (Troy, MI). Other chemicals were of reagent grade and used without further purifiealion. Animals Wistar male rats, weighing 150-200 g, were ubed. The animals were kept ..at 20-22~C on 12 h-lig_ht/12 h-darkness and fed ad |ibimm with a standard diet. Rats were intrapcntoneally injected with 20 m g / k g retinoie acid dissolved in olive oil (controls were injected with olive oil alone). At different time intervals, r~ts were anaesthes[zed with 1O0 m g / k g sodium thiopontul (Farmotul, Farmitalia) and the livers were perfused with cold 0.25 M s u c r o ~ through the portal vein. Liver homogenates (2.5%, w/v) were prepared in 0.25 M sucruse containing 1 mM E D T A (pH 7). Cell cultures Hepatocytes were isolated using the collagenase perfusinn method as described [28]. Ceils were washed three times in RPMI-1640 and then seeded to a cell concentration of 0.5 • 106 cells/ml on collagen coated dishes (10 cm di:lm~ter) in RPMI-1640 containing 10% heat inactivated fetal bovine serum, 10 m g / I insulin, 10 m g / I hydrocortisone, 10 rag/1 inusioe, 60 m g / I gentamicin and 0.5 m g / I fungizune in humidified atmosphere with 5% (v/v) CO 2 at 3"PC. Cell treatment was po_rfurmed by adding 5 ttM R A added from a 5 mM stock solution di~oived in 70% ethanol (ethanol was also supplemented to control cultures). The medium was replaced daily.

Enzyme ax~' and chonical determinations Transglutaminme activity was measured by detecting the incorporation of [3H]putrescine into N,N'-dimethylcasein as previously reported [38]- One enzyme unit is defined as the amount of enzyme linking I nmol of pulrescine to N,N'-dimetbylcasein/h. tTG protein concentration in cell homogenate was as~sayed according to the sandwich ELISA method previously described [29]. Protein was assayed according to Lowry et at. [30] using bovine serum albumin as standard. Total putrescine, spermidin¢ and spctmine were determined as dausyl (5-dimethylaminouaphtalane-Isulfonyl) derivatives in double aliquoLs of cell suspension, hydrol)~ed in 6 M HCI for 16 h at II0~C. The deusylation procedure was as described before, using 1,6-diaminohexune as the Llternal standard and toluene to extract the derivati~es from the reaction mixture [26,31]. The chromatographic assays were performed on a Beckman System Gold high-pressure liquid chromatography apparatus equipped with a 4.6 × 250 mm Ultraspber¢ O D S column (5 /tm particle diameter). Free l~lyarnines and their metabolites were identified by adding 5 p.Ci/ml [~H]putrescine to the culture medium for the last 24 h of the various treatments. ~7~ellswere washed in phusphato-buffcred saline (PISS) and treated with 10% cold TCA. The supernatant obtained by spinning the T C A suspension at 5000 × g was pooled with washings of the resulting pellet. The determinatior~ of radiolabelled free polyamines and protein-bound y-glutamyl derivatives were performed on the acid-soluble fraction and on the enzymatic digest of the aeid-insoh,b!e fraefiun, respectively, using the automated ion-exchange chromatography procedttre previously described [26,31]Partial purification of tmnsgiataminase Livers from rats treated with retinoic acid were perfused in situ with 0.25 M sucrose through the portal vein. The homogenate (25%, w/v), was prepared in 20 mM Tris-HCI buffer (pH 7.5) containing 0.25 M suc r~ e , 1 mM phcnylmethylsulphonyl fluoride (PMSF), 5 mM jS-mercaptoethaunl (thereafter called simply buffer). The whole homogenate was centrifuged at 1 5 0 0 0 x g for 15 rain and, after removing the supernatant, the resulting pellet was washed with buffer and centrifuged again. The supernatants were pooled and centrifuged at 1 0 0 0 0 0 × g for 1 h. The re~.~alting soluble phase was mixed with DE52 resin (approx. 1 g per 100 ~tg of protein) conditioned in buffer. The resin was then packed in a Buchner funnel and rinsed with buffer containing 0.4 M NaC1, and the eluate was concentrated in Amicon cell with YM-10 membrane. The transglutaminase was further purified by chromatograpby on DEAE 5 PW column on FPLC (LKB, Bromma). The final specific activity was about 50-fold

1"/3

TABLE ! Effect of RA oa :issue tmm~lu~aminare protein ~ mt liver and fmdtlv gsalat~ I~mumyte~

Malo Wtstar rats were injeo~xli,p. with olive oil alone (control)or with olive oil mntaining retino¢ acid (20 rag/ks body welshtg 12 h

laler line~ ~ r e per[used with 03-~ M sucrose containing I mM EDTA when li~erhome.hales were to be prepared, or with Hanks' buffer (pH 7.4) containingcogigen~e when hepatoeytcswere to be isolated, tTG protein was assayed as des