Perforin Gene Expression in Granular Lymphocyte Proliferative Disorders

By Kazuo Oshimi, Yoichi Shinkai, KO Okumura, Yoko Oshimi, and Hideaki Mizoguchi

By Northern blot analysis using a cDNA clone of the perforin gene, we studied the levels of perforin mRNA in peripheral blood mononuclear cells from 11 cases of granular lymphocyte-proliferative disorders (GLPDs). The granular lymphocytes studied were characterized by mor- phologic, immunophenotypic, and immunogenotypic analy- ses. Cytolytic functions of the lymphocytes assayed in- cluded nonmajor histocompatibility complex-requiring cytotoxicity, anti-CD3-redirected cytotoxicity, antibody- dependent cellular cytotoxicity, and lectin-dependent cellu- lar cytotoxicity. The results showed that in lymphocytes with strong cytolytic functions high levels of perforin

GROUP OF disorders, collectively termed granular A lymphocyte-proliferative disorders (GLPDs),' chronic Ty-lymphoproliferative disease,* or large granular lympho- cyte l e ~ k e m i a , ~ is characterized by the proliferation of granular lymphocytes (GLs), ie, lymphocytes with azuro- philic granules in the cytoplasm. GLPDs of the CD3+4-8+ phenotype and CD3- 16+ phenotype are common, whereas those of CD3+4-8- and CD3+4+8- are rare. Proliferating GLs sometimes exhibit nonmajor histocompatibility complex (MHC)-requiring cytotoxicity, antibody-dependent cellular cytotoxicity (ADCC), or lectin-dependent cellular cytotoxic- ity (LDCC). Furthermore, CD3+WT31+ GLs exhibit cyto- toxicity for immunoglobulin G Fc receptor-positive target cells when anti-CD3 monoclonal antibody (MoAb) or WT31 MoAb is added during the cytotoxicity assay.4 The mecha- nisms of cytolysis observed in these GLs have not been studied, however.

Perforin is a 70-Kd cytolytic protein demonstrated in the cytoplasmic azurophilic granules of in vitro cultured cyto- toxic T lymphocytes (CTLs) and natural killer (NK) cells, and is considered responsible for their cytolytic Recently, the cDNA and amino-acid sequences of murine and human perforin have been determined by investigators, including o ~ r s e l v e s . ~ ~ ' ~ We report the results of experiments in which a cDNA clone of the human perforin gene was used to study the expression of perforin mRNA in peripheral blood mononuclear cells (PBMNCs) obtained from 11 pa- tients with GLPDs.

MATERIALS AND METHODS

Diagnosis of GLPDs. As described previously,' diagnosis of GLPDs was made when GL lymphocytosis 2 2,00O/pL persisted in

From the Department of Medicine. Division of Hematology, Tokyo Women's Medical College: and Department of Immunology, Juntendo University School of Medicine, Tokyo, Japan.

Submitted January 26,1989; accepted October 9.1989. Address reprint requests to K. Oshimi, MD, Department of

Medicine. Division of Hematology, Tokyo Women's Medical Col- lege, 8-1, Kawada-rho. Shinjuku-ku. Tokyo 162. Japan.

The publication costs of this ariicle were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. section I734 solely to indicate this fact.

0 I990 by The American Society of Hematology. 0006-4971/90/7503-0036$3.00/0

mRNA existed, whereas in lymphocytes with weak or undetectable levels of cytolytic functions, low levels of perforin mRNA existed. Because the levels of perforin mRNA correlated with those of cytolytic functions, perforin is probably a mediator in cytolytic functions of granular lymphocytes in patients with GLPDs. When the lympho- cytes were cultured for 1 day, however, the levels of cytolytic activity were increased, and those of perforin mRNA were decreased. Therefore, we cannot rule out the possibility that factors other than perforin protein are involved in the cytolytic functions of granular lymphocytes. 0 1990 by The American Society of Hematology.

patients for more than 6 months. GLs were defined as lymphocytes with three or more azurophilic granules in the cytoplasm, irrespec- tive of their cell size.'

PB- MNCs isolated by Ficoll-Conray were stained by indirect immuno- fluorescence with the following antibodies: OKT3 (anti-CD3), OKT4 (anti-CD4), and OKT8 (anti-CD8) from Dr G. Goldstein, Ortho, Raritan, NJ; WT31 (which may react with T-cell antigen receptor (TCR)".'2) from Sanbio, Uden, the Netherlands; and CLB FcR gran 1 (anti-CD16) from Dr P. A. T. Tetteroo, the Netherlands Red Cross, Amsterdam. Stained cells were analyzed in a cell sorter as described previously."

Immunogenotypic analysis of TCR-(3 gene rearrangements was performed by Drs Y. Kobayashi and F. Ishikawa at the Department of Medicine, University of Tokyo Faculty of Medicine, as described previo~sly.4~'~

The method of cytotoxic assay is described el~ewhere.4~"~'~ PBMNCs 1 x IO5 were incubated, in triplicate, with 5 x IO3 5'Cr-labeled target cells for 5 hours, after which the supernatant was harvested and the specific "Cr release calculated according to the formula: percentage of release = [(experimental 5 'Cr release - spontaneous 5'Cr release)/(maximum 5'Cr release - spontaneous "Cr release)] x 100. Anti-CD3 MoAb OKT3 (1 pg/mL) and WT31 MoAb (IOpg/mL) were added 1 hour before the "Cr-release assay was started.

To assay ADCC, heat-inactivated rabbit anti-P815 serum was added at a final dilution of 1:900 to 5'Cr-labeled P815 target cells in the cytotoxicity assay. LDCC was assayed against S'Cr-labeled P8 15 or K562 target cells, after addition of 1 :9,000-diluted phytohemag- glutinin (PHA)-P (GIBCO, Grand Island, NY).

Because anti-CD3 and WT31 MoAbs induced non-MHC- requiring cytotoxicity for IgG Fc receptor-bearing targets in patient PBMNCs after 1-day incubation of PMBNCs; cytotoxicity assays were performed using freshly isolated PBMNCs and after 1-day incubation of PBMNCs in RPMI 1640 medium with 10% fetal calf serum (FCS) in microculture plates.

Cytolytic functions and expression of perforin mRNA were studied in normal CD3- 16+ NK cells and CD3+16- cells. These cells were isolated by Percoll density-gradient centrifugation followed by negative selection with the panning method, as described p r e v i o ~ s l y . ~ ~ ' ~ The isolated CD3-16+ NK fraction from a donor contained 97% GLs, 95% CD16' cells, and 4% CD3+ cells. The isolated CD3+16- T-cell fraction contained 3% GLs, 5% CD16' cells, and 97% CD3+ cells.

Cytoplasmic RNA from 1 x IO6 freshly isolated PBMNCs was prepared by the vanadyl-ribonucleoside complexes-NP40 electrophoresed on a 1% formaldehyde-agarose gel, transferred to a Nytron mem- brane (Schleicher & Schuell, Dassel, FRG). RNA blot was prehy-

Immunophenotypic and immunogenotypic gene analyses.

Cytotoxic assays.

Isolation of normal NK and T cells.

Northern blot analysis of perforin mRNA.

704 Blood, Vol 75, No 3 (February 1). 1990: pp 704-708

PERFORIN MRNA IN GLPDs 705

bridized at 65OC for 4 hours in 50% formamide, 6 x SSPE (1 x SSPE = 180 mmol/L NaCl/sodium phosphate, pH 7.7/1 mmol/L EDTA), 2 x Denhardt's solution, 0.5 x sodium dodecyl sulfate (SDS) and 100 pg/mL sheared salmon sperm DNA. Blot was hybridized in the same buffer with human perforin riboprobe (1 x IO8 dpmlpg, 2.5 x IO5 dpm/mL) 32P-labeled by the T3 RNA transcription system (Stratagene Cloning Systems, La Jolla, CA). After hybridization, the blot was washed four times in 0.1 x SSPE, 0.1% SDS at 65OC for 1 hour. The blots were exposed to Kodak XAR-5 film with an intensifying screen at - 7OoC for 2 to 7 days. In some experiments, hybridized filters were rewashed for 30 minutes at room temperature in 2 x SSPE containing 1 pg/mL RNase A and washed twice for 15 minutes at 65OC in 0.1 x SSPE, 0.1% SDS.

The human perforin probe" consisted of a 2.1-kilobase (kb) fragment containing the coding sequence of human perforin cDNA of the clone YT-I 32 in a Bluescript plasmid (Stratagene). Riboprobe was synthesized from YT-132 linearized with EcoRV by T3 RNA polymerase with CT-~*P-UTP (3,000 Ci/mmol).

Because the cytolytic functions were assayed in terms of lympho- cyte number, the levels of perforin mRNA were also assayed in terms of lymphocyte number. Determination of ribosomal RNA showed no substantial loss of RNA contents during the procedure.

RESULTS In 7

of the 11 patients with GLPDs, the clinical features and PBMNC surface phenotypes have been described previo~sly. ' .~, '~ As shown in Table 1, the GL percentages in PBMNCs were high, ranging from 73% to 98%. Phenotypic markers of GLs were CD3+4-8+16+ in 6 patients, CD3'4-8-16+ in one, CD3+4+8-16+ in 1, and CD3-4-8-16+ in 3. The TCR-@ gene was monoclonally rearranged in most patients with CD3+ GLPDs and was in germline configuration in CD3- GLPDs. Non-MHC- requiring cytotoxicity of freshly isolated PBMNCs was either high or low and in CD3+ GLPDs was either enhanced or inhibited by addition of anti-CD3 and/or WT31 MoAbs. Most patients had ADCC and LDCC.

As controls, 1 patient with CD3+4+8- adult T-cell leuke-

Patient characteristics and PBMNC cytotoxicity.

mia (ATL), 2 patients with CDlO+ 19+20+ acute lymphoblas- tic leukemia (ALL), and 1 normal individual were included. In these patients, more than 80% of PBMNCs were leukemic cells, and 52% were GLs. In one patient with ALL, 47% of PB lymphoblasts contained several prominent azurophilic granules in the cytoplasm. PBMNCs of the three leukemia patients exhibited less than 5% of specific "Cr release in cytolytic assays (data not shown). CD3-16+ NK cells of the normal donor exhibited strong non-MHC requiring ADCC and LDCC cytotoxicity (Table 1).

Fig- ure 1 shows the results of Northern blot analysis of cytoplas- mic RNA extracted from freshly isolated PBMNCs of the populations studied. In PBMNCs from patients with GLPDs, a single region of hybridization was detected with the perforin probe at approximately 3 kb; the level of expression differed from patient to patient. In PBMNCs from patients with ATL and ALL, perforin mRNA was rarely detected. In cells from a normal individual, low levels of perforin mRNA were detected in whole PBMNCs and the CD3+16- T-cell fraction and a high level of mRNA was detected in the CD3-16+ NK fraction. On the whole, the data indicate that PBMNCs lacking GLs rarely expressed perforin mRNA whereas PBMNCs containing GLs expressed perforin mRNA. The levels of perforin mRNA did not correlate with the percentage of GLs, however; instead they correlated with the highest level of cytotoxicity obtained.

PBMNC cytotoxicity and perforin mRNA after I-day incubation of PBMNCs. Because we previously showed that GL cytolytic functions were significantly enhanced by 1-day preincubation of GLs?I4 we incubated PBMNCs of some patients for 1 day in RPMI 1640 medium containing 10% FCS and examined their cytolytic functions and per- forin mRNA levels. As expected, the cytolytic functions were enhanced in all patients tested (Table 2), whereas the levels

Northern blot analysis of perforin gene expression.

Table 1. Patient Characteristics and PBMNC Cytotoxicity

PBMNC Cytotoxicity$ GLS in Patient PBMNCs TCR-0 K562 K562 Daudi Daudi P815 K562 P815

NO. (%I Phenotype' Genet K562 a-CD3 Wf31 Daudi a-CD3 Wf31 MOLT-4 Raji P815 ADCC LDCC LDCC

1 2 3 4 5 6 7 8 9

10 11

Normal PB M "3 NK5 T§

98 95 90 86 95 77 91 82 93 90 73

12 97 3

CD3+4-8+16+ CD3+4-8+16+ CD3'4-8'16' C D 3 + 4 - 8 + 1 6 + CD3+4-8+16+ CD3'4-8'16' CD3'4-8-16' CD3+4+8-16+ CD3-4-8-16+ CD3-4-8-16 ' CD3-4-8-16 '

CD3-16+ CD3' 16-

R/R R/R GIR R/R G/R GIG R/R R I R G/G GIG G/G

- -

++++ ++ - -

++++ - -

+++ -

+ ++++

-

+++ +++

++++ +++ - -

++++ - -

+++ -

+ ++++

-

-

++++ +++

+ + -

++ - - +

+++

++ ++++

-

+++ +++

++++ ++++ ++++

+ ++++

-

-

+++ ++++

++ ++++

-

*In patients 1 through 6 and 8, Wf3 1 is positive; in patients 7 and 9 through 1 1, W3 1 is negative. tln TCR-fl gene analysis, R indicates monoclonal rearrangaments and G indicates germline configuration. $Levels of cytotoxicity of freshly iwlated PBMNCs are expressed as -(O% to 9%). + (10% to 29%). + + (30% to 49%). + + + (50% to 69%). and + + + + (70% to

$From a normal individual, whole PBMNCs, CD3-16' NK cells, and CD3'16- T lymphocytes were isolated and their cytotoxicity was assayed. 100%) at an effector-to-target ratio of 20:l. a-CD3 denotes antLCD3 MoAb treatment.

706 OSHlMl ET AL

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Per for in b

185 D

Fig 1. Northorn Mat aru)y.is of perforin mRNA 0bt.M from freshly hd.1.d PBMNCs. Lor).. 1 through 11 are from the patknts shown in fable 1, lane 12 is from a patient with C D 3 ' 4 ' 8 All.. brm 13 and 14 are from patkma with CD10'19'20' common ALL. and lane. 16 through 17 are from normal whole PBMNCs. NK ails. and T alls, rospaxtively, as ahown in Tabla 1. The pationt with ATL (bna 12) hod 2% GLs in PBMNCs; the patient with ALL (bne 131 had 100% lymphoblasts in PBMNCa, with 47% of them containing uveral prominent arurophilic granules in the cytoplasm; and t h . patient with ALL (lane 14) hod 0% GLs in PBMNCs. PBMNCs of these three patients exhibited leas than 6% of specific "Cr release in cytolytic assay.

of perforin mRNA were decreased in five of six patients tested (Fig 2). In patient 6. perforin mRNA level was increased.

DISCUSSION

Although perforin is considered responsible for cytolytic functions mediated by in vitro cultured CTLs and NK cells.' i t s role in vivo has not been delineated except in rat G L tumor'" because of the dificulty in isolating IO" to IO" cytolytic cells required for the biochemical studies.""

Recently. we' '-and other investigators" cloned the cDNA of perforin. The availability of a molecular probe for the perforin gene prompted us to investigate the levels of expression of this genc in freshly isolated cells. We now describe a correlation between perforin gene expression and cytolytic functions in PRMNCs of patients with GLPDr;.

In most patients with GLPDr;. in comparison with a normal individual. high percentages of GLs in PRMNCs were accompanied by high levels of perforin mRNA. In the

normal individual, enrichment of CD3 16' GLs was accom- panied by an increase in the level of perforin mRNA. These results strongly suggest that perforin mRNA and perforin protein are located mainly. i f not exclusively, in GLs. In a few GLPD patients (patients 6.8. and 9) and an A L L patient with azurophilic granules in lymphoblasts (patient 13). however. thc high SL percentages were accompanied by low levels of perforin mRNA. suggesting the existence of GLs containing only minimal. i f any. perforin mRNA and per- forin protein. Furthermore. in these patients. the levels of PRMNC cytotoxicity were low. Together. the levels of perforin mRNA correlated with those of PRMNC cytolytic functions rather than with G L percentages in PRMNCs. Pcrforin mRNA level. therefore. appears to be a good indicator for detection of levels of cytolytic functions.

In this study. cytolytic activities were induced by various procedures (Table I ) . DifTcrent levels of cytotoxicity were induced according to the methods used, and the levels of cytotoxicity d i fTcd from patient to patient. The degree of

PMmlt K582 K562 D u d D u d P815 K562 P815 No. K562 &D3 W 3 1 D u d &DO W 3 1 MOLT4 RaF P815 ADCC LDCC LDCC

- - - - - + + 1 & & - - + + + + 4 4. + + + - + + + - + + + + 6 + c * + - - + + - - + + + + - 7 + + - + + + + - +

10 c + + + + + + + + - .c + * 11 +

- -

- - - - - - - - - - - PBMNCa hom patients 1, 4. 6. 7, 10, end 11 were culNed for 1 day in RPMl 1640 medium containing 10% FCS. Cvotoxic activity ww then

wsaved. end the b w f s of cytotoxicity were canpcmd with those of freshly prepared PBMNCs. The inerewe in percentage cytotoxicity ww expresaed as - ( ~ 9 % ) . 4. (10% to 29%). end + 4. (30% to 49%).

PERFORIN MRNA IN U P D s

1 4 6 7 10

Perfori n4

Fig 2. Comparimm of perforin mRNA I.v.)s between freshly isolatod PBMNCs (I.*) and 1-day cultured PBMNCs (right). Numbers denote the patient number shown in Table 1 and Fig 1. In this experiment. RNase treatment was not dome.

cytotoxicity may be reflected by different factors. including the level of cytolytic molecules."" Recause the factor at the lowest level will limit cytolytic functions. the highest level of cytotoxicity observed in a particular patient implies at least such a level of cytolytic molecules in elfector lymphocytes. Thus, the correlation between the highest levels of cytotoxic- i ty and the levels of pcrforin mRNA can be explained by the hypothesis that pcrforin protein is a mediator for patient PRMYC cytolytic functions.

When the patient PRMNCs were cultured for 1 day. however, the levels of cytotoxicity were enhanced and those of pcrforin mRNA were decreased. As recently reported by Lichtenheld et al.'" interleukin 2 (Il,-2)-induced pcrforin m R N A expression and Il-2-induced lymphokine-activated killer ( L A K ) cell activity do not move in parallel; the peak of LAK activity was observed 12 hours later than that of pcrforin mRWA expression. and high levels of pcrforin appeared to be maintained even when pcrforin mRNA had decreased to a lower stcady-state level. This indicates that the reduced expression of pcrforin mRNA in our culture system docs not always imply a lower level of pcrforin protein contents. Recause changes in the amounts of pcrforin protein in a few cells cannot be monitored. we cannot rule out the possibility of enhanced levels of pcrforin protein in our I-day incubated PRMNCs. Sulica et al" and ltoh et aIm reported that N K activity of normal individuals and GL cytotoxicity of a patient with GLPD were enhanced simply by incubation of the cells. and this enhancement was considered to result from removal of inhibitory IgG present on the cell surface" or in the serum3 after in vitro culture. Therefore. the same phenomenon probably occurred in our culture system. An- other possibility that a pcrforin-indepcndent killing pathway was involved cannot be ruled out bccause this finding has recently been reported in CTLs.'."?'.-':-':

The decrease in pcrforin mRWA levels after incubation can be explained if IL-2 is responsible for pcrforin mRNA expression as has been reported by other investigators"'and if the expression of pcrforin mRNA in GLs depcnds on IL-2 present in the serum. Deprivation of 11,-2 in culture condi- tions may therefore decrease pcrforin mR%A levels. In patient 6. however. incubation of PRMNCs enhanced the pcrforin mRNA level. PRMNCs of this patient were hetero- geneous in surface phcnotypcs and contained a greater number of CD4' cells than did those of the other patients (CD4' cells 163; and CD8' cells 66%).'" The CD4' cells may have produced IL-2 in culture, thereby making CD8' GLs express greater amounts of pcrforin mRNA. Further studies are rquircd to test this hypothesis

The presence of perforin in normal human YK cells has been reported by Liu et al" and 7alman et al." To obtain large numbers of N K cells, Liu et al performed IO-day cultures of PRMNCs with irradiated Daudi cells. IL-2 probably was produced during such culturcs. Zalman et al. on the other hand, added IL-2 exogenously during 2- to 4-week culture periods. Rccuase IL-2 added in vitro induces pcrforin mRNA and pcrforin protein.'.""-': their studies did not ncccsarily indicate the presence of perforin in freshly isolated N K cells. Together with our finding of pcrforin mRNA expression in freshly isolated normal N K cells. however. this indicates that pcrforin is highly likely to be present in normal N K cells and to mediate XK lytic function.

ACKNOWLEDGMENT

We thank Drs Y. Kobayashi and F. lshikawa for gene rearrange- ments. Drs K. Toyama: M. Hattori. S. Sakamoto. M. Omine. S. hsano. S. Nishinarita. and Y. lizuka for providing their patients' blood samples, Drs G. Goldstein and P A T . Tetteroo for the gift of Mohbs. and H. Takasaki for preparing the manuscript.

REFERENCES

I . Oshimi K: Granular lymphocyte proliferative disorders: report of I 2 cascs and review of the literature. Leukemia 2 5 1 7. 1988 2. Reynolds CW. Foon K h : Ty-lymphoprolifcrative dimre and

related disorders in humans and experimental animals: A review of the clinical. cellular. and functional characteristics. Blood M:I146, I984 3. Loughran TP. Starkebaum G: Large granular lymphocyte

leukemia. Report of 38 cam and review of the literature. Medicine 66:397. 1987 4. Oshimi K. Oshimi Y. Akahoshi M. Kobayashi Y. Hirai H.

Takaku F, Hattori M. hsmno S. Kodo H. Nishinarita S. lizuka Y. Mi7oguchi H: Role of T-cell antigens in the cytolytic activities of large granular lymphocytes (LGLs) in patients with LGL lymphocy- tosis. Blood 71:473. 1988

OSHlMl ET AL

5. Henkart P Mechanism of lymphocyte-mediated cytotoxicity. Annu Rev Immunol3:31,1985

6. Young JD-E, Cohn ZA: Cellular and humoral mechanisms of cytotoxicity: Structural and functional analogies. Adv Immunol 41:269, 1987

7. Clark W, Ostergaard H, Gorman K, Torbett B: Molecular mechanisms of CTL-mediated lysis: A cellular perspective. Immu- no1 Rev 103:37,1988

8. Young JD-E, Liu C-C, Persechini PM, Cohn ZA: Perforin- dependent and -independent pathways of cytotoxicity mediated by lymphocytes. Immunol Rev 103:161, 1988

9. Shinkai Y, Takio K, Okumura K: Homology of perforin to the ninth component of complement (C9). Nature 334525, 1988

10. Lichtenheld MG, Olsen KJ, Lu P, Lowrey DM, Hameed A, Hengartner H, Podack ER: Structure and function of human perforin. Nature 335:448, 1988

11. Spits H, Borst J, Tax W, Capel PJA, Terhorst C, de Vries JE: Characteristics of a monoclonal antibody (WT3 1) that recognizes a common epitope on the human T cell receptor for antigen. J Immunol135:1922,1985

12. Van de Griend RJ, Borst J, Tax WJM, Bolhuis RLH: Functional reactivity of WT31 monoclonal antibody with T cell receptor-yexpressingCD3+4-8- Tcells. J Immunol 1401 107,1988

13. Oshimi K, Oshimi Y, Akutsu M, Takei Y, Saito H, Okuda M, Mizoguchi H: Cytotoxicity of interleukin 2-activated lymphocytes for leukemia and lymphoma cells. Blood 68:938, 1986

14. Oshimi K, Hoshino S, Takahashi M, Akahoshi M, Saito H, Kobayashi Y, Hirai H, Takaku F, Yahagi N, Oshimi Y, Horie Y, Mizoguchi H: Ti (WT31)-negative, CD3-positive. large granular lymphocyte leukemia with nonspecific cytotoxicity. Blood 71:923, 1988

15. Oshimi K, Oshimi Y, Satake M, Mizoguchi H: Natural killer-mediated lysis of normal and malignant target cells, and its regulation by monocytes. J Exp Med 162:472, 1985

16. Maniatis T, Fritsch EF, Sambrook J: Molecular Cloning. A Laboratory Manual. Cold Spring Harbor, Cold Spring Harbor Laboratory, 1982, p 187

17. Shinkai Y, Yoshida MC, Maeda K, Kobata T, Maruyama K, Yodoi J, Yagita H, Okumura K: Molecular cloning and chromo- somal assignment of a human perforin (PFP) gene. Immunogenetics (in press)

18. Henkart PA, Millard PJ, Reynolds CW, Henkart MP: Cytolytic activity of purified cytoplasmic granules from cytotoxic rat large granular lymphocyte tumors. J Exp Med 160:75, 1984

19. Sulica A, Gherman M, Galatiuc C, Manciulea M, Herber- man RB: In vitro enhancement of peripheral blood mononuclear cell natural killer activity following short-term incubation with fetal calf serum. J Immunol 128:1031,1982

20. Itoh K, Tsuchikawa K, Awataguchi T, Shiiba K, Kumagai K: A case of chronic lymphocytic leukemia with properties characteris- tic of natural killer cells. Blood 61:940, 1983

21. Liu C-C, Steffen M, King F, Young JD-E: Identification, isolation, and characterization of a novel cytotoxin in murine cytolytic lymphocytes. Cell 51:393, 1987

22. Berke G, Rosen D: Highly lytic in vivo primed cytolytic T lymphocytes devoid of lytic granules and BLT-esterase activity acquire these constituents in the presence of T cell growth factors upon blast transformation in vitro. J Immunol 141:1429, 1988

23. Liu C-C, Perussia B, Cohn ZA, Young JD-E Identification and characterization of a pore-forming protein of human peripheral blood natural killer cells. J Exp Med 164:2061, 1986

24. Zalman LS, Brothers MA, Chiu J, Muller-Eberhard HJ: Mechanism of cytotoxicity of human large granular lymphocytes: Relationship of the cytotoxic lymphocyte protein to the ninth component (C9) of human complement. Proc Natl Acad Sci USA 83:5262, 1986