The distribution of perforin in normal tissues

Immunology Letters, 28 (1991) 195-200 Elsevier IMLET 01576

The distribution of perforin in normal tissues Sanjay V. J o a g 1, L i - M o u Z h e n g 1, Pedro Persechini 1, J o s e f Michl 2, Earl Parr 3 a n d J o h n Ding-E Young IT he Rockefeller University, New York, NY, 2SUNY Health Sciences Center, Brooklyn, NY, 3Southern Illinois University Medical School, Carbondale, IL, U.S.A. (Received 17 January 1991; accepted 29 January 1991)

1. Summary

We describe the production of monoclonal antibodies to murine and human forms of the lymphocyte pore-forming protein (perforin, PFP, or cytolysin), a major granule-localized cytolytic mediator of CTL and NK cells. Antibodies were raised against both murine perforin purified from a CTL line, and human perforin expressed in bacteria as a fusion protein with the Escherichia coli TrpE protein. Antibodies raised against either immunogen inhibited the hemolytic activity of murine perforin, and thus may enable us to identify the pore-forming or selfassociative domain of perforin. One mAb, MP1, was used to study the distribution of perforin in murine tissues under physiological conditions. We found that perforin was expressed in the granular metrial gland (GMG) cells of the pregnant murine uterus, but not in other tissues examined. These results further support the view that perforin is induced only in activated cytolytic lymphocytes, and raise the question whether perforin-containing GMG cells represent an effector of a maternal immune response to the fetus. 2. Introduction

The lymphocyte pore-forming protein (PFP, perforin, or cytolysin) is a potent mediator of cytotoxicity found in the granules of CTL and NK cells [1 - 4]. Key words: Perforin; Monoclonal antibody; Granular metrial gland cell; Lymphocyte Correspondence to: Sanjay Joag, Laboratory of Cellular Physiology and Immunology, The Rockefeller University, 1230 York Avenue, New York, NY 10021, U.S.A. 0165-2478 / 91 / $ 3.50 © 1991 Elsevier Science Publishers BY.

In the presence of calcium, PFP/perforin, a monomer of 70 kDa, polymerizes to form transmembrane tubules, leading to lysis of target membranes. Monoclonal antibodies (mAb) to perforin could be useful for identifying perforin produced by cytotoxic cells in lesions, such as tumors and foci of viral and other infections, and in assigning functional activities to the various structural domains of the perforin molecule. We describe here the production of mAb to both human and murine perforin and report on the results of our immunohistochemical study of murine tissues. 3. Materials and Methods

3.1. Purification of perforin and immunization of mice Perforin was purified to homogeneity from the murine CTL line CTLL-R8 as described [5]. Homogeneity of perforin was confirmed by gel electrophoresis and amino acid sequencing [5]. Perforin was boiled in 1070 SDS, mixed with Freund's adjuvant, and injected i.p. into female BALB/c mice, at 2 - 5 ~g/mouse. Three booster injections were given. A 1.9-kb EcoRV-HindIII fragment of human perforin cDNA clone HP10 [6] was subcloned into the bacterial expression vector pATH 10 [7], generously provided by A. Tzagaloff (Columbia University, New York). On immunoblots, the 97-kDa recombinant fusion protein reacted with anti-perforin antisera raised against purified mouse perforin [5] and against a synthetic peptide corresponding to amino acids 1-19 of human perforin [8]. The TrpEperforin fusion protein was purified by preparative SDS-PAGE and injected i.p. into female BALB/c 195

mice, at 2 5 - 50/zg/mouse. Four booster injections were given.

3.2. Monoclonal antibodies to murine and human perforin Spleen cells from a mouse immunized with murine perforin were fused with NS-1 myeloma cells [9]. Supernatants were tested by ELISA against partially purified perforin, and positive cultures rescreened by ELISA against purified perforin. Reactive clones were subcloned by limiting dilution. One subclone, MP1, was selected for further characterization. Spleen cells from a mouse immunized with recombinant human perforin were fused with P3U-1 myeloma cells. Supernatants were tested by ELISA against bacterial lysates containing the TrpEperforin fusion protein, and separately against lysates containing the TrpE protein, to eliminate those antibodies which reacted against the TrpE portion of the fusion protein, or other bacterial proteins. Positive clones were rescreened by ELISA against IL-2 stimulated human NK cell lysates, generously provided by B. Perussia (Wistar Institute, Philadelphia). Reactive clones were subcloned by limiting dilution. One subclone HP1 was selected for further characterization. Globulin fractions of tissue culture supernatants from clones MP1 and HP1 were precipitated at 50°70 saturation of (NH4)2SO4 and purified by gel filtration through a Sephacryl S-300 column (Pharmacia). Fractions reactive on ELISA separated as a sharp peak, and were dialyzed against PBS and adjusted to 100/~g/ml using a BCA protein assay kit (Pierce). Isotyping of clones was done by ELISA using a mouse monoclonal antibody isotyping kit (Biorad).

3.3. Inhibition of hemolysis Perforin-containing extracts of CTLL-R8 cells were prepared and the hemolytic activity was measured as described [10]. PBS and an ascites preparation of MOPC 104E myeloma (Cappel) containing 200/~g/ml of specific mAb were used as controls. MOPC 104E is a mineral oil-induced murine plasmacytoma producing IgM. As perforin is known to polymerize and lose activity in the presence of free Ca 2÷, perforin-containing CTLL-R8 lysates were 196

prepared in 0.5 mM EGTA and the mAb preparation was dialyzed extensively against calcium-free PBS. Moreover, in preliminary experiments (not shown) we confirmed that a solution of 5 mM CaC12 failed to inhibit lysis in our assay. 3.4. Immunoblotting Proteins were transferred electrophoretically from SDS-PAGE and non-denaturing polyacrylamide gels onto Immobilon-P membranes (Millipore) as described [11], and immunoblots were performed using 125I-labeled anti-mouse Ig (Amersham) as the secondary antibody. 3.5. Immunohistochemistry Cytocentrifuge preparations of CTLL-R8 and YAC-1 cells were prepared, fixed with acetone for 10 min and methanol for 30 s, and stained sequentially with mAb and either FITC-conjugated or biotin-conjugated goat anti-mouse IgM (t~ chainspecific) antiserum (Cappel). Biotinylated antibodies were detected with an ABC immunoperoxidase kit (Vector Labs., Burlingame, CA; prepared according to manufacturer's instructions). Biopsies of lung, liver, spleen, proximal small intestine, external ear, parotid gland, and cervical lymph nodes were taken from C57/BL6 mice (Trudeau Inst.), and immediately frozen in Tissue-Tek O.C.T. (Miles Scientific). Uteri were obtained from pregnant Swiss mice (Charles River Labs.) as previously described [12], and snap frozen in Tissue-Tek O.C.T. Tissue sections 10-12 #m thick were cut with a cryostat and dried onto gelatin-coated slides, and then fixed with acetone for 10 min and methanol for 30 s. Slides were stained as described [13]. All results obtained with mAb were confirmed using specific anti-perforin rabbit polyclonal antisera [5]. 4. Results and Discussion

4.1. Characterization of mA b With an ELISA-based mouse mAb isotyping kit, both MPI and HP1 were found to be IgM with light chains. MP1 was reactive on ELISA against purified perforin. We have previously used the same batch of purified perforin for direct amino acid sequencing

a~

y

1O0

80

_

_

S

_

p I

(0

0

E

60

/

/

//

I s/

I •

4/,..`

40

©

I 20

/

O 0

~.'~' 1 2

, 3

4

5

6

Perforin added (HU)

Fig. 1. Inhibitionof hemolysisby MP1 and HP1. In triplicate on microtiter plates, 120/H CTLL-R8lysates were incubated on ice for 15 min with 20 #1 PBS (closed circles), MP1 (open circles) or HP1 (closed triangles) both at 100 ~tg/ml, or 200 #g/ml MOPC 104E (open triangles). 20 tzl of a sRBC suspension (5 × 108/ml) and 10 tA 20 mM CaC12were added, and incubation continued for 30 min at 37 °C. After centrifugation (200xg, 5 min), 75/zl of the cell-freesupernatant was transferred to a new plate and absorbance measured at 410 nm. RBC were incubated with PBS or dd HzO to determine spontaneous or total release.

[5]. MP1 failed to react against lysates from murine lymphomas YAC-1 and EL-4, and murine mastocytoma P815, and reacted weakly with IL-2 stimulated human NK cell lysates. Similarly, in EL1SA, HP1 reacted strongly with IL-2 stimulated human NK cell lysates and weakly with C T L L R8 lysates, but not with similar preparations from peripheral blood lymphocytes or the erythroleukemia cell line K562. Both MP1 and HP1 failed to react with perforin containing cell lysates on immunoblots performed on non-denaturing and on SDS-containing polyacrylamide gels, under both reducing and non-reducing conditions (data not shown).

Fig. 2. Cytocentrifuge preparation of CTLL-R8 cells stained sequentially with MP1 and biotin-conjugated goat anti-mouse IgM antiserum. The slide was then incubated with avidin-biotinylated peroxidase complex (ABC) as described in Materials and Methods. × 270.

ence of IgM, and is therefore specific to the mAb. Based on its ability to inhibit hemolysis produced by murine perforin, and on its reactivity in ELISA, HP1 appears to recognize perforin specifically. 4.3. Immunocytochemistry Cytocentrifuge preparations of CTLL-R8 (Fig. 2) stained with MP1, but not with HPI. YAC-1 cells were not stained with either antibody (not shown). However, MP1 stained cells only under very mild fixation conditions (cold acetone for 10 min followed by methanol for 30 s). Based on the reactivity with purified perforin on ELISA, staining of perforincontaining cells on immunohistochemistry, and ability to inhibit the hemolytic activity of perforin, MP1 clearly recognizes perforin specifically. 4.4. Distribution of perforin in normal tissues

4.2. Inhibition of hemolytic activity of perforin Both MP1 and HP1 inhibited the hemolytic activity of murine perforin (Fig. 1) when tested at a concentration of 100/~g/ml. A control IgM monoclonal antibody (MOPC 104E) tested at a concentration of 200 lzg/ml dit not inhibit hemolysis. Under the conditions tested, inhibition of hemolytic activity cannot be due either to extraneous Ca 2+ , or to the pres-

The availability of a specific monoclonal antibody to perforin allowed us to study the distribution of perforin in normal tissues. Perforin-containing cells were not seen in liver, lung, spleen, lymph nodes, salivary gland, small intestine, and skin from the external ear (not shown). These findings were confirmed using polyclonal anti-perforin antisera. Perforin is known to be induced by antigenic or 197

Fig. 3. (A) Cross section of the murine uterus at 11 days of pregnancy, showing the decidua (D), embryo (E), labyrinth (L), and metrial gland (MG), x23; (B) PAS stained section of the metrial gland. The granules of GMG cells (arrow) stain strongly with PAS, x320; (C) section of metrial gland, stained sequentiallywith MP1 and FITC-conjugatedgoat anti-mouse IgM antiserum and examinedusing fluorescence optics, × 110. mitogenic stimuli in primary cells in vitro [14], suggesting that it is a marker for activated lymphocytes. Therefore it is likely to be present in vivo only in tissues that contain activated lymphocytes. We have previously suggested [15, 16] that granulated metrial gland (GMG) cells in the pregnant murine uterus may represent activated, perforin-containing large granular lymphocytes. However, those studies were based on polyclonal anti-perforin antisera and posed the question of specificity. The availability of a m A b to perforin enabled us to resolve this question (Fig. 3). The granules of G M G cells are rich in glycoproteins and chondroitin sulfate, and are best seen in periodic acid Schiff (PAS) stained sections (Fig. 3B). In Fig. 3C we show a section of the metrial gland in which the G M G cells are stained with MP1. As perforin ÷ cells in vivo have hitherto been described only in viral or autoimmune diseases [13, 17], where they clearly appear to be a part of an ongoing immune response, their appearance in the pregnant uterus may represent a maternal immune response to the fetus. We have described the production of mAb to perforin and studied its distribution in normal tissues in vivo. These, and similar, reagents can serve to examine the expression of perforin in vivo, and so help clarify the role of cytolytic lymphocytes in gestation, in immune response to infection, and in autoim198

mune diseases. While this manuscript was in preparation, in contrast to our observations Kawasaki et al. [18] reported that mAb to perforin stained a minor population of cells in the spleen, liver, and lymph nodes of normal mice. The differences in these observations and their significance remain to be determined, but it should be noted that to date we have not been able to stain the normal resident tissues reported here using either mAb or polyclonal antisera specific for perforin.

Acknowledgements This work was supported in part by grants from the Irvington Institute for Medical Research, the American Heart Association, New York City Affiliate, and U.S. Public Health Service Grants CA47307 and HD-17480. L. M. Zheng is a Fellow of the Lalor Foundation. J. D.-E Young is a Lucille P. Markey Scholar. We thank Drs. Z. A. Cohn and R. M. Steinman for their advice and support, and for critical reading of this manuscript. We also thank Drs. A. Szary and M. Parr for help in preparing the mAb, Dr. C.-C. Liu for her assistance with the early stages of these studies, and X.-H. Ding for excellent technical assistance.

References [1] Podack, E. R. (1985) Immunol. Today 6, 21. [2] Henkart, P. A. (1985) Annu. Rev. Immunol. 3, 31. [3] Young, J. D.-E and Cohn, Z. A. (1987) Adv. Immunol. 41, 269. [4] Tschopp, J. and Jongeneel, C. V. (1988) Biochemistry 27, 2641. [5] Persechini, P. M. and Young, J. D.-E (1988) Biochem. Biophys. Res. Commun. 156, 740. [61 Liu, C.-C., Rafii, S., Granelli-Piperno, A., Trapani, J. A. and Young, J. D.-E (1989) J. Exp. Med. 170, 2105. [7] Spindler, K. R., Rosser, D. S. E and Berk, A. J. (1984) J. Virol. 49, 132. [8] Young, L. H. Y., Joag, S. V., Zheng, L. M., Lee, C. P., Lee, Y. S. and Young, J. D,-E (1990) Lancet 336, 1019. [9] Kohler, G. and Milstein, C. (1976) Eur. J. Immunol. 6, 511. [10] Jiang, S., Persechini, P. M., Zychlinsky, A., Liu, C.-C., Perussia, B. and Young, J. D.-E (1988) J. Exp. Med. 168,

2207. [11] Young, J. D.-E, Liu, C.-C., Leong, L. G. and Cohn, Z. A. (1986) J. Exp. Med. 164, 2077. [12] Zheng, L. M., Ojcius, D. M., Liu, C.-C., Kramer, M. D., Simon, M. M., Parr, E. L. and Young, J. D.-E (In Press) FASEB J. [13J Young, L. H. Y., Klavinskis, L. S., Oldstone, M. B. A. and Young, J. D.-E (1989) J. Exp. Med. 169, 2159. [14] Joag, S.V., Liu, C.-C., Kwon, B. S., Clark, W. R. and Young, J. D.-E (1990) J. Cell. Biochem. 43, 1. [15] Parr, E. L., Parr, M. B. and Young, J. D.-E (1987) Biol. Reprod. 37, 1327. [16] Parr, E. L., Young, L. H. Y., Parr, M. B. and Young, J. D.-E (1990) J. Immunol. 145, 2365. [17] Young, L. H. Y., Peterson, L. B., Wicker, L. S., Persechini, P. M. and Young, J. D.-E (1989) J. Immunol. 143, 3994. [18] Kawasaki, A., Shinkai, Y., Kuwana, Y., Furuya, A., Iigo, Y., Hanai, N., Itoh, S., Yagita, H. and Okumura, K. (1990) Int. lmmunol. 2, 677.

199