THE JOURNAL OF BIOLOGICAL CHEMISTRY (0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No. 11. Issue of April 15, pp. 7108-7113, 1991 Printed in U. S A .

Differential Regulation of Lymphotoxin and Tumor Necrosis Factor Genes in Human T Lymphocytes*

(Received for publication, November 7, 1990)

B. Keith English$, William M. Weaver, and Christopher B. Wilson From the Departments of Pediatrics and Immunology, University of Washington and the Divisions of Immunology/ Rheumatology and Infectious Disease, Children’s Hospital and Medical Center, Seattle, Washington 98105

Lymphotoxin (LT) and tumor necrosis factor (TNF) are related cytokines that share many biological ef- fects. The genes for LT and TNF are adjacent to each other on chromosome 6 in man, but previous data in- dicate that the kinetics of their production differ mark- edly. To explain the mechanisms for this difference, we compared the regulation of these two genes in hu- man T lymphocytes, isolated from peripheral blood, after stimulation with the mitogens concanavalin A and phorbol myristate acetate. Differences in the ki- netics of protein secretion were paralleled by differ- ences in cognate mRNA accumulation. TNF mRNA accumulated rapidly after stimulation, peaked by 6 h, and returned to unstimulated (base-line) levels by 24 h. In contrast, LT mRNA accumulated slowly after stimulation, usually peaked at -18 h, and remained increased above base-line levels at 48-72 h. By nuclear transcription run-on assays, increased transcription of TNF mRNA and LT mRNA was demonstrated after stimulation. However, TNF transcription peaked ear- lier and appeared to be 4-10 times greater than that of the LT gene. In contrast, the half-life of LT mRNA was 8-10-fold longer than that of TNF mRNA as dem- onstrated by actinomycin D pulse-chase experiments. Cycloheximide did not block LT or TNF mRNA accu- mulation, indicating that new protein synthesis was not required for induction of either gene.

These results suggest strongly that the LT and TNF genes are regulated differently in human T lympho- cytes after mitogen stimulation. TNF mRNA accumu- lates rapidly primarily because of increased transcrip- tion and decreases rapidly related to its brief half-life. In contrast, LT mRNA accumulates more slowly but persists much longer; the accumulation of this mRNA appears to be controlled largely by post-transcriptional mechanisms.

Lymphotoxin (LT)’ and tumor necrosis factor (TNF) are

* This work was supported in part by National Institutes of Health Grants HD18184, AI16760, and HL39157. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Supported by National Research Service Award HD06948. T o whom correspondence should he sent: Dept. of Biochemistry, St. Jude Children’s Research Hospital, 332 N. Lauderdale, Memphis, T N 38105.

’ The abbreviations used are: LT, lymphotoxin; TNF, tumor ne- crosis factor: MC, mononuclear cells; ConA, concanavalin A; PMA, phorbol myristate acetate; IL, interleukin; IFN, interferon; PHA, phytohemagglutinin; GM-CSF, granulocyte-macrophage colony- stimulating factor; Hepes, 4-(2-hydroxyethyl)-l-piperazineetbanesul- fonic acid SDS, sodium dodecyl sulfate.

cytokines with partial molecular homology which bind to the same or similar receptors and mediate multiple common bio- logical effects (1, 2). The human LT and TNF genes are closely linked and are located adjacent to the human major histocompatibility complex on chromosome 6 (3,4). Although originally identified as factors with cytotoxic or cytostatic activity against a variety of tumor cell lines (5-7), the pleio- tropic nature of these mediators has become increasingly apparent (1,2). LT and TNF are involved in multiple aspects of the host inflammatory response including the induction of fever and the hepatic synthesis of acute phase reactants (8, 9). These cytokines inhibit the replication of a variety of DNA and RNA viruses and act synergistically with interferon-y (10, 11). Both LT and TNF act as regulators of the growth and differentiation of myelomonocytic cells (12) and activate mature myeloid cells (13, 14). These cytokines also play important roles in the growth and activation of lymphocytes (15-17). In addition to these physiological roles, overpro- duction of TNF has been implicated as an important mediator of pathophysiologic states including septic shock (18), graft uersus host disease (19), and the cachexia associated with cancer (20). There are less data regarding the role of LT in pathophysiologic states, but some evidence suggests that it may be less toxic (21, 22).

The potential for deleterious as well as beneficial effects indicate the need to regulate the production of these mediators closely. LT is produced primarily by activated T lymphocytes (23, 24). Although TNF was characterized initially as a prod- uct of monocyte/macrophages (25-27), T lymphocytes pro- duce large amounts of TNF under some conditions (28-31). Natural killer cells also produce TNF (32, 33) and may produce LT (34). B cells are also a source of TNF (35). In a previous study comparing the ability of adult and neonatal blood mononuclear cells and T cells to produce LT and TNF, we noted a dramatic difference in the kinetics of the accu- mulation of these two cytokines (30). In both mononuclear cell and T cell preparations, TNF mRNA and protein accu- mulated much earlier than did LT mRNA and protein under all conditions studied.

To gain insight into the mechanisms responsible for this apparent difference in LT and TNF gene regulation, in the present study we examined the molecular basis of the differ- ences in LT and TNF gene expression in human T lympho- cytes. We report here that the earlier accumulation of TNF mRNA after T cell activation appears to result primarily from increased transcription of the TNF gene whereas the delayed accumulation but much longer persistence of LT mRNA results primarily from a greatly prolonged half-life of this cytokine message.

EXPERIMENTAL PROCEDURES

and streptomycin were purchased from GIBCO. RPMI 1640 contain- Materials-Hanks’ buffered salt solution, L-glutamine, penicillin,

7108

Regulation of LT and TNF Genes in Human T Lymphocytes 7109

ing 25 mM Hepes buffer was obtained from Cellgro, Dulles Interna- tional Airport, Washington, D. C., and contained less than 0.3 enzyme units of lipopolysaccharide/ml by limulus amebocyte lysate assay (Pyrotell Associates of Cape Cod, Inc., Woods Hole, MA). Ficoll- Hypaque and concanavalin A (ConA) were purchased from Pharma- cia LKB Biotechnology Inc. 12-0-Tetradecanoylphorbol-13-acetate and cycloheximide were obtained from Sigma. Actinomycin D was obtained from Calbiochem.

Cell Preparations-Mononuclear cells (MC) were isolated from peripheral blood of healthy adult donors by Ficoll-Hypaque density gradient centrifugation as described previously (36), washed twice in Hanks' balanced salt solution, resuspended in RPMI 1640 containing 2 mM L-glutamine, 50 units/ml penicillin G , 50 pg/ml streptomycin, and 5% human AB serum. Purified T lymphocytes were prepared by treatment of MC with T cell Lymphokwik as specified by the manu- facturer (One Lambda, Los Angeles, CA). The compositions of the MC and T cell preparations were assessed by indirect immunofluo- rescence microscopy and/or immunofluorescent flow cytometry after staining with appropriate monoclonal antibodies as described previ- ously (30). T cell preparations contained less than 1% granulocytes, B lymphocytes, and monocytes; 5-10771 of these cells stained positive for the CD16 marker of natural killer cells.

Cell Culture-Blood MC and T cells were cultured in conical 15- ml polypropylene tubes a t 5 X 106/ml and stimulated with 25 pg/ml ConA plus 50 ng/ml PMA. (Preliminary titration experiments indi- cated that these concentrations of ConA and PMA resulted in optimal incorporation of ["Hlthymidine and maximal accumulation of cyto- kine mRNAs by our T cell preparations.) For pulse-chase experi- ments, actinomycin D was added at the indicated times a t a final concentration of 5 pg/ml. In experiments in which cycloheximide was used, it was added to a final concentration of 20 pglml. After the indicated time, supernatants were collected, frozen at -70 "C, and later assayed for LT and TNF.

RNA Isolation and Northern Blot Analysis-Total cellular RNA was isolated from MC or T cells by the guanidinium isothiocyanate/ cesium chloride method and quantitated spectrophotometrically as described previously (37). For blots, 10 pg of RNA was electrophoresed in 2.2 M formaldehyde, 1% agarose gels, transferred to Nytran (Schleicher & Schuell), UV irradiated, and baked at 80 "C for 90 min as described (38). Blots were hybridized with "'P-labeled RNA probes transcribed from the following subclones in pGEM vectors (Promega Biotec, Madison, WI): LT, the 940-base pair EcoRI cDNA fragment (24); TNF, the 800-base pair EcoRI cDNA fragment (27). The specific activity of the probes for LT and TNF ranged from 1.1 to 1.3 X loR cpm and 1.2 to 1.4 X 10' cpm/pg of transcribed cDNA insert, respec- tively. These cDNAs,were originally provided by P. Gray (Genentech, Inc.). The human elongation factor la cDNA probe (provided by R. M. Perlmutter, University of Washington) was ?'P labeled by the random hexamer primer method (39). After hybridization blots were washed at 63 "C with 6 X SSC, 0.1% SDS for 30 min and then with 0.1 X SSC, 0.1% SDS for 30 min and then autoradiographed a t -80 "C. Previously probed Nytran filters were stripped by boiling in 20 mM Tris, pH 8.5, 2 mM EDTA, 0.1% SDS for 15 min, dried, and then reprobed. Filters were stained with methylene blue to detect the ribosomal RNA bands as described (40).

Nuclear Run-on Transcription Assays-Cells were lysed in reticu- locyte standard buffer (0.01 M Tris-HCI, pH 7.4, 0.01 M NaCI, 3 mM MgCI,) containing 0.5% Nonidet P-40 by Dounce homogenization. Nuclei were recovered by centrifugation, enumerated, and then stored in freezing buffer (40% glycerol, 50 mM Tris-HCI, pH 8.3, 5 mM MgC12,O.l mM EDTA) at -80 "C until analysis. Run-on assays were performed exactly as described by Nelson and Groudine (41) by using 2.5 X 10' nuclei/reaction. '"P-Radiolabeled run-on RNA was hybrid- ized to the denatured plasmids described above (or, as controls, to the vector in which they were contained) which had been transferred to Nytran with a slot-blot apparatus (Schleicher & Schuell).

Densitometry-The relative intensity of bands on Northern and run-on blot autoradiographs was determined using a Visage 60 den- sitometer with Sun Microsystems computer (BioImage, Ann Arbor, MI). Although only a single exposure is shown in the figures, multiple exposures were obtained to ensure that analysis was performed in the linear range of the film. When comparisons were made between band intensities a t different times, values for unstimulated cells were used as the base-line values. For the nuclear run-on transcription assays the values obtained by scanning the slots for the pGEM vector (negative control) were used as the background values.

RESULTS

Kinetics of Accumulation of LT mRNA and TNF mRNA in Human MC and T Lymphocytes-We have reported previ- ously that the kinetics of secretion of LT and TNF protein by mitogen-stimulated human blood MC or T cells parallel the kinetics of accumulation of the cognate mRNAs for these two cytokines (30). Fig. 1, a representative Northern blot, illustrates the strikingly different kinetics of accumulation of LT and TNF mRNAs in blood MC or T cells stimulated with ConA + PMA; similar results were obtained when ConA was used alone as a stimulus (data not shown). L T mRNA was not detectable in unstimulated MC or T cells. After stimula- tion, L T mRNA was detected by 6 h but did not peak until 18-24 h; the message remained detectable as late as 72-96 h after stimulation (data not shown). Unlike LT mRNA, T N F mRNA was frequently detectable in freshly isolated, unstim- ulated blood MC and to a lesser degree in unstimulated T cell preparations. Consistent with this, small amounts of T N F protein were often detectable in supernatants of freshly iso- lated, unstimulated MC preparations (by enzyme-linked im- munosorbent assay, range 45-169 pg/ml); in T cell superna- tants TNF protein was either undetectable (<25 pg/ml, seven subjects) or near the limit of detection (29 and 34 pg/ml, respectively, in two subjects). After stimulation, T N F mRNA accumulated rapidly, usually peaked at 6 h, and returned to base-line levels by 24 h (Fig. 1); the kinetics were similar in MC and T cell preparations under these conditions. In other experiments, the kinetics varied somewhat. Nevertheless, ac- cumulation of T N F mRNA always preceded that of L T mRNA. Increased T N F mRNA was detected routinely by 1- 2 h and peaked by 2-6 h after stimulation whereas an increase in LT mRNA usually was not detected until 4-6 h and did not peak until 8-18 h after stimulation.

Transcription Rates of the L T and T N F Genes in Activated Human T Cells-The production of most T cell lymphokines, including IL-2, IFN-7, and IL-4, is controlled primarily a t the level of transcription (42-44). We employed nuclear run-on transcription assays to compare the transcription rates of the LT and TNF genes in activated human T cells. IFN-7 was used as an additional comparison lymphokine gene; the pGEMl vector DNA served as a negative control (back- ground). Fig. 2 shows a representative run-on experiment. In unstimulated T cells, TNF and IFN-7 transcripts were de- tected just above background whereas LT transcripts were

28s m L LT p - 1 -" - -5 I *-* TNF ,e- Ttme(hr) " J

. . . d 0 6 12 18 24

Con A + PMA Con A + PMA

Mononuclear Cells T Cells

FIG. 1. Northern blot analysis of the kinetics of accumula- tion of LT mRNA and TNF mRNA in human mononuclear cells and T lymphocytes. Cells were cultured a t 5 X 10"/ml and stimulated with ConA + PMA as described under "Experimental Procedures." RNA was isolated from the cells at the indicated times. In each lane 10 p g of total RNA was electrophoresed. Preparation of the :v2P-labeled probes for LT and TNF mRNA is described under "Experimental Procedures." T o demonstrate similar loading of RNA/ lane, filters were stained with methylene blue; the 28 S ribosomal RNA bands are shown.

7110 Regulation of LT and TNF Genes in Human T Lymphocytes

IFN 7 (baselhe) e M7.3) 1, . (6.6) , - (1.3) * (2.0)

LT (baseline) ~ 6 . 7 ) - (7.3) (3.8) (5.4)

TNF . (baseline) N11.8) I, (8-0) (4.1) (6.9)

C t r l (baselhe)

_ . ,,...e

T'"= 1 1.5 I (hr) 1.5 6 12 I8

Unstlnulated Con A + PHA

FIG. 2. Transcription rates of the LT, TNF, and IFN-7 genes in nuclei from activated human T lymphocytes. T cells were incubated with ConA + PMA for the indicated times. Nuclei were isolated and run-on transcription assays performed as described under "Experimental Procedures." Each filter was hybridized with 3 X 10" dpm/ml :"P-laheled transcripts, washed, and autoradiographed. The intensity of the autoradiographed bands was quantitated using a Visage 60 densitometer. The control plasmid used in these experiments was pGEM1, the vector in which the LT cDNA was contained. The numbers in parentheses represent the fold increase in the integrated intensity of the hand as compared with the base-line value for each lymphokine message, which was arbitrarily assigned a value of 1.

not detectable. After stimulation, increased transcription of the TNF, IFN-7, and LT genes was demonstrated. As shown here, TNF transcription peaked at 1.5 h and then declined whereas the transcription of IFN-7 and of LT was increased to a similar extent a t 1.5 and 6 h and declined thereafter. In other experiments T N F transcription increased as early as 30 min and peaked as early as 60 min after stimulation, preceding that of IFN-7 and of LT, transcription of which was not detectably increased before 1 h after stimulation (data not shown). Both TNF and LT gene transcription remained de- tectable severalfold above unstimulated (base-line) values for each gene as late as 18 h after stimulation.

Of note, although the TNF and IFN-7 signals exceeded the negative control (background) by 15-20-fold, the maximum LT signals were only 2-4-fold above background. Thus, at any time point the intensity of the TNF signal was 5-8 times greater than that of the LT signal, suggesting that the TNF gene was consistently transcribed at a higher rate than the LT gene in T cells under these conditions. Although it is possible that the binding affinities of these two cytokine messages to their cDNAs differ, this is an unlikely explanation for these run-on results. By Northern blotting using as probes the same cDNA regions that were used to detect run-on transcripts, with the probes labeled to similar specific activi- ties and with similar exposure times to develop autoradi- ographs, LT mRNA abundance appeared to be similar to TNF mRNA abundance by 8-12 h and to exceed T N F mRNA abundance by 12-18 h. Additionally, in four other run-on experiments similar results were obtained; the apparent rate of transcription of the TNF gene in activated T cells was always greater than that of the LT gene.

Determination of L T mRNA and TNF mRNA Half-lives in Activated Human T Cells-Since the striking differences in the kinetics of LT and TNF mRNA accumulation in activated T cells did not appear to result simply from differences in the transcription rates of these genes, we employed actinomycin D in a series of pulse-chase experiments to estimate the half- lives of these two lymphokines messages. Most cytokine mes- sages exhibit a very short half-life (often less than 30 min) in activated lymphocytes or monocytes (42). However, our initial experiments in blood MC stimulated with ConA + PMA suggested that LT mRNA had a much longer half-life than T N F mRNA under these conditions (Fig. 3A). Because T N F is also produced by monocytes after stimuli such as PMA, we next studied purified T cell preparations containing less than 1% monocytes (Fig. 3B). We found that the differences in LT mRNA and TNF mRNA stability were comparable in MC and T cell preparations after stimulation with ConA + PMA. Although there was subject-to-subject variability, in all cases

A

LT

TNF

Time(hr)

B EF

LT

TNF

Tlme(hr)

Con A + PMA Ohr Actinomycin D 6hr

1818.519 22 2 4 - Con A + PMA Ohr

Actinomycin D 18hr

w-

< , ,_

L'I Con A PMA Ohr Actinomycin D 6hr

[ I ' '. 1 < A :, 1 ' .:

Con A PMA Ohr Actinomycin D 18hr

FIG. 3. Determination of LT and TNF mRNA stability in T lymphocytes by actinomycin D pulse-chase experiments. Mononuclear cells ( A ) or T cells ( R ) were cultured a t 5 X 10'/ml in the presence of ConA + PMA. The cells were exposed to 5 &ml actinomycin D, and RNA was isolated from the cells at the indicated times. In each lane 10 pg of total RNA was electrophoresed. The blots were hybridized with the '"P-labeled probes for LT and TNF as described under "Experimental Procedures." T o demonstrate similar loading of RNA/lane, filters were stained with methylene blue (A, not shown) or hybridized with a probe for elongation factor In ( E F ) ( B ) .

( n = 5), the half-life of the LT message was at least 4-fold longer than that of the TNF message. The mean half-lives (+ S.D.) of the messages in activated T cells were as follows: L T mRNA, 334 min (+29); T N F mRNA, 42 min (+13).

Effect of Cycloheximide on L T mRNA and TNF mRNA Accumulation in Activated Human T Cells-New protein syn- thesis may not be required for the transcription of certain T cell lymphokine genes, such as IL-2 (42, 45), although this conclusion is disputed by some (46,47). Because of the lag in accumulation of LT mRNA after T cell activation (compared with TNF as well as IL-2 or IFN-7) (37), we studied the effect of cycloheximide on induction of LT and TNF messages by ConA + PMA. Fig. 4 shows the results of one cycloheximide experiment. Cycloheximide added 2 h before ConA + PMA

Regulation of LT and TNF Genes in H u m a n T Lymphocytes 7111

Time(hr)

C h e x ( h r ) - 0 0 - 2 4

C+P(hr) " 2 0 0 0

. _ ) . j . I

L <-

FIG. 4. Effect of cycloheximide on the accumulation of LT and TNF mRNA in human T lymphocytes. T cells were prepared and cultured a t 5 X 10"/ml as described under "Experimental Proce- dures." The cells were exposed to ConA + PMA (C+P), cycloheximide (Chew), or cycloheximide and ConA + PMA at the indicated times. RNA was isolated from the cells after 12 h. In each lane 10 pg of total RNA was electrophoresed. The blot was hybridized with :"P-labeled probes for LT and TNF as described under "Experimental Proce- dures." To demonstrate similar loading of RNA/lane, filters were stained with methylene blue; the 28 S ribosomal RNA bands are shown.

did not block the induction of either T N F or LT mRNA. Cycloheximide added 2 or 4 h after ConA + PMA resulted in the increased accumulation of both messages. In this experi- ment, cycloheximide alone induced a small amount of L T mRNA but not T N F mRNA. In other experiments ( n = 5), the effect of cycloheximide alone was variable; in three of five experiments a small amount of LT mRNA was induced, and in two of five experiments a small amount of T N F mRNA was induced. However, in all experiments ( n = 4) cyclohexi- mide added before or concomitant with ConA + PMA failed to block the induction of either cytokine message. Similarly, in all experiments ( n = 4) cycloheximide added 2-4 h after ConA + PMA resulted in a greater accumulation of both messages (superinduction). Cycloheximide prolonged the half- life of both LT and TNF messages when added just prior to actinomycin D in pulse-chase experiments (data not shown).

DISCUSSION

LT and TNF are pleiotropic cytokines that bind to the same or similar receptors and appear to mediate similar biological activities with some important exceptions (21, 22, 48). Although most of the interest in TNF has focused on its role as a product of activated monocytes and macrophages, T N F is produced in large quantities by activated T cells (28, 30). In a previous report we noted that the kinetics of accu- mulation of LT and TNF mRNA protein in activated T cells were quite different (30). In this study, we have examined the mechanisms that result in these differences. We found that the increased accumulation of T N F mRNA in activated T cells is regulated primarily at the level of transcription, ac- counting for its rapid accumulation after T cell activation. The relatively short mean half-life (42 min) of T N F mRNA in our studies is similar to that of other, early expressed T cell lymphokines (e.g. IL-2 and GM-CSF) (42) and appears to account for the relatively rapid decline of T N F message after an early peak at -6 h since transcription remained increased as late as 18 h. Conversely, although the transcrip- tion rate of the LT gene increased modestly after T cell activation, the kinetics and magnitude of the accumulation of LT mRNA suggest strongly that this cytokine message is regulated primarily at the level of message stability. The apparent LT mRNA half-life of approximately 5.5 h in our experiments was approximately &fold longer than that of T N F mRNA under the same conditions. Although the peak

accumulation of LT mRNA was much later than that of the message for TNF, new protein synthesis was not required either for LT or T N F gene expression as demonstrated by cycloheximide blocking experiments.

T N F production by activated monocytes/macrophages is known to be regulated at multiple steps. After stimulation of resting monocytes with lipopolysaccharide, transcription of the TNF gene increases 3-8-fold, accumulation of TNF mRNA increases as much as 50-100-fold, and secretion of T N F protein increases as much as 10,000-fold (2,49-52). Our data indicate that transcription of the TNF gene in T cells increases as much as 10-12-fold after activation with ConA + PMA and that TNF mRNA accumulates at least 40-50- fold above unstimulated (base-line) levels in T cells under these conditions. We have shown previously that T cells stimulated with ConA + PMA secrete as much as 10,000- 20,000 pg/ml T N F protein whereas the protein is usually undetectable (e25 pg/ml) in unstimulated T cells (30). Al- though the signals that trigger TNF production in T cells and monocytes are different (28), these data suggest that the control of T N F gene expression in T cells is similar to that in monocytes, but transcriptional regulation may be some- what more important in T cells. The pattern of T N F gene expression in T cells is similar to that of other "early" T cell activation products such as IL-2 and IFN--, (42,46).

LT production by T cells varies with the stimuli employed; in a previous study (30) we found that the combination of ConA + PMA stimulated maximal production of LT mRNA and protein from MC or T cell preparations, so this combi- nation was used in the present study. Our data indicate that the relatively slow accumulation but long persistence of LT mRNA in T cells after activation result from the unusually long half-life of the LT message. Cuturi et al. (28) also reported slow accumulation of LT mRNA after stimulation of T cells with PHA; L T mRNA was detected a t 24 h but not a t 6 h after stimulation. The same authors detected little production of L T mRNA by T cells after stimulation with an analog of PMA (phorbol 12,13-dibutyrate) plus a calcium ionophore but only reported data 6 h after stimulation. In our hands, stimulation of T cells with PMA plus calcium iono- phore results in the production of large amounts of LT mRNA, but the kinetics again lag behind those of T N F mRNA accumulation (data not shown). Turner and Feldmann (53) reported similar kinetics of LT and TNF mRNA accumula- tion after stimulation of MC preparations with PHA and PMA. In their study, TNF and LT mRNA accumulation peaked at approximately 8 h after stimulation and remained detectable as long as 48 h after stimulation. These authors performed actinomycin D pulse-chase experiments in MC stimulated with PHA + PMA and estimated that the half- lives of both LT and TNF mRNA under these conditions were less than 30 min. Our data differ sharply and may reflect differences in the cell preparations as well as the different stimuli employed. We have consistently noted different ki- netics of TNF and LT mRNA accumulation both in T cell and in MC preparations after stimulation with either ConA or ConA + PMA or calcium ionophore + PMA and have found similar differences in LT and TNF message stability in T cells and MC after stimulation with ConA + PMA. Other investigators also have noted delayed accumulation of LT and L T mRNA after stimulation of T cells with staphylococcal enterotoxin A (54), phorbol ester plus calcium ionophore (28), and phorbol ester with or without antibody to CD3 (55). Thus, the pattern of accumulation of these two cytokine messages that we observed after stimulation of T cells appears to be similar to that noted by most other investigators employing a

7112 Regulation of LT and TNF Genes in H u m a n T Lymphocytes

variety of stimuli although it contrasts with the results ob- tained by Turner and Feldmann (53).

We are unaware of any other studies of LT and TNF gene transcription in human T lymphocytes. In preliminary studies in murine T cells, Jongeneel et al. (56) found that LT gene transcription was some 10-fold less than TNF gene transcrip- tion after T cell activation although accumulation of L T mRNA was greater than that of TNF mRNA; these authors did not comment on message stability in their abstract. These transcription data compare quite closely with our data in human T cells and suggest that post-transcriptional mecha- nisms play a relatively greater role in LT than TNF gene expression in human and murine T cells. Recently these workers have also reported differences in the promoter regions of the murine TNF and LT genes which may account for the tissue-specific expression of these two genes as well as their differential regulation in T cells (57).

The results of the cycloheximide experiments parallel those reported for most other T cell lymphokines. Cycloheximide did not block the induction of either TNF or LT by ConA -t PMA, variably induced small amounts of either or both mes- sages, and consistently resulted in the superinduction of large amounts of both messages when added 2-4 h after activation stimuli. This superinduction has been described for other lymphokines including IL-2, IFN-7, and GM-CSF after T cell activation (45). Superinduction of TNF and LT messages in mononuclear cells stimulated with PHA and PMA was also noted by Turner and Feldmann (53). This effect of protein synthesis blockers is generally assumed to result from inhi- bition of production of a protein or proteins involved in the degradative pathways of lymphokine mRNA (42, 45).

The relative instability of lymphokine messenger RNAs appears to be mediated at least in part by conserved AU-rich sequences (AUUUA) in the 3”untranslated regions of these lymphokine genes (58-60). These AU-rich sequences have also been implicated in the phenomenon of superinduction. Although the AUUUA motif appears to be required for rec- ognition of lymphokine messages by the degradative pathway, the recognition site must be more complex because this motif is also present in the 3”untranslated region of some stable mRNAs such as the mRNA of p-globin. Interestingly, the mRNAs of some lymphokine genes which are known to encode messages of very short half-lives (such as human GM-CSF and TNF) contain multiple copies of the AUUUA motif; GM- CSF contains seven copies, and TNF contains six. The human lymphotoxin gene contains only two of these motifs. Our data suggest that the mere presence of the AUUUA motif in the 3“noncoding region of a lymphokine gene is insufficient to mediate the very short half-lives of most lymphokine mes- sages; multiple copies of this motif may be necessary for these effects.

In conclusion, we found that the LT and TNF genes are regulated quite differently in T lymphocytes from human peripheral blood. Whether similar differences also occur within individual T cells or instead reflect differences in the mechanisms and rates of activation of distinct T cells that predominantly produce LT as compared with those that pre- dominantly produce TNF remains to be determined. Never- theless, these results suggest that the role of these two media- tors in the immune response may differ as well. Studies of the in vivo production and function of these cytokines will be required to define these roles.

Acknowledgments-We thank Patrick Gray and Roger M. Perl- mutter for cDNA clones, David B. Lewis for helpful comments, and Barbara Lovseth for assistance in preparation of the manuscript.

REFERENCES 1. Paul, N. L., and Ruddle, N. H. (1988) Annu. Reu. Immunol. 6 ,

2. Beutler, B., and Cerami, A. (1989) Annu. Reu. Immunol. 7 , 625- 655

3. Nedwin, G. E., Naylor, S. L., Sakaguchi, A. Y., Smith, D., Jarrett- Nedwin, J., Pennica, D., Goeddel, D. V., and Gray, P. W. (1985) Nucleic Acids Res. 1 3 , 6361-6373

4. Spies, T., Morton, C. C., Nedospasov, S. A., Fiers, W., Pious, D., and Strominger, J . L. (1986) Proc. Natl. Acad. Sci. U. S. A. 8 3 ,

5. Ruddle, N. H., and Waksman, B. H. (1967) Science 1 5 7 , 1060- 1062

6. Granger, G. A., and Williams, T. W. (1968) Nature 2 1 8 , 1253- 1254

7. Carswell, E. A., Old, L. J., Kassel, R. L., Green, S., Fiore, N., and Williamson, B. (1975) Proc. Natl. Acad. Sci. U. S. A. 7 2 , 3666- 3670

8. Perlmutter, D. H., Dinarello, C. A,, Punsal, P. I., and Colten, H. R. (1986) J. Clin. Invest. 7 8 , 1349-1354

9. Dinarello, C. A., Cannon, J. G., Wolff, S. M., Bernheim, H. A,, Beutler, B., Cerami, A., Figari, I. S., Palladino, M. A,, Jr., and O’Connor, J. V. (1986) J. Exp. Med. 1 6 3 , 1433-1450

10. Wong, G. H. W., and Goeddel, D. V. (1986) Nature 3 2 3 , 819- 822

11. Mestan, J., Digel, W., Mittnacht, S., Hillen, H., Bohm, D., Moller, A., Jacohsen, H., and Kirchner, H. (1986) Nature 3 2 3 , 816- 819

12. Trinchieri, G., Kobayashi, M., Rosen, M’., Loudon, R., Murphy, M., and Perussia, B. (1986) J. Exp. Med. 1 6 4 , 1206-1225

13. Shalaby, M. R., Aggarwal, B. B., Rinderknecht, E., Svedersky, L. P., Finkle, B. S., and Palladino, M. A., Jr. (1985) J. Zmmunol. 135,2069-2073

14. Perussia, B., Kobayashi, M., Rossi, M. E., Anegon, I., and Trin- chieri, G. (1987) J. Immunol. 1 3 8 , 765-774

15. Kehrl, J. H., Miller, A., and Fauci, A. S. (1987) J. Exp. Med.

16. Kehrl, J. H., Alvarez-Mon, M., Delsing, G. A., and Fauci, A. S. (1987) Science 2 3 8 , 1144;1146

17. Scheurich, P., Thoma, B., Ucer, U., and Pfizenmaier, K. (1987) J. Immunol. 138,1786-1790

18. Tracey, K. J., Beutler, B., Lowry, S. F., Merryweather, J., Wolpe, S., Milsark, I. W., Hariri, R. J., Fahey, T. J., 111, Zentella, A., Albert, J . D., Shires, G. T., and Cerami, A. (1986) Science 2 3 4 , 470-474

19. Piguet, P. F., Grau, G., Allet, B., and Vassalli, P. (1987) Immu- nobiology 1 7 5 , 27 (abstr.)

20. Cliff, A,, Defeo-Jones, D., Boyer, M., Martinez, D., Kierfer, D., Vuocolo, G., Wolfe, A., and Scoher, S. H. (1987) Cell 5 0 , 555- 563

21. Locksley, R. M., Heinzel, F. P., Shepard, H. M., Agosti, J., Eessalu, T. E., Aggarwal, B. B., and Harland, J. M. (1987) J. Immunol. 1 3 9 , 1891-1895

22. Broudy, V. C., Harlan, J. M., and Adamson, J. W. (1987) J. Immunol. 138,4298-4302

23. Aggarwal, B. B., Henzel, W. J., Moffat, B., Kohr, W. J., and Harkins, R. N. (1985) J. Biol. Chem. 260, 2334-2344

24. Gray, P. W., Aggarwal, B. B., Benton, C. V., Bringman, T. S., Henzel, W. J., Jarrett, J. A., Leung, D. W., Moffat, B., Ng, P., Svedersky, L. P., Palladino, M. A,, and Nedwin, G. E. (1984) Nature 312, 721-724

25. Aggarwal, B. B., Kohr, W. J., Haas, P. E., Moffat, B., Spencer, S. A., Henzel, W. J., Bringman, T. S., Nedwin, G. E., Goeddel, D. V., and Harkins, R. N. (1985) J. Biol. Chem. 260, 2345- 2354

26. Beutler, B., Greenwald, D., Hulmes, J . D., Chang, M., Pan, Y.-C. E., Mathison, J., Ulevitch, R., and Cerami, A. (1985) Nature

27. Pennica, D., Nedwin, G. E., Hayflick, J. S., Seeburg, P. H., Derynck, R., Palladinio, M. A., Kohr, W. J., Aggarwal, B. B., and Goeddel, D. V. (1984) Nature 312, 724-729

28. Cuturi, M. C., Murphy, M., Costa-Giomi, M. P., Weinmann, R., Perussia, B., and Trinchieri, G. (1987) J. Exp. Med. 165,1581- 1594

29. Green, L. M., Reade, J. L., Ware, C. F., Devlin, P. E., Liang, C.- M., and Devlin, J . J. (1986) J. Immunol. 1 3 7 , 3488-3493

30. English, B. K., Burchett, S. K., English, J . D., Ammann, A. J.,

407-438

8699-9702

166,786-791

316,552-558

Regulation of LT and T N F Genes in Human T Lymphocytes 7113

Wara, D. W., and Wilson, C. B. (1988) Pediatr. Res. 24, 717- and Wilson, C. B. (1991) J. Clin. Inuest. 87, 194-202 722 45. Shaw, J., Meerovitch, K., Elliott, J. F., Bleackley, R. C., and

31. Steffen, M., Ottmann, 0. G., and Moore, M. A. S. (1988) J . Paetkau, V. (1987) Mol. Irnrnunol. 24, 409-419 Irnrnunol. 140,2621-2624 46. Shaw, J.-P., Utz, P. J., Durand, D. B., Toole, J . J., Emmel, E. A.,

3 2 . Degliantoni, G., Murphy, M., Kobayashi, M., Francis, M. K., and Crabtree, G. R. (1988) Science 241, 202-205 Perussia, B., and Trinchieri, G. (1985) J . Exp. Med. 162,1512- 47. Crabtree, G. R. (1989) Science 243, 355-361 1530 48. Browning, J., and Ribolini, A. (1989) J . Irnrnurzol. 143, 1859-

3 3 ,

40.

41.

42. 43.

44.

Peters, P. M., Ortaldo, J . R., Shalaby, R., Svederesky, L. P., Nedwin, G. E., Bringman, T . S., Hass, P. E., Aggarwal, B. B., Herberman, R. B., Goeddel, D. V., and Palladino, M. A., Jr. (1986) J. Irnmunol. 137, 2592-2598

Paya, C. V., Kenmotsu, N., Schoon, R. A,, and Leibson, P. d. (1988) J . Irnrnunol. 141, 1989-1995

Sung, S.-S. J., Jung, L. K. L., Walters, J. A., Chen, W., Wang, C. Y., and Fu, S . M. (1988) J . Exp. Med. 168, 1539-1551

Wilson, C. B., and Remington, J. W. (1979) Clin. Exp. Irnrnunol. 3 6 , 511-520

Lewis, D. B., Larsen, A., and Wilson, C. B. (1986) J. Exp. Med.

Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, pp. 187-209, Cold Spring Har- bor Laboratory, Cold Spring Harbor, NY

Feinberg, A. P., and Vogeistein, B. (1983) Anal. Biochern. 132, 6-13

Herrin, D. L., and Schmidt, G. W. (1988) Biotechniques 6, 196- 200

Nelson, J . A,, and Groudine, M. (1986) Mol. Cell. Bid. 6, 452- 461

Taniguchi, T. (1988) Annu. Reu. Irnrnunol. 6, 439-464 Kronke, M., Leonard, W. J., Depper, J. M., and Greene, W. C.

Lewis, D. B., Yu, C. C., Meyer, J., English, B. K., Kahn, S. J.,

163, 1018-1023

(1985) J . Erp. Med. 161, 1593-1598

1867 49. Beutler, B., Krochin, N., Milsark, I. W., Luedke, C., and Cerami,

A. (1986) Science 232, 977-990 50. Sariban, E., Imamura, K., Luebbers, R., and Kufe, D. (1988) J .

Clin. Invest. 81, 1506-1510 51. Collart, M. A., Belin, D., Vassalli, J.-D., De Kossodo, S., and

Vassalli, P. (1986) J. Exp. Med. 164, 2113-2118 52. Burchett, S. K., Weaver, W. M., Westall, J. W., Larsen, A,,

Kronheim, S., and Wilson, C. B. (1988) J . Irnmunol. 140 , 3473-3481

53. Turner, M., and Feldmann, M. (1988) Biochern. Biophys. Res. Comrnun. 153, 1144-1151

54. Fischer, H., Dohlstein, M., Andersson, U., Hedlund, G., Ericsson, P., Hansson, J., and Sjogren, H. 0. (1990) J . Irnrnunol. 1 4 4 , 4663-4669

55. Sung, S-S. J., Bjorndahl, J. M., Wang, C. Y., Kao, H. T., and Fu, S. M. (1988) J. Exp. Med. 167,937-953

56. dongeneel, C. V., Shakhov, A. N., Nedospasov, S. A,, and Cerot- tini, J. C. (1989) J. Cell. Biochern. 13, 230 (abstr.)

57. Shakhov, A. N., Collart, M. A., Vassalli, P., Nedospasov, S. A,, and Jongeneel, C. V. (1990) J. Exp. Med. 171, 35-47

58. Shaw, G., and Kamen, R. (1986) Cell 46, 659-667 59. Caput, D., Beutler, B., Hartog, K., Thayer, R., Brown-Shimer,

S., and Cerami, A. (1986) Proc. Natl. Acad. Sci. U . S. A . 8 3 , 1670-1674

60. Brawerman, G. (1989) Cell 57,9-10