~

Cancer Immunol Immunother (1991) 34: 49- 52 0340700491000955

ancer mmunolggy mmunotherapy

© Springer-Verlag 1991

Production of tumor necrosis factor ~z and interferon ], in interleukin-2-treated melanoma patients: eorrelation with elinieal toxieity* James S. Economou 1, Mary Hoban 1, Jeffrey D. Lee 1, Richard Essner 1, Stephen Swisher 1, William McBride 2, Dave B. Hoon 1, and Donald L. Mortonl 1 Division of Surgical Oncology, Department of Surgery, John Wayne Cancer Clinic, and 2 Department of Radiation Oncology, Jonsson Comprehensive Cancer Center, UCLA Medical Center, Los Angeles, CA 90024, USA Received 31 December 1990/Accepted 9 May 1991

Summary. Interleukin-2 (IL-2)-based immunotherapy regimens are accompanied by dose-limiting toxicity consisting o f fever, tachycardia, chills and capillary leak syndrome. We hypothesized that the toxicity was caused by the induction and release of endogenous cytokines such as tumor necrosis factor ~ (TNFc0 and interferon "/(IFNy). W e measured the serum levels of TNFc~ and IFN 7 in IL-2treated m e l a n o m a patients and attempted a correlation with clinical toxicity. A total o f 23 patients received either 6 × 106 IU or 12 × l06 IU Cetus IL-2/m 2 by i. v. bolus daily for 5 consecutive days on weeks 1, 3 and 5. Serum T N F ~ and 1FN7 levels were measured by enzyme-linked immunosorbent assay. Clinical toxicity was scored each day by objective measurements of hypotension, tachycardia, fever and chills/rigors. Clinical toxicity and IFN7 levels correlated nicely, peaking on the 5th day o f each treatment cycle. The kinetics and magnitude of TNFcz production, however, were not predictable and did not correlate with either I F N 7 o r toxicity. Some patients had modest increases in TNFc~ production while others had markedly increased levels during the second and third treatment weeks. Remarkably, these high levels persisted during nontreatment weeks and after completion of therapy. This clinical study demonstrates novel kinetics for immunoreactive TNF~x in IL-2 cancer patients, which do not correlate well with toxicity.

Introduction Interleukin-2 (IL-2)-based immunotherapy regimens are associated with dose-limiting toxicity, which gives this biological a suboptimal therapeutic index when used as a single agent [14, 23]. It is n o w well recognized that IL-2 induces the production of a variety of cytokines by lymphocytes and macrophages, both in vivo and in vitro [2, 7]. These include I L - l , tumor necrosis factor c~ (TNFc0, TNF]3 (lymphotoxin), interferon 7 (IFNy) and IL-6. M a n y of the toxic side-effects o f IL-2 administration are reminiscent o f those observed in sepsis and quite similar to those associated with TNFc~, IL-6 and IL-1 [7, 13]. Thus, the hypothesis has been proposed that one or more of these secondarily produced cytokines plays a role in IL-2 toxicity [7]. W e elected to study the kinetics o f TNFcc and I F N 7 production in IL-2--treated m e l a n o m a patients to determine whether they correlated with clinical toxicity.

Materials and methods Study design. Prior approval for this protocol was obtained from the

* This work was supported by NIH Grants CA 50780 (J. E.) and CA 29 605, CA 12582 (D. L. M.) and the U. C. Tobacco-Related Disease Research Program RT-62 (J. E.). J. E. is the recipient of an NCI Clinical Investigator Award (KO8-01360) and is a Dorothy and Leonard Straus Scholar at UCLA

UCLA Human Subjects Protection Committee (no. 88-08-351) and the Food and Drug Administration (no. 32 869), and informed consent was obtained from patients. Stage II and III melanoma patients were given systemic Cetus (Emeryville, Calif.) recombinant human IL-2 and melanoma tumor cell vaccine in the following regimen. IL-2 was given by bolus intravenous infusion once daily, Monday through Friday on weeks 1, 3 and 5. One group of patients received 6 x 106 IU/m2 and the second received 12x 106 IU]m2 with each infusion. Patients also received an intradermal inoculation of an irradiated allogeneic melanoma cell .vaccine on the Wednesdays of weeks 1, 3, 5, 9, 13 and 17. The composition and antigenic characteristics of this vaccine and the immunotherapy regimen have been described [10]. Patients were treated as outpatients in the John Wayne Clinical Research Center, Division of Surgical Oncology, UCLA Medical Center. The toxic side-effects of IL-2 treatment were treated with intravenous hydration, acetaminophen or meperidine. Three patients (all receiving the higher IL-2 dose regimen) required a single overnight hospital admission for management of tachycardia and hypotension.

Offprint requests to: J. S. Economou, Division of Surgical Oncology, University of California, Los Angeles, 10833 LeConte Avenue, Los Angeles, CA 90024-1782, USA

Cytokine assays. Human TNF~ and IFNy levels in patient sera were measured using commercially available assays. Patients' blood was drawn 2 h after cytokine infusion. The TNF« assay is an enzyme-linked

Key words: IL-2 - Toxicity

50 immunosorbent (ELISA) assay (Endogen, Boston, Mass.), adapted to Terasaki plates with a sensitivity of 12 pg/ml. The IFN7 assay is a radioimmunoassay (Centocor, Malvern, Pa.) with a sensitivity of 0.5 U/ml. Patient sera were collected in heparinized tubes, separated and stored at -70 ° C. TNF is stable at -70 °C for months and can undergo two to three freeze/thaw cycles before losing activity. All assays were performed on samples thawed only once.

o >

21000

18000

15000

~~" 12000

Clinical toxicity. A clinical toxicity score was calculated daily for each patient and based upon objective measurements of hypotension, tachycardia, fever and chills/rigors. Each of these four parameters was given a score of 0, 1, 2 or 3 according to the following parameters: (a) hypotension: 0, no change; 1, >20 mm Hg (2.7 kPa) reduction in baseline systolic pressure; 2, >30 mm Hg (4 kPa); 3, >40 mm Hg (5.3 kPa); (b) tachycardia: 0, no change; 1, 100-120 beats/min; 2, 121-140 beats/min, 3, 141-160 beats/min; (c)fever: 0, no change; 100-101°F (37.838.3 °C); 2, 101 - 103°F (38.2- 39.4°C); 3, 103 - 105°F (39.4-40.6 °C); and (d) chills/rigors: 0, none: 1, mild; 2, moderate; 3, severe requiring meperidine. Each of these four parameter scores was totalled each day to give a composite clinical toxicity score that could range from 0 to 12.

E

go00 U. I--

3000

IL-2

IL-2

WEEK

>

1

2

3

IL-2 4

5

6

7

/v'v'V~ IL-2 5

.........

8

9

21000-

18000-

Results Groups of melanoma patients received either 6 x 10 o I U / m 2 or 12 x 106 I U / m 2 IL-2 for 5 consecutive days on weeks 1, 3 and 5 in this protokol. A total o f 23 patients was treated. These included patients with either stage II (nodal disease completely resected) or stage III disease (distant disease resected or unresected). W e obtained serial serum samples from approximately half of these patients and measured cytokine levels. W e elected to use either E L I S A or radioimmunoassays because of our concern for the specificity of bioassays. Serum TNFo~ levels were serially measured over an 8-week period. Patients receiving 6 x 106 IU/m2 (Fig. l A ) or 12 x l06 I U / m 2 (Fig. 1B) had a variable T N F response. Several patients in each group had relatively modest increases in T N F (patients 3, 4, 5, 6, 10) while others had very markedly increased levels. Patients who produced TNFc~ had small increases towards the end of the first IL-2 treatment week, but significant increases in the second and third IL-2 treatment weeks. Quite remarkably, these high levels frequently persisted over a several-week period even after some patients had completed IL-2 treatments. A greater number of patients appeared to have elevated T N F levels in the higher IL-2 treatment dose (4/5 vs 2/6). W e did not detect any elevation of T N F levels in 10 m e l a n o m a patients undergoing treatment with the vaccine alone (not shown). I F N 7 levels were measured in the same serum samples. As shown for groups of patients receiving 6 x 106 IU IL-2/m2 (Fig. 2 A ) and 12 x 106 I U / m 2 (Fig. 2B), increases in I F N 7 serum levels were quite predictable. All patients produced IFN7 with peak levels on the 5th treatment day. I F N 7 levels were undetectable in between and after treatment cycles. There appeared to be no correlation between the magnitude o f the I F N y responses and T N F responses and no real differences between the two IL-2 dosage groups. The clinical toxicities o f this IL-2 regimen include fever, chills, rigors, hypotension, fatigue, tachycardia and malaise. Several o f the more c o m m o n toxicities were selected to generate a daily "toxicity score". This is graphi-

6000

15000-

E

tl Z I.-,-

12000.

9000'

6000.

3000,

0 ' Nk, M~ IL-2 WEEK 1

.........

2

tvVv~ IL-2 3

.........

4

6

7

8

9

Fig. 1. Tumor necrosis factor a (TNF) production in melanoma patients receiving 6 × 106 IU IL-2/m~ day-1 (A) or 12 x 106 IU IL-2/m-2 day 1 (B). Patient numbers: [], 1; O, 2; g , 3; I~, 4; B, 5; n, 6; A, 7; O, 8; x, 9; [], 10; A, ll

cally presented for each of the 11 patients studied (Figs. 3 A, B). These toxic side-effects were more severe towards the end of each IL-2 treatment cycle. Toxicity was cumulative, being somewhat more severe in the second and third treatment cycles. Also, clinical toxicities were transieht and resolved within 24 h. The magnitude of clinical toxicity did not correlate with the magnitude of T N F production.

Discussion The clinical toxicity o f IL-2 in humans is cumulative and dose-related. Effects include fever, chills, hypotension and capillary leak syndrome. The exact causes of this toxicity are not known, but are probably due to the induction and release of secondary mediators. W e previously demonstrated that peripheral blood leukocytes (PBL) from IL-2-treated cancer patients and mice produced greatly exaggerated quantities of T N F [7, 15]. Peak in vivo production required several days of IL-2 infusion. H u m a n T cells, B cells and peripheral monocytes can be directly stimulated by IL-2 to produce TNFc~, m R N A and protein [24, 25]. Since the biological effects of

51 10

20

g ~6 ~:

~!2a

~0

I-| Z IJ.

¢9

5

"7

201 WEEK

IL-2 1

2

IL-2 3

4

IL-2 5

6

7

8

9

WEEK

IL-2 1

2

IL-2 3

4

IL-2 5

6

7

8

9

10"

15

~

10'

è. ¢3

AA. . . . . . . . . ~

o

IL-2 WEEK 1

2

IL-2 3

4

......... IL-2 5

6

7

8

• "AA~A. . . . . . . . . 9

WEEK

IL-2 1

2

AA,q~A. . . . . . . . . ~ IL-2 3

4

AA. . . . . . . .

-

6

9

IL-2 5

7

8

Fig. 2. Interferon 7 (1FN7) production in melanoma patients receiving 6 x 106 IU IL-2 m-2 day-1 (A) or 12 x 106 IU IL-2 m-2 day I (B). Patient numbers identified as in Fig. 1

Fig. 3. Daily toxicity score in melanoma patients receiving 6 x 10 6 IU IL-2 m-2 day-1 (A) or 12 x 106 IU IL-2 m-2 day-t (B). Patient numbers: [],1; O, 2; n, 3; I>, 4; æ,5;[~,6; A,7;O,8; x,9; [],10; A, 11

TNF administration so closely resemble those of IL-2 administration, TNF is a candidate as a mediator of IL-2 toxicity. This hypothesis is supported by Fraker et al. [9], who demonstrated that the toxic effects of IL-2 in mice were partially ameliorated by an anti-murine TNF antibody. However, a number of other cytokines and mediators - IL-6, IL-l, IFN 7, thromboxane A2 - are induced by IL-2 in vitro and in vivo and could also contribute to toxicity [5, 11, 16]. For these reasons we elected to study the induction of two candidate cytokines - TNFo~ and IFNy. Our results demonstrating elevated levels of TNFc~ in the serum of IL-2-treated cancer patients confirm and expand upon our earlier findings and those of others [6, 7]. The kinetics of TNF production are interesting and have not been previously described. Patients were treated for 5 consecutive days with 6 x 10 6 or 12 x 106 IU m -2 day -1 on weeks 1, 3 and 5. During week 1, we did not observe increases in TNF serum levels until days 4 or 5. This several-day delay is similar to our findings in mice [7, 15]. Although the kinetics of IL-2-induced TNF production in vitro is quite a bit more rapid (peak m R N A at 6 h, peak protein at 1 2 - 1 8 h), our blood samplings, timed several hours after IL-2 infusion, did not detect possible transient TNF elevations. Interestingly, Michie [18] also did not

observe elevations in TNF shortly after an infusion of IL-2. We occasionally observed elevated TNF levels during the first non-treatment week. By the second IL-2 treatment week, several patients had earlier and more exaggerated TNF responses, suggesting that PBL were "primed" during the first week of treatment. Several patients had very pronounced levels of TNF that persisted into non-treatment weeks. Of great interest was out observation that a number of patients had markedly elevated levels of TNF several weeks after their last IL-2 treatment. The production of IFN7 by IL-2-treated melanoma patients was far more predictable. All patients produced IFNy towards the end of each IL-2 treatment week and IFN7was not detectable during non-treatment weeks. Also, in contrast to TNF production, there was no apparent "priming" effect generating greater IFN 7 levels with subsequent IL-2 doses. The clinical toxicity experienced by our patients was fairly predictable. Side-effects were more severe on days 4 and 5 of each treatment week, increasing during later treatment weeks and with larger doses of IL-2. The most common side-effects were chills, fever, rigors, hypotension and tachycardia. These largely resolved within 24 h of cessation of treatment. There was certainly no clear-cut correla-

52 tion b e t w e e n T N F levels for a particular patient and their toxicity score. Patients with m o d e s t elevations in T N F were not spared f r o m significant side-effects. M o r e o v e r , T N F levels persisted, in s o m e cases for weeks, long after I L - 2 - r e l a t e d effects had resolved. These observations certainly raise the question w h e t h e r this is b i o l o g i c a l l y active T N F [22]. T N F appears to p l a y a m a j o r role in the pathogenesis o f g r a m - n e g a t i v e shock [20] and induces fever, m e t a b o l i c acidosis, diarrhea, h y p o t e n s i o n and d i s s e m i nated intravascular c o a g u l a t i o n [8]. T h e s e observations are even m o r e curious since the half-life o f infused T N F is quite short. Certainly, the induction o f T N F in h u m a n s b y e n d o t o x i n or O K T 3 (in transplantation) is transient [4, 19]. H o w e v e r , in these reports, there was not a repetitive induction signal being d e l i v e r e d to the patient. A l l patients prod u c e d I F N y and these kinetics m o r e closely a p p r o x i m a t e d toxicity. The absolute m a g n i t u d e o f I F N 7 levels did not, h o w e v e r , predict the severity o f side-effects. These results d e m o n s t r a t e clear differences in the kinetics o f TNFc~ and I F N 7 generation in I L - 2 - t r e a t e d m e l a n o m a patients. Since the s a m e s e r u m s a m p l e was used for both assays, the spectrum o f responses for T N F cannot be attributed to m a j o r differences in s a m p l e handling and storage since the I F N 7 r e s p o n s e was so predictable. Clinical toxicity correlated better with I F N 7 p r o d u c t i o n than with TNFo~. It is difficult to e x p l a i n persistent levels o f i m m u n o c h e m i c a l l y reactive T N F during non-treatm e n t w e e k s and in patients w h o were not clinically ill. It is p o s s i b l e that this reactive species is not b i o l o g i c a l l y active [22]. The induction o f TNFc~ and I F N 7 b y I L - 2 in v i v o m a y have therapeutic implications. B l a y et al. [3] have suggested that I L - 2 - t r e a t e d cancer patients w h o generate higher levels o f T N F e x p e r i e n c e bettet t u m o r r e s p o n s e rates. Also, there is c o n s i d e r a b l e interest in c o m b i n i n g b i o l o g i c a l s in the treatment o f h u m a n cancers. Preclinical a n i m a l studies have shown clear benefits o f various c o m b i nations o f T N F c q I F N 7, IFNc~A and IL-2 [17, 21]. The p r e d i c t a b l e and in s o m e cases p r o n o u n c e d induction b y I L - 2 o f TNFc~, I F N 7, I L - l , I L - ó and p o s s i b l y other cytokines should be carefully c o n s i d e r e d in the d e s i g n and interpretation o f b i o l o g i c a l therapy trials.

References 1. Agar R, Malloy B, Sherrod A, Mazumder A (1988) Successful therapy of natural killer-resistantpulmonary metastases by the synergism of 7-interferon with tumor necrosis factor and interleukin-2 in reite. Cancer Res 48:2245 2. Belldegrun A, Kasid A, Uppenkamp M, Topalian SL, Rosenberg SA (1989) Human tumor infiltrating lymphocytes. Analysis of lymphokines in RNA expression and relevance to cancer. Immunology 148:4520 3. Blay J-V, Favrot MC, Negrier S, Comboret V, Chousib S, Morcatello A, Kaemmerlen P, Franks CR, Thierry P (1990) Correlation between clinical response to interleukin-2 therapy and sustained production of tumor necrosis factor. Cancer Res 50:2371 4. Chatenoud L, Ferran C, Legendre C, Thovara I, Merito S, Reuter A, Gevaert Y, Kreis H, Franchimont P, Bach J-F (1990) In vivo cell activation following OKT3 administration. Transplantation 49:697 5. Demetri GD, Spriggs DR, Sherman ML, Arthur KA, Imamura K, Kofe DW (1989) A phase I trial of recombinant human tumor necrosis factor and interferon-gamma: effects of combination cytokine administration in vivo. J Clin Oncol 7:1545

6. Economou JS, Essner R, Rhoades KR, McBride WH, Golub S, Hohnes EC, Morton DL (1988) Molecular mechanisms of interleukin-2 induction of human tumor necrosis factor gene expression in macrophages. Surg Forum 39:426 7. Economou JS, McBride WH, Essner R, Rhoades KR, Golub SH, Holmes EC, Morton DL (1989) Tumor necrosis factor production by interleukin-2 activated macrophages in vitro and in vivo. Immunology 67:514 8. Fong Y, Lowry SF (1990) Short analytical review: tumor necrosis factor in the pathophysiology of infection and sepsis. Clin Immunol Immunopatho155:157 9. Fraker DL, Langstein HN, Norton JA (1989) Passive immunization against tumor necrosis factor partially abrogates interleukin-2 toxicity. J Exp Med 170:1015 10. Hoon DSB, Foshag LJ, Nizze AS, Bohman R, Morton DL (1990) Suppressor cell activity in a randomized trial of patients receiving active specific immunotherapy with melanoma cell vaccine and low dosage of cyclophosphamide. Cancer Res 50:5358 11. Klausner JM, Schoof DD, Kobzik LM, et al. (1987) Interleukin-2 leads to lung lymphosequestration, thromboxane generation, and rapid increase in permeability. Surg Forum 38:428 12. Kovacs EJ, Beckner SK, Longo DL, Varesio L, Young HA (1989) Cytokine gene expression during the generation of human lymphokine-activated killer cells: induction of interleukin-1 ~ by interleukin-2. Cancer Res 49:940 13. Lotze MT, Matory YL, Rayner AA, et al. (1986) Clinical effects and toxicity of interleukin-2 in patients with cancer. Cancer 58:2764 14. Lotze MT, Chang AE, Seipp CA, Simpson C, Vetto JT, Rosenberg SA (1986) High-dose recombinant interleukin-2 in the treatment of patients with disseminated cancer responses, treatment-related morbidity, and histologic findings. JAMA 256:3117 15. McBride WH, Economou JSS, Nayersina R, Comora S, Essner R (1990) Influences of interleukin 2 and 4 on tumor necrosis factor production by murine mononuclear phagocytes. Cancer Res 50:2949 16. Mcintosh JK, Jablons DM, Mule JJ, Nordon RP, Radikoff S, Lotze MT, Rosenberg SA (1989) In vivo induction of IL-6 by administration of exogenes, cytokines and detection of de novo serum levels of IL-6 in tumor-bearing mice. J Immunol 143:162 17. McIntosh JK, Mule JJ, Krosnick JA, Rosenberg SA (1989) Combination cytokine immunotherapy with tumor necrosis factor-cq interleukin-2, and c~-interferon and its synergistic anti-tumor effects in mice. Cancer Res 49:1408 18. Michie HR, Eberlin TJ, Spriggs DR, Manogue KR, Cerami A, Wilmore DW (1988) Interleukin-2 initiates metabolic responses associated with critical illness in humans. Ann Surg 208:4993 19. Michie HR, Manogue KR, Spriggs DR, Revhaug A, O'Dwyer SO, Dinarello CA, Cerami A, Wolff SM, Wilmore DW (1988) Detection of circulating tumor necrosis factor after endotoxin administration. N Engl J Med 318:1481 20. Natarson C, Eichenholtz PW, Danner RL, Eichacher PQ, Hoffman WD, Kuo SC, Banks SM, MacVittie TJ, Parrillo JE (1989) Endotoxin and tumor necrosis factor changes in dogs stimulate the cardiovascular profile of human septic shock. J Exp Med 169:823 21. Nickimura T, Ohtu S, Soto N, Togashi Y, Goto M, Hashimoto Y (1989) Combination tumor immunotherapy with recombinant tumor necrosis factor and recombinant intefleukin-2 in mice. Int J Cancer 40:255 22. Petersen CM, Moller BK (1988) Immunological reactivity and bioactivity of tumor necrosis factor (letter). Lancet 1: 934 23. Rosenberg SA, Lotze MT, Muul LM, et al. (1982) A progress report on the treatment of 157 patients with advanced cancer using lymphokine-activated killer cells and interleukin-2 or high-dose interleukin-2 alone. N Engl J Med 316:889 24. Sung S-SJ, Bjorndall JM, Wang CY, Kao HT, Fu SM (1988) Production of tumor necrosis factor/cachectin by human T cell lines and peripheral blood T lymphocytes stimulated by phorbol myristate acetate and anti-CD3 antibody. J Exp Med 167:937 25. Sung S-SJ, Jung LKL, Walters JA, Chen W, Wang CY, Fu SM (1988) Production of tumor necrosis factor by human B cells and tonsillar B cells. J Exp Med 168:1539