Cancer Immunol Immunother (1987) 24:72-75 ancer mmunolggy mmunotherapy

© Springer-Verlag 1987

Modulation of natural killer cell activity in stage I postmenopausal breast cancer patients on low-dose aminoglutethimide

John Berry 1, Bridget J. Green 2, and David S. Matheson 2

~ Department Medicine, Medical Oncology, Tom Baker Cancer Centre, 1331 -29th St. N.W., Calgary, Alberta, T2N 4N2 2 Department Paediatrics, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada, T2N 4N1

Summary. Natural killer (NK) cells are important in sur- veillance against malignant cells. The activity of NK cells can be modulated by naturally occurring mediators; inter- feron, interleukin-2, and hormones. Low-dose aminoglu- tethimide (Ag 250 mg/day) inhibits the peripheral aroma- tization of androstenedione hence decreasing circulating estrogens. Of ten patients treated, seven were evaluable. There was a statistically significant increase in NK activity (P = 0.0025) following the administration of Ag. There was no consistent shift in N K cell number (Leu-l lb positive cells). In vitro Ag did not alter N K activity whereas 17-[~- estradiol did. These data are consistent with an indirect ef- fect of Ag on NK activity. Hence in vivo Ag which causes a reduction in serum estrogens in postmenopausal pa- tients, also induces an increase in NK activity.

Introduction

Natural Killer (NK) cells may have a role in the protection against malignant or virally infected cells [13], and the de- gree of NK cytotoxicity is modulated by endogenous me- diators including hormones. In vitro estrogen acts as an in- hibitor of human NK cell cytolysis [3]. Postmenopausal patients have a higher N K activity than their premeno- pausal counterparts, and there is a variation in premeno- pausal NK activity relative to the menstrual cycle, being lower in the first phase and increasing to normal levels af- ter ovulation suggesting an inverse correlation of NK ac- tivity with circulating estrogen levels [20].

There is evidence that breast cancer patients have low- er NK activity compared to controls [1], and patients with breast cancer who have reduced N K activity are more pro- ne to disease recurrence [9]. Aminoglutethimide (Ag) is a proven hormonal therapy in postmenopausal breast can- cer [5, 6]. It acts at two levels; it inhibits pregnanalone for- mation from cholesterol in the adrenal gland, and it blocks the peripheral aromatization of androgenic precursors to estrogens [15, 16]. The latter mechanism would seem more important in postmenopausal women as adrenal and ovar- ian secretion of estrone and estradiol are insignificant in terms of total estrogen production [4]. It has been shown that low-dose Ag can reduce serum estrogen levels without corresponding adrenal suppression [2, 7]. We report here the effect of low-dose Ag in vivo and its modulation of NK activity in postmenopausal patients.

Offprint requests to: John Berry

Patients and methods

Ten consenting postmenopausal stage I breast cancer pa- tients, mean age 57 years (range 47-69), who had not re- ceived any therapy (radiotherapy or chemotherapy) in the preceding 3-month period, entered into the study. The Ag (125 mg) was given twice daily by mouth from week 0 to 4. Blood was sampled at weeks 0, 4, and 6, for analysis of the following; NK activity, cell marker studies, estradiol, an- drostenedione, thyroid stimulating hormone, thyroxin, prolactin, complete blood count and differential.

Peripheral blood mononuclear (PBM) cells were ob- tained by density centrifugation of heparinized blood over Ficoll-Paque (Pharmacia, Montreal, Quebec). K-562, an erythroleukemia line, was maintained in continous culture of minimum essential medium (MEM)-alpha and 10% fe- tal calf serum (Flow Laboratories, Mississauga, Ontario) [8]. A NK assay was performed by incubating 10 4

51chromium (Amersham, Montreal, Quebec)-labeled tar- get cells (K-562) with varying numbers of PBM effector cells for 18 h at 37 °C, in a 5% CO 2 air atmosphere in MEM-alpha medium as described previously [10]. The as- say was performed at effector-to-target cell ratios of 30:1, 10:l, 3:1, and 1:1. One lytic unit was defined as the num- ber of cells required to achieve a particular level of lysis, and the number of lytic units per 10 7 effector cells was de- termined. Multiple predrug N K levels were not performed as each patient acted as her own control and, except for intercurrent illness, N K levels are relatively stable for each individual ([11] and personal observations).

In order to assess N K function in vitro, purified Ag (CIBA Pharmaceuticals), or purified 17-[3-estradiol (Stera- loids Inc., Wilton, NH) was dissolved in acetone and then diluted in medium before incubating with 2 x 10 6 effector cells for 2 h in 0.6 ml of medium. Cells were then dispersed into microtiter plates containing 10 4 labeled K-562 cells for an 18-h chromium release assay.

Cell surface markers were assessed, using appropriate monoclonal antibodies, to determine the precentage of OKT3 (T cells), sIg (B cells) and Leu-l lb positive (NK cells) [21].

Results

Two patients were withdrawn from the study due to devel- opment of skin rash and lethargy during drug therapy. An- other patient was found, on hormonal evaluation, to be

73

cycling with respect to her estrogen levels and was there- fore excluded from analysis.

The percentage lysis at 10:1 effector to target ratio is shown in Table 1. Clearly each pat ient had a normal N K activity pr ior to the ingest ion of Ag and, in all cases, the percentage lysis increased during drug therapy. Except for patients 6 and 7, the percentage lysis d ropped to near the predrug levels by 2 weeks after the drug was discontiuned. By pa i red t-test statistics there was a statistically signifi- cant difference between the pre and during (P=0.0025) and between the during and post (P = 0.025) lysis values.

Perhaps a more accepted means of expressing the data is to convert the lysis curves to lytic units/107 effectors cell [12]. This was done and then each pat ient ' s predrug level was set at 100%. Subsequent lytic values were then ex- pressed as a percentage of the predrug value for each pa- tient. Hence, each pat ient acted as her own control.

As shown above, low-dose Ag enhanced N K activity (Fig. 1). Statistical analysis, per formed by pa i red t-test (P = 0.0025) indica ted a significant increase in N K activity during drug therapy compared to baseline.

Hormone analysis did not confirm the abil i ty of low- dose Ag to decrease serum estrogen levels since 40 pmol/1 was the lower limit of our assay. However, for all patients whose estrogen levels were above 40 pmol /1 pr ior to the drug adminis t ra t ion, the levels d ropped to less than 40 pmol /1 on taking the drug. Some patients had serum es- trogen levels less than 40 pmol /1 pr ior to drug adminis t ra- t ion, and al though none increased to over 40 pmol /1 dur- ing the Ag therapy, we cannot confi rm that their estrogens decreased further. The serum andros tendione increased as expected in all pat ients [2, 7]. There was no effect on the levels of thyroxin, thyroid st imulat ing hormone, prolact in, white b lood cell count or differential during the course of this short s tudy (data not shown).

Cell surface markers were analyzed in four patients in whom enough lymphocytes were retrieved. There was no consistent shift in L e u - l l b posit ive cells during the per iod of drug adminis t ra t ion compared to the drug-free mea- surements (Table 2). Similar ly the number of OKT-3 and sIg posit ive cells were unchanged (data not shown).

Ag had no significant effect on the N K lytic activity of human lymphocytes in vitro (Table 3), whereas estrogen reduced N K lysis (Table 4). The Ag at 10 -5 M decreased N K function by less than 30%, but estrogen decreased N K function by aproximate ly 50% at 10-7M. Control assess- ment with acetone did not show any influence on basel ine chromium release.

3 0 0

z

0

0 v,

I.i. 0 uJ 0

2 0 0

1 0 0 Z U.I U n,- U.I a. * N.S.,

I I I Pre During Pos t

A M I N O G L U T E T H I M I D E

Fig. 1. NK cell function was assessed before, at the end of 4 weeks of oral Ag administration, and 2 weeks after the Ag was discontin- ued. All NK data prior to the drug therapy were then set at 100% and other NK values were adjusted to that predrug level for each patient; i.-e., each patient's predrug NK value acts as the base- line from which subsequent NK determinations deviate. The verti- cal bars represent mean and standard error of the mean. The as- terisk (*) indicates that there was a significant (P=0.0025) in- crease during the drug therapy over the pretherapy baseline and N.-S. indicates that there was no significant difference in NK le- vel after 2 weeks off Ag compared to predrug values. The patient with the highest percentage of background in the postdrug inges- tio, assessment developed an upper respiratory tract infection 1 day after the blood sample had been obtained

Table 2. Effect of Ag in vivo on NK phenotype

Patient Pre During Post

1 14 11 10 2 10 12 14 3 9 10 10 4 10 12 13

(Results as%Leu- l lb positive cells)

Table 1. NK activity at 10:1 effector to target ratio

Patient Pre During Post

1 26 59 45 2 48 65 45 3 50 63 45 4 47 58 44 5 43 62 39 6 23 30 35 7 42 55 56 Mean (±SD) 40(±10) 56(±11) 44(±6)

Results are given as percent lysis for 10:1 effector to target ratio. The mean percentage lysis (and SD) for 30 different normal female controls in our laboratory is 37 (± 11)

Table 3. Effect of Ag in vitro on NK cell activity

Ag (M) Expt 1 Expt 2

10 -4 34 37 10 -s 114 92 10 -6 141 93 10 7 143 103 10 -8 160 90 l0 -9 148 90 0 158 88

(Results in lytic units/107 cells)

Background lysis of the labeled target cells in the absence of effec- tors was not altered by any of the Ag concentrations assayed

74

Table 4. Effect of in vitro 17-~-Estradiol on NK cell activity

17-[3-Estradiol (M) Expt 1 Expt 2

10 -6 44 N.T. 10 -7 72 63 10 -8 95 88 10 -9 152 91 10 -1° 148 93 0 150 102

(Results are expressed aslytic units 107 PBMCat30%lysis)

Background lysis of labeled targed cells in the absence of effector cells was not altered by the 17-~-estradiol. N.T. = not tested

Discussion

Our da ta are consistent with low-dose Ag reducing ciculat- ing estrogen levels in pos tmenopausa l patients and sup- por t the concept that estrogen may inhibit N K activity. Previous investigators have documented that estrogen in vivo could inhibit murine N K activity albeit in supraphy- siological doses [17]. Others have suggested that the varia- t ion in N K activity in menstruat ing women, most pro- nounced during the per iovula tory period, is related to f luctuation in the levels of lutenizing hormone rather than estrogens [18, t9]. However, Ag does not influence the se- cretion of either lutenizing or follicle st imulating hormone by the pi tui tary [14]. Hence, it is unlikely that the modula- t ion of N K activity seen could be ascribed to a f luctuation in pi tui tary hormones in this pos tmenopausa l group of pa- tients. We were unable to rule out the effect of increased androgeneic precursors on N K activity al though the in vi- vo and in vitro data per ta in ing to estrogens are consistent with estrogenic modulat ion.

Ag probab ly does not directly modula te N K activity as there was no effect on N K activity in vitro. However , N K activity was inhibi ted by pharmacologica l (10-7M), if not physiological (10-9M), doses of estrogens in vitro. The down regulat ion of N K activity seen in vivo by even the low levels of estrogens in pos tmenopausa l patients could be modula ted by decreasing circulating free estrogens with low-dose Ag. This p roduced a doubl ing of basel ine N K ac- tivity. There was no evidence of an increased number of cells with N K cell phenotypes (Leu-1 lb positive) responsi- ble for the increase in N K activity secondary to the Ag. Our results are therefore consistent with the hypothesis that decreasing the level of estrogens augments N K activi- ty but p roo f of that mechanism requires further study.

Unfor tunate ly despite using low-dose Ag there was a difficulty with side effects and this contr ibuted to the small number of patients that completed this trial. This would restrict the value of low-dose Ag as a biological response modif ier in pos tmenopausa l breast cancer patients.

More extensive studies are clearly needed to confirm the modula t ion of N K activity by hormone manipula t ion. It is not certain whether N K cells have any direct signifi- cance for the host tumor relat ionship. Should this be the case, then data such as ours could provide an insight into the beneficial effects of hormone therapy. In addi t ion, de- l iberate at tempts to modify N K activity may be indica ted and measurement of this parameter could be an impor tant component in clinical studies of endocr ine therapy for breast cancer.

Acknowledgements. This work is supported in part by the Alberta Children's Hospital Foundation and, the Alberta Cancer Board. DSM is a scholar of the Alberta Heritage Foundation for medical research. We wish to thank Dr. Y. Lefebre for providing the 17-[~- estradiol and reviewing the manuscript with Drs. D. A. Hart and J. Miller.

References

1. Cunningham-Rundels R, Fillipa DA, Braun Jr DW, Antonelli A, Ashikari H (1981) Natural cytotoxity of peripheral blood lymphocytes and regional lymph node, cells in breast cancer in women. J Natl Cancer Inst 67 (3): 585

2. Dowsett M, Santer S J, Santan RJ, Jeffcoate SL, Smith TE 0985) Effective inhibition by low-dose aminoglutethimide of peripheral aromatization in post-menopausal breast cancer patients. Br J Cancer 52:31

3. Ferguson MM, McDonald FG (1985) Oestrogen as an inhibi- tor of human NK cell cytolysis. FEBS 191 (1): 145

4. Grondin JM, Siiteri BK, MacDonald PC (1972) Modes of es- trogen production in post-menopausal women. J Clin Endoc- rinol Metab 36 (2): 207

5. Harris AL, Dowsett M, Smith IE, Jeffcoate SL (1983) Amino- glutethimide induced hormone suppression in response to therapy in advanced post-menopausal breast cancer. Br J Cancer 48:585

6. Harris AL, Powles TJ, Smith IE, Coombes RD, Ford HT, Gazet JC, Harmer CL, Morgan M, White H, Parsons CA, McKinna JA (1983) Aminoglutethimide for the treatment of advanced post-menopausal breast cancer. Eur J Cancer Clin Oncol 19:11

7. Harris AL, Dowsett M, Smith IE, Jeffcoate SL (1983) Endo- crine effects of low-dose aminoglutethimide alone in adv- anced post-menopausal breast cancer. Br J Cancer 47:621

8. Lozzio CV, Lozzio BB (1975) Human chronic myelogenous leukemia cell line with positive Philadelphia chromosome. Blood 45:321

9. Mackay IR, Goodyear MDE, Riglar C, Penschow J (1983) Ef- fect on natural killer cell and antibody dependant cellulocyto- toxicity of adjuvant cytotoxic chemotherapy including mel- phalan in breast cancer. Cancer Immunol Immunother 16:98

10. Matheson DS, Wilson RW, Green B (1982) Trimethoprim-in- duced augmentation of human natural killer cell activity in vitro. Immunopharmacology 4:73

11. Pross HF, Baines MG 0982) Studies of human natural killer cell in vivo parameters affecting normal cytotoxic function. Int J Cancer 29:383

12. Pross HF, Baines MG, Rubin P, Shragge P, Patterson MS (1981) Spontaneous human lymphocytic mediated cytotoxi- city. J Clin Immunol 1 : 51

13. Roder JC, Haliotis T (1980) Do natural killer cells play a role in anti-tumor surveillance? Immunol Today 1 : 96

14. Santen RJ, Wells SA (1980) The use of aminoglutethimide in the treatment of patients with metastatic carcinoma of the breast Cancer 46:1066

15. Santen RJ, Samoklik E, Lipton A, Harvey H, Ruby EB, Wells SA, Kendall J (1977) Kinetic hormonal and clinical studies with aminoglutethimide in breast cancer. Cancer 39:2948

16. Santen RJ, Santer S, Davis B, Veldhius J, Samojlik E, Ruby E (1978) Aminoglutethimide inhibits extraglandular estrogen production in post-menopausal women with breast carcino- ma. J Clin Endocrinol Metab 47:1257

17. Seaman W, Blackman MA, Gindhash TD, Roubinian JR, Loeb JM, Talal N (1978) Beta estradiol reduces natural killer cells in mice. J Immunol 121 : 2193

18. Sulke AN, Jones DB, Wood PJ (1985) Variation in natural killer activity in peripheral blood during the menstrual cycle. Br Med J 290:884

75

19. Sulke AN, Jones DB, Wood PJ (1985) Hormonal modulation of tiuman natural killer cell activity in vitro. J Reprod Immu- nol 7:105

20. White D, Jones DB, Cooke P, Kirkham N (1982) Natural kill- er (NK) activity in perhipheral blood lymphocytes of patients with benign and malignant breast disease. Br J Cancer 46:611

21. Zipf TF, Lauzon GJ, Longenecker BM (1983) A monoclonal antibody detects a 39,000 mw molecule that is present on B lymphocyted and chronic lymphocytic leukemia cells but is rare on acute lymphocytic leukemia blasts. J Imrnunol 131: 3064

Received December 30, 1985/Accepted August 20, 1986