Combined preoperative and postoperative immunotherapy for murine C1300 neuroblastoma

Combined preoperative and postoperative immunotherapy for murine C1300 neuroblastoma

Combined Preoperative and Postoperative Immunotherapy Murine Cl300 Neuroblastoma for By Carol t. Fowler, Stephen P. Brooks, Jon E. Rossman, and Don...

937KB Sizes 0 Downloads 5 Views

Combined

Preoperative and Postoperative Immunotherapy Murine Cl300 Neuroblastoma

for

By Carol t. Fowler, Stephen P. Brooks, Jon E. Rossman, and Donald R. Cooney Lexington, Kentucky, Buffalo, New York, and Columbus, Ohio

D

l Preoperative treatment of murine C1300-neuroblastoma (C1300) with triple immunotherapy using low-dose cyclophosphamide (CY), retinyl palmitate (RP), and interleukin-2 (IL2). followed by tumor resection leads to significant initial tumor control and prolonged survival. However, because long-term tumor recurrence is 67%. the efficacy of continued postoperative immunotherapy is now evaluated. Thirty-two A/J mice with 1 cm subcutaneous Cl300 tumors were treated for 13 days with CY-100 mg/kg. intraperitoneally (IP), on day 2 of treatment then 25 mg/kg on day 9, UP-2500 IU IP 2 x/week, and IL2 1.6 x IO5 U IP BID on days 4 to 9 and 11 to 13. On day 14, mice were divided into five treatment groups: (1) OP (operated-tumor resection, n = 6); (2) OP+CY (resection and postoperative CY, n = 7); (3) OP+CY+RP (resection and postoperative CY+RP, n = 7); (4) OP+CY+RP+ILZ (resection and postoperative CY+RP+ILZ, n = 7); and (5) CY+RP+lL2 (continued CY+RP+lL2 with no resection, n = 5). Survival and postoperative tumor recurrence were followed for 60 days. The cure rates were group 1 33% (2/6), group 2 43% (3/7), group 3 29% (2/7), group 4 71% (5/7), and group 5 20% (l/5). After surgery, tumors that recurred did so in 6 to 22 days, with no statistical difference noted between groups. MHC class I antigenic expression of tumors resected on day 14 and recurrent tumors was determined with monoclonal antibodies and flow cytometry. In tumors resected on day 14, class I expression measured by mean fluorescence, was 374.6 f 27.40. Recurrent tumors in all groups expressed significantly less class I antigen (230.7 2 36.32) than the original tumors. Because class I expression is important for cell-mediated tumor lysis, the low expression in the recurrent tumors suggests that the mechanism of their recurrence and resistance to preoperative immunotherapy was through escape of host antitumor immunosurveillance. Because the best overall results, with 71% cures, were obtained with continued postoperative triple immunotherapy despite the presence of viable immunoresistant cells, the success of continuing the triple-drug regimen suggests that the mechanism of action of the regimen may not be purely immunotherapeutic but chemotherapeutic as well. Copyright o 1993 by W.5. Saunders Company

ESPITE VAST improvements in chemotherapy and irradiation, the prognosis of children with advanced stages of neuroblastoma has not improved significantly in the past two decades. Because neuroblastoma is an immunogenic tumor, immunotherapy has been investigated as an alternate therapy to chemotherapy.1-5 Using murine C1300-neuroblastoma (C1300), a well-established model for human disease, we previously demonstrated significant tumor regression and increased survival using tripleagent immunotherapy prior to tumor resection.6 The regimen utilized low-dose cyclophosphamide (CY), retinyl palmitate (RP), and interleukin-2 (IL2). However, despite grossly complete tumor resections and prolonged survival, many tumors recurred, resulting in a 33% late cure rate. The present investigation evaluates the value of continuing postoperative therapy using combinations of the same three immunomodulating agents (CY, RP, and IL2).

INDEX WORDS: Neuroblastoma, phosphamide; retinyl palmitate.

Cyclophosphamide

C1300; interleukin-2;

cyclo-

From the Section of Pediatric Surgery, University of Kentucky Medical Center, Lexington, m; the Department of Molecular Medicine and Immunology, Roswell Park Cancer Institute, and the Department of Pediatric Surgery, Children S Hospital of Buffalo, Buffalo, h?‘j; and the Department of Pediatric Surgery, Children’s Hospital, Columbus, OH. Presented at the 23rd Annual Meeting of the American Pediatric Surgical Association, Colorado Springs, Colorado, May 13-16, 1992. Address reprint requests to Carol L. Fowler, MD, Section of Pediatric Surgery, University of Kentucky Medical Center, Lexington, KY 40536-0084. Copyright 0 1993 by W.B. Saunders Company 0022-3468/93/2803-0024$03.00/0 420

MATERIALS AND METHODS

Mice Female A/J mice, 8 to 10 weeks old, were obtained from Jackson Labs (Bar Harbor, ME). Animals were housed at a constant temperature and provided ad lib standard rodent chow and water.

Tumor Cl300 has been maintained in our laboratory by serial passage, by both in vivo and in vitro techniques. For passage in vivo, tumor cell suspensions were prepared by aseptically excising subcutaneous tumors, then mechanically disaggregating them in phosphatebuffered saline (PBS). After viability of cells was determined by trypan blue exclusion, a suspension of 1 x lo6 live tumor cells was injected subcutaneously in the flanks of syngeneic A/J mice.

Each dose of CY (Sigma Chemical Co, St Louis, MO) was administered as a single intraperitoneal (IP) injection per day of either 25 mg/kg or 100 mg/kg as indicated in the treatment schedule (Fig 1).

Interleukin-2 Highly purified recombinant human IL2 from E coli was obtained from the Cetus Corporation (Emmeryville, CA). It was reconstituted in sterile PBS to a dilution of 1.6 x IO6 UimL and injected IP as 0.1 mL doses of 1.6 x lo5 U.

Retinyl Palmitate Aquasol A parenteral water-soluble

form; Armor

Journat offediafric

(retinyl palmitate, Pharmaceutical

Surgery, Vol28,

50,000 JUimL

Co, Kankakee,

No 3 (March),

in

IL) was

1993: pp 420-427

421

COMBINEDIMMUNOTHERAPY

rRP

TUMOR RESECT

c;ylOo RP 1

I

I

2

4

DAY OF RX

I

I

8

9

WEEK 1

I

I

11

1

13 14

WEEK 2

---

IL2 1

I

RP 15

7

loo

16

CY , 25

RP

Tumors were manually disaggregated in PBS and cell suspensions were counted. A final suspension of lo5 cells in 0.05 mL was incubated on ice with anti-Hz monoclonal antibody (Hybirtech, San Diego, CA) for 30 minutes. After washing with PBS, the cells were incubated on ice with FITC-conjugated second antibody for 30 minutes before two final washes with PBS. Flow cytometric analysis was determined with EPICS (Coulter Corp, Hialeah, FL). Class I expression was reported as mean fluorescence intensity.

I

RP

RP

18

--r-C---+---H 22 23 25 27

DAY OF RX - WEEK 3

MHC Class I Expression

Serum Vitamin A Levels 28

WEEK 4 -6

DOSES: CY-25 or 100 mglkg, _ ._____ ~~ -Fig 1. Treatment regimen. Preoperative immunotherapy with CY, RP, and IL2 was administered for 13 days to mice with l-cm Cl300 tumors. Tumor resection was performed on day 14 in four of the five groups, and postoperative therapy with various combinations of CY, RP, and IL2 were continued in four groups. Drug doses and treatment schedules are shown.

Using a previously reported6 modification of methods published by Neeld and Pearson’ and McCormick,s serum retinoic acid levels were determined in control untreated mice, and weekly in all groups of mice.

Statistical Methods Survival data were compared with a generalized Wilcoxon test. Other data are presented as group means +- SEM. ANOVA was used for group comparisons of class I expression. P values of 5 0.05 were considered significant. RESULTS

injected IP as indicated in the treatment schedule in doses of 0.05 ml (2,500 IU) (Fig 1).

Treatment Schedules Thirty-two mice with l.O-cm subcutaneous flank tumors (7 days after implantation) were treated for 13 days with CY, RP, and IL2. CY was administered 100 mg/kg on day 2, and 25 mg/kg on day 9. RP was given twice per week as 2,500 IU per dose. IL2 was injected twice a day as 1.6 x 105U/dose on days 4 to 9 and 11 to 13. On day 14, animals were divided into five treatment groups: (1) OP (operated-tumor resection, n = 6); (2) OP+CY (resection and postoperative CY, n = 7); (3) OP+CY+RP (resection and postoperative CY+RP, n = 7); (4) OP+CY+RP+IL2 (resection and postoperative CY+RP+IL2, n = 7); and (5) CY+RP+IL2 (continued CY+RP+IL2 with no resection, n = 5). On day 14, tumors were resected in groups 1 to 4 under sterile conditions using sodium pentobarbital anesthesia. Resections were performed as completely as possible, and included adjacent areas of involved skin, but major resections of involved muscle were not attempted. Postoperative therapy was continued in groups 2 to 5 for 30 days. Drug doses were: CY 100 mgikg on day 16, then 25 mg/kg once per week; RP 2500 IU twice per week; IL2 1.6 x lo5 U/dose twice a day for 6 consecutive days each week with 1 day of rest, beginning 1 week postoperatively.

Suwival and Cure Rate

The postoperative cure rate was 33% (2/6) in group OP, 42.8% (3/7) in OP+CY, 28.6% (2/7) in OP+CY+RP, and 20% (l/5) in CY+RP+IL2 (Fig 2). There was no statistical difference between these results; however, the 60-day postoperative cure rate was 71% in OP+CY+RP+IL2. Survival was significantly longer in this group compared to all other groups (P < .Ol), except OP+CY, in which the difference in survival closely approached significance (P < .06). Survival curves are plotted in Fig 3. Tumor Growth Rate and Recurrence

Untreated Cl300 measures 1 cm 7 days after implantation, grows rapidly, and proves uniformly fatal within 2 to 3 more weeks. In all groups receiving the 13 day preoperative course of CY +RP+lL2, the tumor growth tapered and did not significantly in_____.~~

TREATMENT-GROUP 1. OP

Daily measurements of tumor size were determined and recorded as the product of the tumor’s greatest width and diameter. Survival and tumor recurrence were followed in all groups for 60 days after tumor resection.

Histology Resected tumors were horizontally divided into two sections. Half was fixed in formalin and submitted for histological examination and the other half was analyzed for class I expression.

L

216

(33%)

3t7

(43%)

211

(29%)

OP

+cY

3. OP

+CY

+RP

4.

+CY

+RP

+IL2

511

(71%)

+RP

+IL2

l/5

(20%)

2.

Tumor Growth and Recurrence, Survival

NO. (%) CURED

5.

OP

CY

Fig 2. All mice received preoperative therapy with CY, RP, and 112. Operation (OP) to resect tumor was performed in groups 1 to 4, but not in group 5. The cure rates are noted after various combinations of postoperative CY, RP. and Il.2 were administered. OP+CY+RP+ILZ. had the highest cure rate, withf < .Ol as compared to all other groups except OP+CY (P < .06).

422

100

FOWLER ET AL

Percent

ple of recurrent tumor growth is shown in Fig 4. Despite recurrences, metastastic lesions were not observed in any group, which is typical of Cl300 neuroblastoma.

op+ * CY+RP+ILZ

op+CY OP OP+CY+RP CY+RP+ILZ

Fig 3. trated.

Survival curves of the five experimental

!

groups are illus-

crease from days 5 to 14 of treatment. The group that received continued triple therapy but no tumor resection (CY + RP+ IL2) demonstrates the clinical courses of the tumor after the period of initial growth stasis. In this group, the tumor appears to regain momentum in the growth rate starting on day 18. Although the tumor then grew at a slower rate than the untreated tumor, the tumor proved fatal to 80% of the mice in the group within 3 weeks. This result was similar to that of the group that received tumor resection on day 14 but no postoperative therapy (OP, with 67% mortality). In this group (OP) and the three postoperatively treated groups (OP+ CY, OP+ CY + RP, OP+CY +RP+IL2), tumors often reappeared from 8 to 22 days postoperatively, despite apparent complete tumor resections at the time of surgery (Fig 4). In all groups, the recurrent tumor growth was rapid and fatal despite continued postoperative therapy, and similar to untreated tumors. A representative examTUMOR

SIZE

Tumor Appearance and Histology

All tumors pretreated with the triple-drug therapy were small and easily resectable at 14 days of treatment. This is in comparison to the large, bulky, invasive tumors in untreated mice, which in many cases were lethal prior to day 14 when surgery was scheduled (Fig 5). Advanced untreated Cl300 tumors are large, irregular, invasive masses with central areas of hemorrhagic coagulation necrosis and abundant mitoses with marked pleomorphism. A striking difference was noted in the gross appearance of the triply treated tumors, which were typically small and fibrotic. Microscopically, large areas of coagulation necrosis were seen. In these tumors, the mitosis rate varied from few to abundant, and cellular pleomorphism with vesicular nuclei was common. The recurrent tumors all resembled each other with coagulation necrosis, vesicular nuclei, and high mitosis rates. A gross microscopic estimate of the tumors’ viability demonstrated that the percent of viable tumor varied considerably from 0 to 90% in all groups, but did not correlate with the ultimate prognosis of the individual

(cm2)

8r(IUNTREATED 6-

CY+RP+ILZ 4-

-:/'

POSTOP

*

DAY OF TREATMENT Fig 4. Tumor growth rate. Tumors were measured daily in all mice during the preoperative period of treatment, and postoperatively in all recurrent tumors. Note period of growth stasis from day 5 to 18 in tumors treated preoperatively with CY+RP+ILZ, as compared to the uninhibited, rapid growth of untreated tumors. The curve marked CY+RP+lLP depicts the clinical course of tumors treated with continued CY, RP, and IL2 but without surgical resection. Tumors were resected in the other groups on day 14 of treatment. To the right is demonstrated the growth of a representative recurrent tumor.

Fig 5. Tumor appearance. Cl300 tumors were resected on day 14 of treatment. Tumor seen at top of figure is from an untreated animal. Note the contrast between this large, friable hemorrhagic tumor and the smaller, fibrotic tumor below that was resected from an animal treated with CY+RP+ILP.

423

COMBLNED IMMUNOTHERAPY

mice. Tumor infiltrating lymphocytes were not identified in any of the Cl300 tumors. MHC Class I Expression

As measured by mean fluorescence intensity, class I expression of cells from tumors preoperatively treated with CY + RP+ IL2 and resected on day 14 of treatment, measured 374.8 ? 27.4 (n = 16), which was significantly higher than recurrent tumors from all groups (230 +_38.3, n = 7, P = .Ol, Fig 6). Untreated tumors also express low levels of class I surface antigen (272.3 2 11.1, n = 12). These values compare to control MHC class I expression of lymphocytes from peripheral blood (225 ‘_ 40.9, n = 5). Serum Vitamin A Levels

Serum retinoic acid levels in untreated mice averaged 20.5 ug/dL. Weekly serum levels were determined in all mice that received RP as part of their therapy. These ranged from 29 to 46 ug/dL, which were nontoxic levels. DISCUSSION

In a previous study on murine Cl300 neuroblastoma, a 2-week preoperative course of triple immunotherapy using CY, RP, and IL2 followed by tumor resection resulted in the best survival rate and the greatest inhibition of tumor growth, as compared to results using any single or double combinations of the drugs.‘j The proposed mechanisms of action of this triple-drug regimen were both immunogenic and directly cytotoxic and have been discussed in the prior study.h Unfortunately, 67% of these preoperatively treated tumors eventually recurred within 3 weeks 400

LUORESCENCE

200

100

0 14 DAY TREATMENT

CM+/-SEM)

I

RECURRENT

UNTREATED

MHC class I expression. Using monoclonal antibodies, class of treated, untreated, and recurrent tumors is measured by mean fluorescence intensity from flow cytometry. The class I expression of tumors resected on day 14 of CY+RP+IL2 treatment is significantly higher than that of either untreated or recurrent tumors. Note that the expression of recurrent tumors is similar to untreated tumors. Fig 6.

I expression

after surgery. Therefore, it is evident that viable tumor was still present at the time of tumor resection. This was confirmed by the presence of viable cells found histologically in the resected tumors. A longer course of continued triple-drug therapy without tumor resection was not effective in eradicating all viable tumor. This was demonstrated by the fact that 80% of tumors in this treatment group (CY + RP+ IL2) resumed rapid growth after the initial period of growth stasis, which finally led to the demise of the hosts (Figs 2, 3, and 4). The best results (71% cure) were produced with combination therapy consisting of preoperative triple-drug therapy, tumor resection, then postoperative triple therapy (OP + CY + RP+IL2). Surgical cytoreduction thus appeared important to the outcome. To understand the continued success of the triple regimen when applied postoperatively, MHC class I antigenic expression was compared in untreated, treated, and recurrent tumors. Neuroblastomas, like other malignant tumors particularly of epithelial origin, normally express low levels of class I antigen.9J0 Because class I expression is a requisite for host immune recognition and lysis of tumor cells9 the success of immunotherapy for tumors with low class I expression is probably limited unless upregulation can be produced by treatment. However, the number of cell surface MHC molecules that are needed to stimulate cytolytic cell activity has not been determined; therefore increased MHC levels may not be biologically significant. It has been demonstrated that although class I expression of neuroblastoma cells can be increased by interferon-y in vitro, cytolytic cell killing is not simultaneously increased.” However, neuroblastoma cells can be killed by allotypic effectors when the cells were coupled to the cytolytic cells by lectins. Although class I expression is just one of many determinants of the antitumor immune response, a clinical correlation between human neuroblastoma response and class I expression has been suggested by Squire et al, who noted the greatest class I expression in stage IV-S as compared to the other stages.i2 IV-S neuroblastoma exhibits a good prognosis related to a high rate of spontaneous resolution of metastatic lesions, which may represent an effective immune-mediated antitumor response by the host. Cl300 tumors were resected at 14 days because this was during the period of growth stasis, and therefore at the time of peak clinical response to therapy. Class I surface antigen of tumors resected at this time was significantly upregulated as compared to untreated and recurrent tumors (Fig 6). Class I expression in the recurrent tumors was low, similar to untreated tumors. The low class I expression of recurrent

424

tumors may signify that these tumor cells were able to proliferate because they were able to escape early host antitumor immunosurveillance. Cl300 is a heterogeneous tumor of cells of both high and low expression of MHC molecules. It is possible that the recurrent tumors, expressing low levels of class I molecules, expanded from cells of low class I expression that were immunoresistant. The upregulation of class I expression in treated tumors as compared to untreated tumors is evidence of a modification of cell surface antigenic expression. This may have resulted from IL2-induced autologous interferon activity, as suggested in other studies.13 In experimental murinel and human14 neuroblastoma cell lines, class I expression has been increased by exogenous interferon treatment. RP may have augmented these effects, as it can induce cellular antigenic modulation in neuroblastoma cell cultures.14s15 Retinoids have also been demonstrated to induce morphologic cellular differentiation and maturation changes in both animal and human neuroblastoma cell cultures.16,17 Although there were individual variations, the growth rate of the recurrent tumors was generally rapid and similar to control untreated tumors. This indicates that the recurring tumors were developing from residual therapy-resistant clones of cells. All recurrent tumors were evident from 8 to 22 days after resection. Comparing the treatment groups, there was no statistical difference either between the time of postoperative tumor reappearance, or tumor growth once the tumor had recurred. The major difference between groups was the recurrence rate. This directly affected the survival rate because the tumors, once they recurred, were all lethal and resistant to any further therapy attempted. Histological examination of the recurrent tumors did not show any significant differences between treatment groups. Large central areas of coagulation necrosis with variable mitosis rates were found. Often, rims of viable tumor were seen around a necrotic tumor center. While regression of a viable rim of tumor is dependent upon host immunity, central hemorrhagic necrosis does not significantly depend upon host immunity. l* Coagulation necrosis usually results from ischemic changes secondary to deprivation of blood flow produced by therapy. The pattern of tumor necrosis was similar to that experimentally produced by treatment of tumors with either interferonlg or tumor necrosis factor.20 Tumor infiltrating lymphocytes have not been demonstrated in murine Cl300 neuroblastoma. Because a complete antitumor response was demonstrated postoperatively in 71% of the group that

FOWLER ET AL

received continued postoperative triple therapy after tumor resection (OP+CY+RP+IL2), this indicates that the cells in the residual tumor burden were still able to respond to the triple therapy by some mechanism, whether immune or cytotoxic-mediated. Although the early (preoperative) antineoplastic effects of the triple regimen may have been immune related (as evidenced by the upregulation of MHC class 1 expression), the postoperative antitumor effects of the same regimen may have been due to additional chemotherapeutic effects of the higher 100 mg/kg dose (but still considered low dose) of CY, as compared to the purely immunotherapeutic effects of the lower CY dose (25 mg/kg). In our previous experiments, growth of Cl300 was associated with increased T suppressor activity, and low dose CY was able to abrogate this suppressor activity while demonstrating an antitumor effect.4 The higher dose of CY (100 mg/kg) may not only physically debulk the tumor, but contribute to an antitumor immune response by releasing soluble antigen from the tumor. This soluble antigen might activate the necessary effector and accessory cells to produce a clinical antitumor response. However both 25 and 100 mg/kg of CY have a primarily immunomodulating effect, as demonstrated in that the antitumor effect can be eliminated with antilymphocyte serum.4 Tumor cytoreduction both by CY and by surgery may be effective by decreasing the suppressive cytokines released by the tumor. One of us (S.P.B.) has demonstrated active TGF-l3 in serum-free media of the cell lines Cl300 and TBJ (a Cl300 clone) (personal communication). Vitamin A analogs have experimentally demonstrated both tumoricidal and as well as immunomodulatumoristatic properties 21-24 tion.21*25-27 Synergistic antineoplastic effects have been reported between retinoids and CY,28-30as well as between CY and IL2.31-34The retinoids exert direct cytotoxic effects that may have been active in all groups in the postoperative period. Although some groups did not receive postoperative RP, all groups had been treated with RP for 2 weeks preoperatively. The serum retinoic acid level remained somewhat elevated in all groups when followed for 4 weeks postoperatively, even in the groups that did not receive postoperative RP. This potential for prolonged vitamin A effect exists because retinoids are stored in the liver and slowly released into the serum. Some proposed mechanisms of retinoid antitumor activity involve labilization of cell membranes and release of lysosomal enzymes with subsequent cell lysis.21 Labilization of cell membranes may cause the tumor cell to be more susceptible to chemotherapy,

COMBINED

425

IMMUNOTHERAPY

thus producing an enhanced effect when combined with CY therapy. Experimentally, additive antitumor effects have been noted when retinoid therapy is combined with radiation or chemotherapeutic agents including CY .24J8-30 This work suggests that although the combination of three agents (CY,RP,IL2) administered at low nontoxic doses can form an initially effective antitumor therapy, resistant clones of tumor cells can survive. Probably many immune factors and cytotoxic

effects contribute to the observed therapeutic results. Combination therapy, such as the present one that is cytotoxic as well as immunopotentiating, has many theoretic advantages over any single agent or modality therapy. Future studies may demonstrate that if initial immunotherapy fails, a different mechanism of destruction for the residual tumor, such as high-dose chemotherapy, may be more rewarding in the postoperative period.

REFERENCES 1, Sigal RK, Reynolds JV, Markmann JF, et al: Upregulation of MHC class I: Effect on growth and LAK sensitivity of neuroblastoma. Surg Forum 39572-5751988 2. Necheles TF, Tefft M, Weinberg V: Randomized trial of immunotherapy in the treatment of advanced neuroblastoma, in Terry WD, Rosenberg SA (eds): Immunotherapy of Human Cancer. New York, NY, Elsevier, 1982, pp 377-383 3. Bruce J, Brooks SP, Rich GA, et al: Effects of immunostimulation on host survival in A/J mice with transplantable Cl300 neuroblastoma. Curr Surgery 45:17-19, 1988 4. Fowler CL, Brooks SP, Rossman JE, et al: Postoperative immunotherapy of murine C1300-neuroblastoma. J Pediatr Surg 25:229-237, 1990 5. Brooks SP, Cooney DR, Fowler CL: Treatment of murine Cl300 neuroblastoma with IL-2/LAK cell therapy. Presented at the meeting of Association of Academic Surgeons, Salt Lake City, UT, November 1988 6. Fowler CL, Brooks SP, Squire R, et al: Enhanced resection and improved survival in murine neuroblastoma (C1300-NB) after preoperative immunotherapy. J Pediatr Surg 26:381-388,199l 7. Neeld JB Jr, Pearson WN: Macro- and micromethods for the determination of serum vitamin A using trifluoroacetic acid. J Nutrition 79454-462, 1963 8. McCormick DB: Vitamins, in Tietz NW (ed): Fundamentals of Clinical Chemistry. Philadelphia, PA, Saunders, 1986, pp 927964 9. Tanaka K, Yoshioka T, Bieberich C, et al: Role of the major histocompatibility complex class I antigens in tumor growth and metastases. Annu Rev Immunol6:359-380, 1988 10. Lampson LA, Fisher LA, Whelan JP: Striking paucity of HLA-A,B,C and @microglobulin on human neuroblastoma cell lines. J Immunol 130:2471-2478, 1983 11. Main EK, Monos DS, Lampson LA: IFN-treated neuroblastoma cell lines remain resistant to T cell-mediated allo-killing, and susceptible to non-MHC-restricted cytotoxicity. J Immunol 141: 2943.2950,1988 12. Squire R, Fowler CL, Brooks SP, et al: The relationship of class I MHC antigen expression to stage IV-S disease and survival in neuroblastoma. J Pediatr Surg 25:381-386, 1990 13. Merigan TC: Is recombinant interleukin-2 the best way to deliver interferon-gamma in human disease? J Interferon Res 7:635-639, 1987 14. Gross N, Beck D, Favre S, et al: In vitro antigenic modulation of human neuroblastoma cells induced by IFN-gamma, retinoic acid and dibutyryl cyclic AMP. Int J Cancer 39:521-529, 1987 15. Reynolds CP, Maples J: Modulation of cell surface antigens accompanies morphologic differentiation of human neuroblastoma cell lines. Prog Clin Biol Res 175:13-37, 1985 16. Shea TB, Fischer I, Sapirstein VS: Effect of retinoic acid on

growth and morphologic differentiation of mouse NB2a neuroblastoma cells in culture. Dev Brain Res 21:307-314,1985 17. Side11N: Retinoic acid-induced growth inhibition and morphologic differentiation of human neuroblastoma cells in vitro. J Nat1 Cancer Inst 68:589-592,1982 18. North RJ, Have11 EA: The antitumor function of tumor necrosis factor (TNF): II. Analysis of the role of endogenous TNF in endotoxin-induced hemorrhagic necrosis and regression of an established sarcoma. J Exp Med 167:1086-1099,198s 19. Dvorak HF, Gresser I: Microvascular injury in pathogenesis of interferon-induced necrosis of subcutaneous tumors in mice. J Nat1 Cancer Inst 81:497-502,1989 20. Asher A, Mule JJ, Reichert CM, et al: Studies on the anti-tumor efficacy of systemically administered recombinant tumor necrosis factor against several murine tumors in vivo. J Immunol 138963-974, 1987 21. Eccles SA: Effects of retinoids on growth and dissemination of malignant tumors: Immunological considerations. Biochem Pharm 34:1599-1610,1985 22. Felix EL, Cohen MH, Loyd BC: Immune and toxic antitumor effects of systemic and intralesional vitamin A. J Surg Res 21:307-312,1976 23. Watson RR, Moriguchi S, Gensler HL: Effects of dietary retinyl palmitate and selenium on tumoricidal capacity of macrophages in mice undergoing tumor promotion. Cancer Lett 36:181187,1987 24. Seifter E, Rettura G, Padawer J, et al: Regression of CSHBA mouse tumor due to X-ray therapy combined with supplemental p-carotene or vitamin A. J Nat1 Cancer Inst 71:409417,1983 25. Moriguchi S, Werner L, Watson RR: High dietary vitamin A (retinyl palmitate) and cellular immune functions in mice. Immunology 56:169-177,1985 26. Katz DR, Drzymala M, Turton JA, et al: Regulation of accessory cell function by retinoids in murine immune responses. Br J Exp Path01 68:343-350, 1987 27. Prabhala RH, Maxey V, Hicks MJ, et al: Enhancement of the expression of activation markers on human peripheral blood mononuclear cells by in vitro culture with retinoids and carotenoids. J Leukoc Biol45:249-254, 1989 28. Nathanson L, Maddock CL, Hall TC: Exploratory studies of vitamin A and cyclophosphamide in tumor-bearing mice. J Clin Pharm 9:359-373, 1969 29. Levenson SM, Rettura G, Seifter E: Effects of supplemental dietary vitamin A and p-carotene on experimental tumors, local tumor excision, chemotherapy, radiation injury, and radiotherapy, in Butterworth CE Jr, Hutchinson ML (eds): Nutritional Factors in the Induction and Maintenance of Malignancy. New York, NY. Academic, 1983, pp 169-203

FOWLER ET AL

30. Cohen MH, Carbone PP: Enhancement of the antitumor effects of 1,3-bis (2-chloroethyl)1-nitrosourea and cyclophosphamide by vitamin A. J Nat1 Cancer Inst 48:921-926,1972 31. Lee K, O’Donnell RW, Marquis D, et al: Eradication of palpable intradermal murine bladder tumors by systemic interleukin-2 and cyclophosphamide in C3H mice. J Biol Response Mod 7:32-42,1988 32. Hosokawa M, Yabiku T, Ikeda J, et al: Effects of a combination of cyclophosphamide and human recombinant interleukin-2 on pulmonary metastases after the surgical removal of a

3-methylcholanthrene-induced primary tumor in autochthonous mice. Jpn J Cancer Res 79:1147-1154,1988 33. Kolitz JE, Wong GY, Welte K, et al: Phase I trial of recombinant interleukin-2 and cyclophosphamide: Augmentation of cellular immunity and T-cell mitogenic response with long-term administration of rIL-2. J Biol Response Mod 7:457-472, 1988 34. Topalian SL, Solomon D, Avis FP, et al: Immunotherapy of patients with advanced cancer using tumor-infiltrating lymphocytes and recombinant interleukin-2: A pilot study. J Clin Oncol 6:839853,1988

Discussion P.R. Exelby (New York, NY): I am very pleased to open the discussion on this interesting paper which continues the work of Drs. S. Fowler, Cooney, and colleagues on the effects of immunotherapy on the Cl300 model of murine neuroblastoma. Improved survival using the preoperative regimen of cyclophosphamide, retinyl palmitate, and interleukin-2 prior to complete surgical resection improves survival. Because the recurrence rate was so high, used the same four groups of mixed immunotherapy postsurgery and showed improved survival. The best result was in the group with surgical removal followed by cyclophosphamide, retinyl palmitate, and IL-2. My question starts at the end of their discussion which is, are we sure the improved survival is due to an immune response, since the alkylating agent retinyl has killed cancer cells. Could this be more a chemotherapy rather than an immune therapy phenomenon. The other concern that I have is that Cl300 certainly differs somewhat from human neuroblastoma. It doesn’t metastasize and I don’t think it has shown ganglionic maturation during treatment in vivo, although it has in the tissue culture. Is it therefore truly comparable to human neuroblasts? My colleague Michael La Quaglia has looked at this paper and raised two other questions. One is, if the up-regulation of class I expression is important, how specific is the anticlass I antibody used in this study? The validity of the statistical analysis of 32 mice in 5 different treatment groups was raised. With all these reservations however, we think the work is extremely fascinating and would like to know if you have in the works a phase II study in human neuroblastoma. G.M. Huase (Denver, CO): This interesting neuroblastoma murine model has been used over the years to demonstrate many things. However, if you try to follow clinical correlates from each of these studies, you may be disappointed. I am concerned with the low class I antigen expression that was shown implying that the immunotherapy may not be so effective. This particular model has provided better survival

rates with combination chemotherapy than was noted here with low-dose chemotherapy and immunomodulation. I wonder if you considered repeating this study with better chemotherapy in addition to the immunotherapy which would more closely mimic the clinical situation. C.L. Fowler (response): Thank you for those comments. First of all, Dr Exelby, this is a continued study. We have done many previous studies to show that the mechanism of cyclophosphamide is dose dependent. The first study showed that low dose cyclophosphamide in the doses we are using is an immunomodulator. Low-dose cyclophosphamide is an immunomodulator in that it decreases an increased T-suppressor function that is produced by the neuroblastoma cells. We have shown that this increased T-suppressor function produced by neuroblastoma cells is reversed by low-dose cyclophosphamide. If you use antithymocyte serum which eliminates the T cells and then treat tumors with low-dose cyclophosphamide in the doses that we have used, the antitumor effect is in fact abrogated, so that does show this is an immunogenic response we are seeing. Also, we had previously reinjected tumor in animals cured with low-dose cyclophosphamide, and these animals were completely resistant to reinjected tumor growth. This shows that there was a learned antibody response. If you reinject other animals that were cured of Cl300 tumor with high-dose cyclophosphamide, which is a nonimmunogenic cure, then the tumors will grow readily. It has been shown that vitamin A increases accessory cell function and macrophage killing. It is a very complex mechanism because, as in human neuroblastomas, Vitamin A analogs increase cell maturation and differentiation in culture. Vitamin A is also tumoricidal and tumoristatic. IL2 of course is a lymphokine that increases effector cell function. As far as human versus mouse neuroblastoma, Cl300 is a well established in-vivo or in-vitro model of the human tumor. It’s a good model because although

COMBINED

IMMUNOTHERAPY

Cl300 is non-metastatic, it is extremely aggressive. It’s immunogenic and it does express N-myc and ras oncogenes like human tumors. There is a clone called TBJ which is a metastatic variant but it has low immunogenicity. Presently, I have studies on-going to define the mechanisms responsible for the differences between the metastatic potential of the two clones. As for the questions about class I antigen, the monoclonal antibody used is specific for class I antigen. Theoretically, increased T-cell killing corresponds to increased class I antigenicity. However, this has not been shown in in vitro studies, especially after class I has been increased by interferon. When children with neuroblastoma have been treated with interferon, a clinical response has not been seen. However, the reason some studies indicate that class I upregulation is not very effective is because neuroblastoma also expresses low adhesion molecules. So the next step would be to use gene transfection to increase adhesion molecules so that the cytotoxic T-cells will stick to the neuroblastomas and the neuroblastoma cells will be good targets for killing.

427

There is a correlation between class I antigenicity and prognosis in that, in a separate study that we had done, low class I expression is normally seen in human neuroblastoma as well as the mouse neuroblastoma. There is a better prognosis in stages I and IV and we have shown that those stages do have a higher class I antigen expression than the higher stages with a bad prognosis. In rat neuroblastoma, high N-myc amplification corresponds to a low class I antigenicity, so there is more correlation between prognosis and class I antigen. We do not have any on-going studies on humans at this time. As to Dr Haase’s question, yes, we know that we can have better results using high-dose chemotherapy. However, the whole point of these studies is not to use high-dose chemotherapy. High-dose chemotherapy is toxic and immunocompromising. In humans, there are secondary tumors that develop long term. The point of immunotherapy is that, at least with the doses we are using, it’s nontoxic and potentially better in advanced stages of neuroblastoma since chemotherapy has failed to improve the prognosis of these children in the past two decades.