Inhibition of “Tissue” Transglutaminase Increases Cell Survival by Preventing Apoptosis*

Treatment of the human promonocytic cell line U937 with all-trans-retinoic acid (RA) commits these cells to apoptosis, which can be triggered by simply increasing intracellular calcium levels by the ionophore A23187. RA treatment of U937 cells is characterized by a decrease in Bcl-2 and marked induction of “tissue” transglutaminase (tTG) gene expression. In this study, we show that the inhibition of tTG expression in U937 cells undergoing apoptosis prevents their death. In fact, U937 cell-derived clones transfected with the human tTG gene in the antisense orientation showed a pronounced decrease in apoptosis induced by several stimuli. These findings demonstrate that the Ca2+-dependent irreversible cross-linking of intracellular proteins catalyzed by tTG represents an important biochemical event in the gene-regulated cell death in monoblasts. In addition, our data indicate that the apoptotic program in promonocytic cells is strictly regulated by RA and that a key role is played by the free intracellular calcium concentration.


Treatment of the human promonocytic cell line U937
with all-trans-retinoic acid (RA) commits these cells to apoptosis, which can be triggered by simply increasing intracellular calcium levels by the ionophore A23187. RA treatment of U937 cells is characterized by a decrease in Bcl-2 and marked induction of "tissue" transglutaminase (tTG) gene expression. In this study, we show that the inhibition of tTG expression in U937 cells undergoing apoptosis prevents their death. In fact, U937 cell-derived clones transfected with the human tTG gene in the antisense orientation showed a pronounced decrease in apoptosis induced by several stimuli. These findings demonstrate that the Ca 2؉ -dependent irreversible cross-linking of intracellular proteins catalyzed by tTG represents an important biochemical event in the gene-regulated cell death in monoblasts. In addition, our data indicate that the apoptotic program in promonocytic cells is strictly regulated by RA and that a key role is played by the free intracellular calcium concentration.
"Tissue" transglutaminase (tTG) 1 catalyzes a Ca 2ϩ -dependent reaction in which ␥-carboxamide groups of peptide-bound glutamine residues serve as acyl donors and primary amino groups of several compounds function as acceptor substrates (1,2). The reaction results in post-translational modification of proteins by establishing ⑀-(␥-glutamyl)lysine cross-linkages and/or covalent incorporation of polyamines and histamine into proteins (2)(3). Di-and polyamines may also participate in cross-linking reactions through the formation of N,N-bis(␥-glutamyl)polyamine cross-bridges (3). tTG-dependent formation of stable cross-linking determines protein polymerization, conferring resistance to mechanical breakage and chemical attack to the polypeptides involved in the linkage; in fact, these polymers can be destroyed only by proteolytic degradation of the protein chains (3,4).
In addition to the Ca 2ϩ -dependent protein cross-linking activity, tTG binds guanine nucleotides and hydrolyzes GTP and ATP. Nakaoka et al. (5) have demonstrated that the 74-kDa ␣-subunit (G␣) associated with the 50-kDa ␤-subunit of the GTP-binding protein G h is tTG. This dimer acts in association with the ␣ 1 -adrenergic receptor in a ternary complex. Thus, tTG is a multifunctional protein that not only acts as a transglutaminase, but activates phospholipase C after receptor stimulation (5).
It has been well established in various experimental systems that tTG is one of the few genes induced during apoptosis (6 -10). tTG protein is undetectable in the majority of cells, and its mRNA is transcribed as a consequence of the onset of apoptosis (6,8). Overexpression of tTG primes cells for suicide, and the clones resistant to tTG transfections are highly susceptible to apoptosis induced by various agents (9 -13). Since the intracellular tTG cross-linking activity is inhibited by GTP and nitric oxide, the accumulation of tTG protein inside the cell is not necessarily associated with the activation of its crosslinking activity (1,14). However, the Ca 2ϩ -dependent activation of the enzyme leads to the formation of a detergentinsoluble cross-linked protein scaffold in cells undergoing programmed cell death (4). This insoluble protein scaffold may stabilize the integrity of the dying cells before their clearance by phagocytosis, thus preventing the nonspecific release of harmful intracellular components (i.e. lysosomal enzymes, nucleic acid, etc.) and consequently inflammatory responses and scar formation in bystander tissues (9,10). Although several laboratories have shown that tTG gene expression characterizes cells undergoing apoptosis both in physiological and experimental settings, the precise position of tTG in the cascade of events leading to establishment of the death phenotype has not yet been fully clarified (4,6,7,14). Several studies indicate that induction of tTG parallels Bcl-2 down-regulation and is not sensitive to its inhibitory effect (12).
To investigate to what extent the overexpression of the tTG gene is a key biochemical event in the death program, we studied the effect on apoptosis of the all-trans-retinoic acid (RA)-dependent induction of tTG in cell lines derived from promonocytic U937 cells stably transfected with plasmids containing tTG cDNA in the antisense orientation.
Cell Culture and Treatments-The U937 parental cell line and its derived transfected sublines (UNeo, UAS1, and UAS2) were grown in RPMI 1640 medium supplemented with 10% (v/v) heat-inactivated fetal calf serum, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin and incubated at 37°C in a humidified atmosphere of 5% CO 2 in air.
In various experiments, cells (seeded at a density of 3 ϫ 10 5 cells/ml) were first exposed or not to all-trans-retinoic acid (1 M from a 5 mM stock solution dissolved in 70% ethanol) and treated as follows. (a) In calphostin C (0.25 M from a 0.25 mM stock solution in Me 2 SO) studies, cells pretreated or not with 1 M RA for 36 h were incubated for an additional 11 h in growth medium containing 0.5% serum in the presence or absence of 1 M RA. The cells were finally treated with calphostin C for 1 h. After calphostin C addition, cells were exposed to light for 10 min, and the incubation was allowed to proceed in the dark for an additional 50 min. (b) A23187 calcium ionophore (1 M from a 1 mM stock solution in Me 2 SO), apoptosis-inducing anti-CD95 mAb (100 ng/ml in culture medium), and cycloheximide (4 g/ml) were added to cells pretreated or not with 1 M RA for 36 h, and incubation was carried out for an additional 12 h. (c) In staurosporine experiments, cells pretreated or not with 1 M RA for 43 h were incubated for the last 5 h in the presence of staurosporine (2 M from a 2 mM stock solution in Me 2 SO). (d) Z-VAD (100 M) was added 5 h before treatment with A23187 (1 M) to cells pretreated or not with RA for 36 h, and the incubation was allowed to proceed for an additional 12 h in the presence of the calcium ionophore. (e) Mycobacterial infections of U937 cells were performed as described previously (15). Briefly, 1 ϫ 10 6 cells were incubated for 24 h with Mycobacterium tuberculosis strain H37Rv in RPMI 1640 medium, 10% FCS, and 2 mM L-glutamine (complete medium); washed; and maintained at 3 ϫ 10 5 cells/ml in complete medium. (f) For time courses, cycloheximide (4 g/ml) and actinomycin D (5 ng/ml) were added, and samples were analyzed after 6, 16, 24, and 48 h by flow cytometry.
Antisense tTG Vector Construction and Expression-Two different antisense tTG-expressing vectors were used in this study. pSG5-AS-tTG carries the first 5Ј 1.0 kilobase pair of the human tTG cDNA (human endothelial cell tTG, clone hTG-1) (12) cloned in the antisense orientation into the EcoRI site of the pSG5 vector (Stratagene) under the control of the SV40 early promoter.
For the pRC2 vector, after PCR amplification with primers carrying EcoRI restriction sites, the entire coding region for human tTG was cloned into the EcoRI site of the pEGFP vector (CLONTECH), and clones with the insert in the antisense orientation were selected. Digestion with NheI and EcoRV restriction endonucleases and subsequent religation of the linearized blunt-end vector gave rise to the 5Ј 1.3kilobase pair tTG cDNA in the antisense orientation under the control the cytomegalovirus promoter. The integrity of the tTG gene was controlled by sequencing.
To assess expression of antisense transcripts, 5 g of total RNA extracted from cells (16) were subjected to reverse transcription-PCR. Reverse transcription (Superscript II, Life Technologies, Inc.) with primer SO2 (5Ј-AATTCTATGGCCGAGGA), specific for antisense tTG RNA, was performed following the supplier's conditions. cDNA PCR amplification with TaqGold DNA polymerase (Perkin-Elmer) and primers SO1 (5Ј-ATGGCCGAGGAGCTGGTCTTAG) and TG1R (5Ј-CGGGGTGGTGAGCTGCAGCG) was carried out as follows: 10-min activation step at 94°C, followed by 25 cycles of denaturation for 30 s at 94°C, annealing for 30 s at 65°C, and extension for 30 s at 72°C using actin as an internal control. PCR products were visualized on 2% agarose alongside with DNA molecular mass marker V.
Transfections and Single Clone Selection-Antisense UAS1 and UAS2 clones were established by cotransfecting U937 cells with pSG5-AS-tTG and pSV2-Neo. DNA transfections were performed using Lipofectin accordingly to the supplier's conditions. About 10 7 cells on a 60-mm plate were transfected with pSG5-AS-tTG (10 g) and pSV2-Neo (1 g) or pSV2-Neo (10 g) in the presence of Lipofectin (70 g). For pRC2 antisense-expressing clones, 2 ϫ 10 6 cells (seeded on 6-wells plates) were transfected with 2 g of DNA using 1,2-dimyristyloxypro-pyl-3-dimethyl-hydroxyethyl amomonium bromide and cholesterol accordingly to the supplier's instructions. In all cases, after 5 h of transfection in serum-free RPMI 1640 medium, cells were shifted to complete medium. 48 h after transfection, cells (5 ϫ 10 3 /well) were plated onto 96-well plates in RPMI 1640 medium and 10% FCS supplemented with G418 (500 g/ml). After 1 month of G418 selection, resistant cells were reseeded at limiting dilution to achieve single clones.
Tissue Transglutaminase Activity Assay-Cells treated with 1 M all-trans-retinoic acid were washed in phosphate-buffered saline (PBS) without Ca 2ϩ and Mg 2ϩ ; resuspended in 50 mM Tris-HCl, pH 8.4, and 1 mM Na 2 EDTA; and sonicated at 4°C for 20 s. tTG activity was measured by detecting the incorporation of [ 3 H]putrescine into N,NЈ-dimethylcasein as previously reported (6).
Apoptosis Detection-Apoptosis evaluation at different days after treatment and/or infections was carried out using flow cytometric analysis. For propidium iodide staining, aliquots of 5 ϫ 10 5 cells were washed twice in PBS; placed on ice; gently resuspended with 0.5 ml of solution containing 50 g/ml propidium iodide, 0.1% Triton X-100, and 0.1% sodium citrate; and left at 4°C for 30 min in the dark before analysis. For detection of phosphatidylserine exposure on cell membranes, FITC-conjugated annexin V was added at a final concentration of 2 g/ml to 2.5 ϫ 10 5 cells. After washing twice in PBS, cells were resuspended in binding buffer (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, and 2.5 mM CaCl 2 ) and incubated with FITC-conjugated annexin V for 10 min in the dark at room temperature. Cells were washed twice and resuspended in binding buffer containing propidium iodide (1 g/ml final concentration) to exclude necrotic cells accordingly to the method suggested by the supplier. For each sample, 5 ϫ 10 3 viable cells were gated following size (forward scatter; FCS) and granularity (side scat- Cytofluorometric Detection of tTG-5 ϫ 10 5 U937 and RA-treated cells were washed in PBS, pH 7.2, supplemented with 1% (w/v) bovine serum albumin and 0.1% NaN 3 ; fixed in 4% p-formaldehyde for 20 min at room temperature; washed in PBS; and then incubated for 15 min in 0.5 M NH 4 Cl at room temperature. After a final wash in PBS, cells were permeabilized by incubation in PBS/bovine serum albumin/NaN 3 containing 0.05% (w/v) saponin detergent for 15 min at room temperature. Cells were then stained intracellularly with anti-tTG mAb in saponin detergent buffer for 30 min at room temperature. After washes with saponin detergent buffer, cells were incubated with FITC-conjugated goat anti-mouse Ig for 30 min, washed and fixed with 1% formaldehyde in PBS, applied to a FACSCalibur flow cytometer, and analyzed as described above. Western blotting was carried out as described previously (12) on aliquots of total protein (20 -50 g) extracted from cells after the different treatments.

Effect of RA on Apoptosis and tTG Expression in U937
Cells-RA treatment has drastic effects on the cell cycle, blocking the U937 cells in the G 1 phase and triggering differentiation (17,18). In parallel with these events, RA treatment induced an early and dramatic accumulation of tTG (Fig. 1), which was paralleled by a decrease in Bcl-2 protein levels (see Fig. 3D) (19). The data reported in Fig. 2 show that the RA treatment of U937 cells is associated with their commitment to apoptosis. In fact, the RA-treated cells, which express high tTG protein levels, were particularly susceptible to apoptosis induced by different stimuli (calcium ionophore A23187, calphostin C, staurosporine, and cycloheximide), whereas untreated cells were much more resistant to death ( Fig. 2A). Although U937 cells express high levels of CD95 receptor protein, which is not affected by the RA treatment (data not shown), these cells were unable to undergo apoptosis when treated with agonistic anti-CD95 IgGs (Fig. 2A). Interestingly, pretreatment of U937 cells with RA committed these cells to apoptosis, which could be triggered by increasing their intracellular calcium levels with the A23187 ionophore (Fig. 2B). The ionophore-induced apoptosis of RA-pretreated cells was mediated by caspases, as indicated by the death inhibitory effect displayed by Z-VAD (Fig. 2C) and the parallel proteolytic cleavage of poly(ADPribose) polymerase (data not shown). Staining with FITC-conjugated annexin V revealed that the dying U937 cells exposed PS molecules on the outer leaflet of the plasma membrane (Fig.  2B). The externalization of PS was not prevented by Z-VAD (Fig. 2D), confirming that this event in human leukemic U937 cells undergoing apoptosis is independent from caspase activation (20). Taken together, these data show that, in U937 cells expressing high tTG levels, an increase in the intracellular free calcium levels is able to induce a complete apoptotic program, including the surface changes related to the clearance of the dead cells.
Isolation and Characterization of U937 Cell Lines Containing Antisense tTG Transcripts-To investigate whether the priming effect brought about by RA on U937 cells was dependent on tTG, we attempted to block the enzyme expression by an antisense-based strategy. U937 cells were transfected with two different vectors expressing an antisense RNA complementary to the 5Ј-portion of the human tTG gene as described under "Experimental Procedures." A large number of clones were collected from U937 cells stably transfected with pSG5-AS-tTG or pRC2 and from cells transfected only with the control plasmids pSV2-Neo (UNeo) and pEGFP. The individual cell lines obtained were screened for the expression of the antisense tTG RNA by reverse transcription-PCR as described under "Experimental Procedures." As shown in Fig. 3A, the selected clones (UAS1 and UAS2) expressed high levels of antisense RNA, which were comparable to the actin mRNA levels.
The effect of the antisense construct on tTG gene expression was verified by measuring protein levels and enzyme activity before and after RA treatment (Fig. 3, B and C, respectively). Whereas the U937 cell line and the clones transfected with pSV2-Neo showed the typical RA-dependent increase in tTG expression and activity, cell lines cotransfected with pSG5-AS-tTG and pSV2-Neo showed a drastic reduction in both the tTG protein levels and enzyme activity after induction by RA (Fig.  3), thus indicating the efficient inhibition achieved by the antisense strategy in the U937 cell line. To verify that the inhibitory effect on tTG expression was indeed due to the specific action of the antisense construct and not to clonal selection, we studied the effect of RA on Bcl-2 protein levels in U937 cells and its derived transfectants. As reported in Fig. 3D, U937 cells and pSV2-Neo-transfected and pSG5-AS-tTG/pSV2-Neo-cotransfected clones showed comparable basal Bcl-2 protein levels, which were significantly decreased by RA treatment, thus confirming the specificity of tTG inhibition observed in the antisense transfectants. Similar results were obtained with the pRC2 vector (data not shown).
To further assess the specificity of our antisense construct, we treated the U937 parental cell line and its derived transfected clones with agents capable of inducing cell death by nonspecifically blocking transcription and transduction. Fig. 4 shows that treatment of cells with cycloheximide (4 g/ml) or actinomycin D (5 ng/ml) induces a marked cell death, which was comparable in both the U937 parental cell line and its derived clones. Higher doses of cycloheximide (50 g/ml) or actinomycin D (50 ng/ml) induce a significant level of cell death (ϳ60 -80%) already after 24 h of treatment in all cells (data not shown). These data demonstrate that the antisense tTG construct did not interfere with cellular processes, including cell death, that do not require gene expression and protein synthesis.
To determine whether the priming effect shown in Fig. 2 was dependent on tTG expression, we treated the antisense tTG transfectants with the same apoptosis-inducing agents. The data indicate that the reduced tTG expression observed in the RA-pretreated antisense tTG clones is associated with a net decrease in the sensitivity to various apoptotic stimuli. In fact, the antisense transfectants were much less prone to apoptosis induction elicited by calphostin C, calcium ionophore, and cycloheximide (Fig. 5, A-B). It is interesting to note that the tTG inhibition was less effective in preventing apoptosis induced by cycloheximide than that elicited by the Ca 2ϩ ionophore A23187 (Fig. 5A). This phenomenon can be explained by assuming that a continuous tTG synthesis is needed for apoptosis in U937 cells and/or that the induction of additional pro-apoptotic proteins is required to complete the death program.
M. tuberculosis replicates in cells of the M/M lineage (including U937), where, under some circumstances, it induces apoptosis and tTG expression (15). Consistent with these observations, macrophages obtained by bronchoalveolar lavage from patients with reactive pulmonary tuberculosis and from AIDS patients with disseminated pulmonary tuberculosis show increased levels tTG protein, which is mainly expressed in cells undergoing apoptosis (7,15). Since tTG has been shown to modulate the release of M. tuberculosis and intracellular components from the infected macrophages into the medium, apoptosis of the host cells may represent a very efficient means to prevent spreading of M. tuberculosis infection (21,22). Based on these observations, we investigated the effect of tTG inhibition on the ability of M. tuberculosis to induce apoptosis in U937 cells. The induction of apoptosis by M. tuberculosis requires viable bacteria, is dose-dependent, and is restricted to H37Rv (15). The data reported in Fig. 5C indicate that the inhibition of tTG reduce M. tuberculosis-dependent cell death in U937 cells, thus confirming that tTG is an important effector element of apoptosis in M. tuberculosis-infected M/M cells.
The data reported in this paper indicate the tTG-mediated protein cross-linking is a critical event in the apoptotic path- FIG. 5. Inhibition of apoptosis in U937 cell-derived antisense tTG clones. A, inhibition of apoptosis induced by A23187 and cycloheximide in RA-pretreated control cells (UNeo) and antisense tTG clones (UAS1 and UAS2). Cells were pretreated for 36 h with 1 M RA (black bars) and subsequently treated overnight with 1 M calcium ionophore A23187 (dark gray bars) or with 4 g/ml cycloheximide (light gray bars). After treatment, cells were collected and stained with propidium iodide, and apoptosis was analyzed by flow cytometry. B, inhibition of apoptosis induced by calphostin C in antisense tTG transfectants. Control pSV2-Neo-transfected cells (UNeo) and pSG5-AS-tTG antisense transfectants (U937 cell-derived clones UAS1 and UAS2) were maintained for 36 h in normal medium and then shifted to 0.5% serum-containing medium for overnight exposure (12 h) and incubated (gray bars) or not (white bars) with 0.25 M calphostin C as described under "Experimental Procedures." Apoptosis was evaluated as described above by flow cytometric analysis. C, U937 cell-derived antisense tTG transfectants are resistant to M. tuberculosis-induced apoptosis. The untransfected U937 cell line and transfected clones (UNeo, UAS1, and UAS2) were infected with M. tuberculosis strain H37Rv; and at the different time intervals indicated, apoptosis was detected by flow cytometry using the propidium iodide staining method. Ⅺ, U937; f, UNeo; ‚, UAS1; OE, UAS2. Data are the means Ϯ S.E. of triplicate determinations from three different experiments. way in U937 cells. The transfection studies carried out on U937 cells clearly show that the blocking of tTG expression in these cells significantly inhibits apoptosis, thus suggesting that tTG is one of the genes responsible for the commitment of these promonocytic cells to undergo cell death by apoptosis. Furthermore, these data indicate that the apoptotic program in these mesoderm-derived cells is strictly regulated by RA and that a key role is played by the free intracellular calcium concentration. It is well known that both tTG activation and the onset of the apoptotic program require sustained high Ca 2ϩ levels (1). Thus, it is plausible to hypothesize that, at the low level of free cytosolic calcium normally present in viable RA-treated U937 cells, the accumulated tTG protein might be in its G-protein configuration, which inhibits the cross-linking activity required for apoptosis (18,23).
In conclusion, we have presented data demonstrating that the tTG gene product is one of the effector elements of the apoptotic program in promonocytic cells. tTG activation, leading to the assembly of intracellular cross-linked protein polymers, may irreversibly modify cell organization, contributing to the determination of those ultrastructural changes typical of cells undergoing apoptosis.