IL 1-a synthesis by pulmonary fibroblast subpopulations

Eur. J. Immunol. 1990. 20: 1723-1727

1723

Differential expression of interleukin la by Thy-l+ and Thy-l- lung fibroblast subpopulations: enhancement of interleukin la production by tumor necrosis factor-a*

Richard P. Phippsov, Clare Baecherovo, John G. Freliger.", David P. Permeyon, Peter Kengoa and Deborah Brown." University of Rochester Cancer Center. and Departments of Microbiology and Immunologyo, Pathology and Laboratory Medicinen, and Radiation Oncologya, University of Rochester School of Medicine and Dentistry, Rochester

The purpose of this investigation was to determine whether subpopulations of murine lung fibroblasts produced interleukin 1 (IL 1). We previously identified two major populations of pulmonary fibroblasts based on the presence or absence of Thy-1. Thy-l+ and Thy-l- subsets synsthesize fibronectin and type I and I11 collagen, but only the Thy-l- population displays class I1 major histocompatibility complex antigens after stimulation with interferon -y and presents antigen to T helper clones. Interestingly, in the current study we determined that only Thy-l- fibroblast lines and clones synthesized IL 1. Although constitutive production was low, tumor necrosis factor - a (TNF-a) stimulated 5-20-fold increases in IL 1production inThy-1- fibroblasts. The Thy-l+ fibroblasts did not produce IL 1 even after TNF-a treatment. Northern blot analysis of TNF-a treated cells revealed that in the Thy-l- subset increased mRNA levels for IL la were detected, while IL 1p mRNA was not detected. Furthermore, IL 1 activity from TNF-a-treated Thy-l- fibroblast membranes and supernatants was completely neutralized by IL la-specific antibodies. These observations support the hypothesis that the antigen-presenting Thy-l- subset is important for promoting the inflammation associated with pulmonary fibrosis. In addition, the existence of functional subsets of lung fibroblasts is further substantiated by differential expression of IL 1.

1 Introduction We recently identified two major populations of fibroblasts isolated from murine lung [l-31. Subsets of cells expressing or lacking the Thy-1 surface antigen were isolated by fluorescence activated cell sorting. Stable lines and clones were established. Thy-l+ and Thy-l- adherent pulmonary fibroblasts synthesize fibronectin and type I and I11 collagen and lack the characteristics of hematopoietic, epithelial and endothelial cells. Although both subsets responded to rIFN-y by up-regulating class I MHC expression, only the fraction which lacked Thy-1 expression displayed class I1 MHC. This population also presented antigen to T lymphocyte clones. This observation suggested that the Thy-lsubset might be involved in promoting chronic inflammation, which is associated with developing pulmonary fibrosis [4].

be capable of producing IL 1[7,8]. It is not known whether this IL 1 is synthesized by all lung fibroblasts or only a subset. The purpose of this investigation was to determine whether or not Thy-l+ and Thy-l- murine lung fibroblast populations synthesize IL 1. In assessing cytokines which might induce IL 1 production, we focused on TNF-a since it was shown to stimulate IL 1 production in human skin fibroblasts [9]. Moreover, pulmonary TNF-a mRNA levels are dramatically increased after treatment with bleomycin, a potent inducer of pulmonary fibrosis [lo]. Our data clearly demonstrate that TNF-a induces IL la expression in Thy-1-, but not in Thy-l+ fibroblasts. This observation supports our hypothesis that the antigen-presentingThy-1subset is important in promoting chronic pulmonary inflammation. Selective production of IL 1 by Thy-lfibroblasts also serves to reinforce the concept that functional subsets of fibroblasts exist.

IL 1is a potent proinflammatory polypeptide hormone [ 5 ] . This cytokine has been found in lung following chemotherapy-induced damage [6] and may be responsible, in part, for promoting pulmonary fibrotic disease [7]. Although 2 Materials and methods alveolar or interstitial M a are likely sources of IL 1 production, other lung cells including fibroblasts may also 2.1 Fibroblasts [I 83921

*

This research was supported by U.S.P.H.S. grants HL-39949, CA-42739, BRSG-STRR05403-27, 5-P3O-CA-11198 and ACS IM 535. This is publication 59 from the Immunology Unit of the Cancer Center. Supported by training grant T32-A1-07285.

+

Correspondence:Rick Phipps, Box 704, Cancer Center, University of Rochester School of Medicine and Dentistry, Rochester, NY 14642, USA 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheirn, 1990

The details of the preparation and characterization of Thy-l+ and Thy-l- murine lung fibroblast lines and clones have been described [l]. Stable lines (Fib2-T-3+,Fib2-T-4-) and clones (.1C6, .5C9, .05G9, .1D1) of these fibroblasts (between passage 5 to 20 after establishment) were used for the experiments described in this report. The cells are cultured in RPMI 1640/10% FBS and are maintained by weekly passage. The phenotype and characteristics of these fibroblasts have not changed, even with prolonged (i.e. 1 year) in vitro passage.

-

0014-2980/90/0808-1723$3.50+ .2YO

1724

Eur. J. Immunol. 1990. 20: 1723-1727

R. I? Phipps, C. Baecher, J. G. Frelinger et al.

2.2 Reagents

and the 636-bp IL 1p (murine) Bam Hl/Hind I11 fragment [19] were isolated on a 0.8% low-melt agarose gel and labeled via the random primer method [20] to specific activities of 3 x lo9-4 x lo9 cpm/pg.The RNA filters were prehybridized for 4-8 h, hybridized overnight at 42 "C with either the IL la or the IL 10 probe at 1 x 106 cpm/ml as described [17], and washed at a final stringency of 0.1 x SSC, 0.1% SDS 50°C.

Human rIL la and rIL l p were obtained from Genzyme Corp. (Boston, MA). Murine IL la was generously provided by Dr. Richard Chizzonite (Hoffmann-La Roche Inc., Nutley, NJ). In preliminary experiments fibroblasts were incubated with LPS (E. coli 055 : B5, Sigma Chemical Co.,St. Louis, MO); however, IL 1 activity was not induced. Murine rTNF-a was generously supplied by Genentech (South San Francisco, CA) and was used at 1-200 ng/ml. The activity of this preparation was 3.3 x lo7 3 Results U/mg. A second preparation of murine rTNF-a obtained from Genentech also stimulated IL 1 production by lung 3.1 Induction of IL 1 synthesis in Thy-1- pulmonary fibroblasts. An anti-murine IL la mAb (161.1) was genfibroblasts by TNF-a erously supplied by Dr. David Chaplin (Howard Hughes Medical Institute, Washington University, St. Louis, MO). SN from unstimulated Thy-1- (Fib2-T-4-) and Thy-lf The characteristics of the mAb have been described in (Fib2-T-3+) lines were examined for their ability to produce detail [ll]. mAb 161.1 was used at a final concentration of IL 1as measured by D10.G4.1 proliferation [12]. Small but 10 pg/ml in our studies. A hamster IgG antibody (Cappel, reproducible increases in the proliferation of D10.G4.1 Westchester, PA) was used as a specificity control at a cells were noted when SN from Fib2-T-4- cells were concentration of 100 pg/ml. A polyclonal goat anti-mouse examined (Fig. 1 and data not shown). No activity above IL la was generously supplied by Dr. Richard Chizzonite. background levels was found in Fib2-T-3+ SN. It was This antibody was used at a final dilution of 1/100 and possible that synthesis of an inhibitory molecule may have completely neutralized the activity of human or mouse rIL masked IL 1 activity. However, addition of rIL la to la. fibroblast SN failed to demonstrate the presence of inhibitory compounds (data not shown). Two substances, namely LPS and TNF-a have been shown to be potent inducers of 2.3 Assay for IL 1 IL 1[5,9]. Experiments using LPS showed no induction of IL 1activity by either subset of fibroblast (data not shown). The bioassay for IL 1 using D10.G4.1 cells has been TNF-a, however, proved to be a powerful IL 1stimulus for described elsewhere [12, 131. An assay for "membrane" Fib2-T-4- cells. Fig. 1 shows that SN from unstimulated IL 1 was also used and has been described by Kurt-Jones Fib2-T-4- contained a small amount of IL 1, which was et al. [131 and has been previously performed by us [ 141. In greatly increased 24 h after TNF-a treatment. In contrast, brief, fibroblasts treated for 6-72 h with TNF-a, were fixed no activity was found in SN from theThy-1+, Fib2-T-3+line, with 1% paraformaldehyde for 15 min, washed and incu- either before or after TNF-a treatment. Two other recombated for 24 h with culture medium [14]. On the day of binant murine cytokines were tested (IFN-y, IL4) and assay, 1 x 10"fibroblasts were washed with Hepes-buffered these failed to induce IL 1 activity in fibroblasts. medium and incubated with D10.G4.1 cells (2 x 104/well) in the presence or absence of Con A (2.5 pglml; Sigma). D10.G4.1 cells did not proliferate in the absence of Con A. 3.2 Clones of Thy-1-, but not Thy-l+ lung fibroblasts synthesize IL 1 in response to TNF-a In all experiments determinations were performed in triplicate and IL 1 activity in SN or on membranes was quantitated by comparison to dose-response curves We previously developed several clones of Thy-l+ and obtained using human rIL la (1000 U/ml, Genzyme Corp). Thy-1- fibroblasts that were derived from the parental lines The results are expressed as U IL 111 x lo6fibroblasts.The used in Fig. 1[11.We examined the SN from these clones for curves obtained with D10.G4.1 cells stimulated with IL la were not significantly altered by addition of TNF-a. All IL 1 (UNITS) experiments were repeated a minimum of three times with 1 2 3 4 similar results. CELLTYPEno.u0s FlbZ-T-4'

14

2.4 Northern hybridization Poly(A)+ RNA was isolated from Thy-1- and Thy-l+ fibroblasts and from LPS-treated P388D 1by the technique described by Badley et al. [15]. The relative quantity of poly(A)+ RNA from each cell line was determined by slot blot analysis of serial twofold dilutions of each RNA sample, which was probed with polynucleotide kinase 32P-labeledoligo-dT and washed at the reduced stringency of 2 x SSC at 45 "C, as described [161. For RNA blot analysis equal amounts of poly(A)+ RNA were analyzed on formaldehyde-containing 1% agarose gels, electrophoresed, and transferred to nitrocellulose filters [17]. For use as probes, the 603-bp IL l a (murine) Eco RI/EcO RV fragment [18]

FibZ-T-3*

24

Figurel. TNF-a enhances IL 1 synthesis by Thy-1-, but not Thy-l+ pulmonary fibroblast.Thy-1- (Fib2-T-4-) or Thy-l+ (Fib2T-3+) fibroblasts were treated with buffer [El) or with 10 (a),100 ) or 200 ng (U) per ml TNF-a.The SN were harvested at 24,48, or 72 h and assayed for IL 1 activity using the bioassay described in Sect. 2.3. Results are expressed as U IL 1/1 x lo6 fibroblasts and were comouted using a commercial IL la standard.

-

IL 1-asynthesis by pulmonary fibroblast subpopulations

Eur. J. Immunol. 1990. 20: 1723-1727 IL 1 (UNITS)

~~0

1

FIb2-1-4' .05G9

24

Flb2-T-4,101

24

Fib2-T-3' .1C6

24

FibZ-T-3'

24

.scs

1725

this activity could be increased two-to sixfold by addition of 10-100 ng/ml of rTNF-a. Membrane IL 1 activity was not detected in the Thy-l+ fibroblasts, even after TNF-a treatment. Several clones of both Thy-l+ and Thy-lfibroblasts were also tested. Uniformly, the clones mimicked the pattern of the parental lines (data not shown). That is, only Thy-1- clones expressed membrane IL 1 and could be induced for enhanced IL 1 expression after incubation with TNF-a.

2

3.4 Identification of the IL 1 activity in SN and on fixed Thy-1- fibroblasts as IL la

Figure 2. Clones of Thy-1-, but not Thy-l+ pulmonary fibroblasts synthesize IL 1 in response to stimulation by TNF-a. Fibroblasts were treated with buffer (El) or with 1(Q), 10 (B) or 100 (W) ng/ml TNF-a. Legend is similar to that of Fig. 1 except that clones of fibroblasts were used and SN were assayed at 24 h. Fibroblast clones were also tested at 48 and 72 h and only the Thy-1- clones showed any IL 1 activity (data not shown).

We wanted to determine whether the IL 1-like activity detected in the D10.G4.1 assay in fact was IL 1 and if so, whether it was IL l a , IL l p or both. Antibodies which specifically neutralize murine IL la were employed to help answer these questions [ll]. Fig. 4 shows the results using the 161.1 anti-murine IL la mAb. This mAb totally abrogated activity in both the SN and membrane fraction of constitutive IL 1 activity. Typically, a low level of IL 1 Thy-1- fibroblasts. A second reagent, a polyclonal goat activity was detected in SN fromThy-1- clones (e.g. .05G9, anti-mouse IL la antiserum, also completely neutralized .1D1, Fig. 2). However, after stimulation for 24 h with the activity in the SN and in paraformaldehyde-fixed cells TNF-a, the SN from Thy-1- clones contained substantial (data not shown). These observations indicate that the IL 1 activity (Fig. 2). Thy-l+ clones did not produce IL 1, activity detected in the D10.G4.1 bioassay is solely due to even after treatment for up to 4 days with TNF-a. In the production of IL l a , and not KL l p or some other addition to the clones shown in Fig. 2 two other clones from cytokine. each subset have been examined and again only Thy-lclones produced IL 1 (data not shown). 3.5 Induction of IL la, but not IL l p mRNA by TNF-a in Thy-1- fibroblasts 3.3 Induction of membrane IL 1 by TNF-a An additional approach for detection of cell cytokine IL 1 has been described as a protein which may either be profiles utilizes analysis of mRNA. Even though theThy-1secreted or exist as a membrane-associated form [13]. In fibroblasts failed to synthesize IL 1p protein, it was possible the mouse, IL la is associated with the membrane-bound that they were producing IL l p mRNA [8]. Similarly, the form, although it can be found in SN [21,22]; IL 1p has only Thy-l+ subset, which was not induced by TNF-a to produce been found in the soluble form [5,13]. Since two molecular IL 1, may nonetheless transcribe IL 1 mRNA. These forms of IL 1 have been identified it was possible that the possibilities were assessed by analyzing RNA for the Thy-l+ lines and clones produced only the membrane- presence of IL la and p transcripts. Hybridization of IL 1 bound form which may not be detected in SN.Therefore, IL la activity was assessed using paraformaldehyde-fixed cells IL 1 (UNITS) IL 1 NEUTRALIZING [13, 141. In the unstimulated Thy-1- lines membrane 2 SOURCE ANTIBODY . . . 3 . 4 . 5 . 6I 1(- 0.7U) was consistently detected (Fig. 3). Moreover, I

I

1

I

1

I

NONE HAMSTER

MEMBRANE IL I (UNITS) 0

1

2

3

4

4 MEMBRAN

I

1 Figure3. TNF-a enhances membrane IL 1 on Thy-1-, but not Thy-lf, pulmonary fibroblasts. Symbols as in Fig. 2. Fibroblast subpopulations were stimulated with 0, 1, 10, or 100 ng/ml TNF-a for 24-72 h. Cells were fixed with paraformaldehyde, washed, incubated overnight, rewashed and examined for membrane IL 1 activity (see Sect. 2.3 for details).

100 161.1

B

NONE HAYSTER

100 161.1

b

Figure 4. IL lais produced only by Thy-1- fibroblasts.TheThy-1fibroblastline, Fib2-T-4-,wastreated for 24 h with 1(ffl), 10 (a), or 100 (W) ng/ml TNF-a.The SN was harvested and incubated for 60 min with buffer, hamster IgG (as a negative control) or the 161.1 anti-IL la mAb. The SN was then assayed for IL 1 activity. For membrane-boundactivity the TNFa-treated fibroblasts were fixed with paraformaldehydeand 161.1antibody was added just prior to assay for rIL 1. Experiments were also performed using a polyclonal goat anti-mouse IL la antiserum. These results were identical to those shown using 161.1.

1726

R. €! Phipps, C. Baecher, J. G. Frelinger et al.

Eur. J. Immunol. 1990. 20: 1723-1727

Kurt-Jones et al. [24] found only IL la on adult human dermal fibroblasts after TNF stimulation. These investigators also found that LPS did not induce fibroblast IL 1 production, a finding which agrees with our result showing that LPS fails to stimulate lung fibroblast IL 1production. The difference between our findings and those of Le et al. [9] may be another reflection of fibroblast heterogeneity. For example, it is possible that foreskin fibroblasts represent a unique fibroblast population which may be present only early in development, or may be tissue specific. Our studies do not exclude the possibility that there are other subsets of lung fibroblasts that are capable of IL l p synthesis or that under the action of other stimuli Thy-l+ or Thy-1- fibroblasts might transcribe the IL l p gene. Figure5 IL la mRNA is increased after stimulation of Thy-1fibroblasts with TNF-a. Thy-l+ or Thy-1- fibroblasts were stimulated for 6 h with 10 nglml TNF-a. LPS-treatedP388D1 were used as positive control. RNA was extracted and hybridized with either IL la or IL If3cDNA. Note the increase in mRNA for IL la after TNF-a treatment of Thy-1- fibroblasts.

probes to LPS-induced P388D1 showed the presence of two species corresponding to IL la and 1L 1p message. In unstimulated or TNF-a-treated Thy-l+ fibroblasts, mRNA for either IL 1 species was not detected. However, IL la message was readily detected in the Thy-1- subset. After treatment with rTNF-a abundant IL l a , but not IL l p message was detected only in the Thy-1- subset (Fig. 5). These results show that Fib-2-T-4- only produces IL la and that Thy-l+ fibroblasts (Fib2-T-3+) are not induced to transcribe IL la or p mRNA.

It has been proposed that IL l p is secreted, whereas IL la is membrane bound [ l l , 131. For this reason we initially assumed that the IL 1 activity found in the SN of Thy-1fibroblasts would be IL 1p. However, two experiments pointed to the fact that this presumption was incorrect. First, the activity in the SN was totally neutralized by antibodies specific for IL la. Second, IL l a , but not IL 1p mRNA was detected in Thy-1- fibroblasts and was upregulated by TNF-a. The increase in IL la message reflected the increases in IL la detected using the bioassay. Although initially described as membrane-bound, recent evidence from two research groups support the concept that IL l a , which lacks a transmembrane sequence, is secreted by cells. The studies of Streck et al. [21] and of MinnichCarruth et al. [22] indicate that IL la is continually secreted. IL la was found in M a SN even after fixation with paraformaldehyde. It was concluded that the relevant form of IL la was as a secretory and not membrane-bound protein.

TNF-a may be a critically important cytokine for the development of pulmonary fibrosis. Recently, it was shown that shortly after bleomycin treatment (an agent which The fibroblast has often been considered a cell type induces pulmonary fibrosis), a massive and persistent important only in wound healing, fibrotic disease and as a increase in TNF-a mRNA occurs in murine pulmonary structural element to maintain tissue integrity. The fibro- tissue [10].The synthesis of large quantities of TNF-a may blast, however, is rapidly proving to be an important cell play a key role in promoting inflammation, which is closely capable of regulating immune responses in various tissues associated with pulmonary fibrosis. The importance of [ l , 8,231. Our previous research demonstrated that murine TNF-a may stem from its ability to stimulate IL 1produclung fibroblasts were divisible into two subsets [l-31. tion by interstitial fibroblasts. As shown by our study, Interestingly, only the Thy-1- lines and clones could be TNF-a is a potent enhancer of IL 1 synthesis by murine induced to express class I1 and present antigen to Th cells Thy-1- fibroblasts. IL 1 is a powerful promoter of fibro[l]. Data presented in this report shows that only the blast proliferation [25] and can stimulate fibroblast syntheThy-1- subset appears capable of producing IL 1, an sis of collagen [26].Thus, this single cytokine can drive two observation which supports the hypothesis that functional of the most important aspects of pulmonary fibrosis: subsets of fibroblasts exist. The capacity of the Thy-1- expansion of fibroblast populations and increased collagen population to express la antigens and t o synthesize IL la synthesis. In support of the critical role of IL 1 in fibrosis, reinforces the hypothesis that, in murine lung, it is the our preliminary studies show that both Thy-l+ and Thy-1Thy-1- fraction which participates in pulmonary immune subsets possess IL 1R and that IL la is a potent stimulator responses. Chronic inflammation is associated with pul- of murine lung fibroblast collagen synthesis [27]. We monary fibrosis [4], and Thy-l- fibroblasts might be propose that following damage to the lung from radiation important contributors to that process by producing IL 1 or chemotherapeutic agents, up-regulation of Ma-derived and by stimulatingTcel1 proliferation and cytokine produc- TNF-a will stimulate IL 1 production by pulmonary fibroblasts. In our model, production of IL 1would be restricted tion. to the Thy-1- subset. However, IL la could affect both The observation that neither Thy-l+ nor Thy-1- fibroblasts populations of fibroblasts eliciting proliferation and produce IL 16 is a surprising finding. As demonstrated by enhancing collagen synthesis. Moreover, IL la production the studies of Le et al. fibroblasts from human foreskin by Thy-1- fibroblasts would contribute to regional T cell rapidly elevate the level of IL 1p mRNA after TNF-a activation, a critical component for lung fibrosis to develop stimulation [9]. However, in support of our findings [28, 291.

4 Discussion

Eur. J. Immunol. 1990. 20: 1723-1727

In conclusion, our experiments demonstrate that the Thy-1- subset of murine pulmonary fibroblasts responds t o TNF-a by up-regulating IL 1 production. In conjunction with prior research showing that only the Thy-1- fraction presents antigen to T cells a picture is emerging which delegates to Thy-1- murine lung fibroblasts the role of activating T cells. The separation of fibroblasts into functional subsets based on the synthesis of cytokines such as IL 1,provides further evidence in support of the concept of fibroblast heterogeneity. We thank Dr. David Chaplin for the 161.1 monoclonal anti-lL l a antibody and for plasmids pMuIL I a and pMulL I p, Dr. Richard Chizzonite for goat anti-murine IL I a, Dr. H. Quill and D.Phipps for reviewing the manuscript, and D. Romaniw and M. Palmierefor expertly preparing this document. Received March 9. 1990.

5 References Phipps, R. I?, Penney, D., Keng, I?, Quill, H., Paxhia, A., Derdak, S. and Felch, M. E., A m . J. Respir. Cell Mol. Biol. 1989. I : 65. Phipps, R. I?, Penney, D. P., Keng, P., Quill, H., Paxhia, A., Derdak, S. and Felch, M., J. Cell Biochem. 1989. 13A: c148. Penney, D. I?, Phipps, R. P. and Keng, I? C., Anat. Rec. 1989. 223: 89A. Reiser, K. M. and Last, J. A., Exp. Lung Res. 1989. 10: 331. Oppenheim, J. J., Kovacks, E. J., Matsushima, K. and Durham, S. K., Immunol. Today 1986. 7: 45. Suwabe, A.,Takahashui, K.,Yasui, S., Arai, S. and Sendo, F., Am. J. Pathol. 1988. 132: 512. Bochner, B. S., Landy, S. D., Plant, M., Dinarello, C. and Schleimer, R. I?, I. J. Immunol. 1987. 139: 2297. Elias, J., Reynolds, M., Kotloff, R. M. and Bean, J., Proc. Natl. Acad. Sci. USA 1989. 86: 6171.

IL 1-a synthesis by pulmonary fibroblast subpopulations

1727

9 Le, J.,Weinstein, D., Gubler, U. and Vilcek, J., J. Immunol. 1987. 138: 2137. 10 Piguet, P. F., Collart, M. A., Grau, G. E., Kapanci, Y. and Vassalli, I?, J. Exp. Med. 1989. 170: 655. 11 Fuhlbrigge, R. C., Sheehan, K., Schreiber, R. D., Chaplin, D. and Unanue, E., J. Immunol. 1988. 141: 2643. 12 Kaye, J., Gillis, S., Mizel, S. B., Shevach, E. M., Mulch,T. R., Dinarello, C. A., Lachman, L. B. and Janeway, C. A., J. Immunol. 1984. 133: 1339. 13 Kurt-Jones, E. A., Beller, D. I., Mizel, S. B. and Unanue, E. R., Proc. Natl. Acad. Sci. USA 1985. 82: 1204. 14 Quill, H., Gaur, A., Brown, D., Infante, A. J. and Phipps, R. I?, J. Immunol. 1989. 143: 2242. 15 Badley, J. E., Bishop, G. A., St. John,T. and Frelinger, J. G., Biotechnology 1988. 6: 114. 16 Harley, C. B., Gene Anal. Tech. 1987. 4: 17. 17 Thomas, I? S., Proc. Natl. Acud. Sci. USA 1980. 77: 5201. 18 Lomedico, P., Gubler, U., Hellman, C. P., Dukovich, M., Giri, J. G., Pan,Y. C., Collier, K., Semionow, R., Chua, A. 0. and Mizel, S. B., Nature 1984. 312: 458. 19 Gray, I? N., Glaister, D., Chen, E., Goeddel, D. and Pennica, D., J. Immunol. 1986. 137: 3644. 20 Feinberg, A. I? and Vogelstein, B., Anal. Biochem. 1983. 132: 6. 21 Streck, H., Giinther, C., Beuscher, H. and Rollinghoff, M., Eur. J. Immunol. 1988. 18: 1609. 22 Minnich-Carruth, L., Suttles, J. and Mizel, S. B., J. Immunol. 1989. 142: 526. 23 Umetsu, D. T., Katzen, D., Jabara, H. H. and Geha, R. S . , J. Immunol. 1986. 136: 440. 24 Kurt-Jones, E. A,, Fiers, W. and Pober, J. S., J. Immunol. 1987. 139: 2317. 25 Schmidt, J. A., Mizel, S. B., Cohen, D. and Green, I., J. Immunol. 1982. 128: 2177. 26 Dinarello, C. A., Rev. Infect. Dis. 1984. 6: 51. 27 Derdak, S., Silvera, M., Felch, M., Penney, D., Keng, I? and Phipps, R. P., A m . Rev. Resp. Dis. 1990. 141: A703. 28 Stein-Streilein, J., Lipscomb, M. E, Fisch, H. and Whitney, P., Am. Rev. Resp. Dis. 1987. 136: 119. 29 Schrier, D. J., Phan, S. H. andMcGarry, B. M., A m . Rev. Resp. Dis. 1983. 127: 614.