Postoperative immunotherapy of murine C1300-neuroblastoma

Postoperative

Immunotherapy

By Carol L.

of Murine C1300-Neuroblastoma

Fowler, Stephen P. Brooks, Jon E. Rossman, and Donald R. Cooney Buffalo,

New

0 Low-dose cyclophosphamide (CY) is an immunomodulating agent that down-regulates T suppressor cell function. This study investigates postoperative immunotherapy with CY as an alternate treatment for advanced immunogenic tumors such as neuroblastoma that typically respond poorly to traditional high-dose chemotherapy. A/J mice with 1.5cm subcutaneous Cl 3W-neuroblastoma (Cl 3WNB) tumors were divided into the following treatment groups: I, untreated (n = 14); II, 85% tumor resection (n = 18): Ill. sham-operated (n = 18); IV, multiple-dose CY (n = 6); V, 85% resection and single-dose CY (n = 14); VI, 85% resection and multiple-dose CY (n = 14). CY (100 mg/kg, intraperitoneally) was given initially 24 hours postoperatively to groups IV, V. and VI. Groups IV and VI also received weekly maintenance doses of 25 mg/kg CY. Results showed significantly increased survival (log-rank test) in CY-treated groups (IV, V, VI) compared with control groups ~1.11.111). Cures were observed only in groups receiving partial resection plus CY (V, 7%; VI, 29%). Although surgical debulking of tumor alone (II) did not enhance survival, the procedure normalized depressed total lymphocyte counts and the subpopulation of Lyt 2.3+ (T suppressor/cytolytic cells) in the immediate postoperative period during which immunotherapy with CY was instigated. This may have contributed to the success of CY immunotherapy. To characterize the tumor-host immune interaction, additional studies were performed. Results showed the following. (1) Mice cured by debulking plus CY (from groups V and VII could not be successfully reimplanted with Cl 3W-N8, demonstrating immunologic mediation by CY. (2) Injection of antithymocyte serum abrogated the therapeutic effects of CY in tumor-bearing mice, indicating that CY action is T cell-mediated. (3) T cell Concanavalin-A blastogenesis assay demonstrated suppression of T cell stimulation in tumor-bearing mice. This tumor-associated suppression was eradicated by CY treatment (188 mg/kg) of tumor-bearing mice. (4) Treatment of tumor-bearing mice with intravenous anti-IJx serum significantly delayed tumor growth (P = 882). demonstrating a role of IJK cells in C13W-N8 growth. These IJK cells are part of the T suppressor network. (5) An increased polymorphonuclear cell count in the peripheral blood was a marker for the presence of tumor tissue, even before the gross tumor was noted. Because this evidence suggests that tumor-associated T suppressor populations contribute to experimental neuroblastoma growth, a trial of treatment of neuroblastoma patients with low-dose CY to inactivate these populations and support host antitumor mechanisms may be indicated, especially in the face of treatment failures with high-dose CY. @ 1990 by W.B. Saunders Company. INDEX WORDS: Cyclophosphamide; immunotherapy; neuroblastoma.

T

HE CURRENT treatment of advanced neuroblastoma with high-dose chemotherapy and irradiation is unsatisfactory because it has not significantly Journal of Pediatric Surgery, Vol 25, No 2 (February), 1990: pp 229-237

York

increased overall patient survival in the past two decades.’ Furthermore, although palliation is occasionally achieved, it is usually accompanied by toxic side effects.2 A more efficacious and less toxic form of therapy for neuroblastoma is needed, and immunotherapy may provide such an alternative. Neuroblastoma evokes host immune responses both clinically and in experimental models.3*4Occasionally, spontaneous tumor regression occurs,’ and is possibly a consequence of the host immune response.6 Because these observations suggest that neuroblastoma is immunogenic, it has been postulated that immunotherapy might improve survival of affected patients by enhancing the host antitumor response.‘,* The syngeneic murine neuroblastoma, C 1300-NB, should provide an acceptable tumor model on which to test immunotherapies, because it has been shown previously that Cl 300-NB has immunogenic properties.2 While cyclophosphamide (CY) is traditionally used in high doses to treat neuroblastoma, there is evidence that in lower doses it is an immunomodulating agent that down-regulates T suppressor (Tsupp) cell function.‘-” This property of low-dose CY is potentially useful in tumor therapy because one mechanism by which tumors may evade the host antitumor response is to stimulate Tsupp function, which in turn inhibits the host antitumor effector cells.‘2*13Experimentally, preferential down-regulation of tumor-associated Tsupp cell function has mediated regression of established tumors in animal models.‘4*‘5 In this study, postoperative treatments using immunoregulating doses of CY in a clinically relevant model of incompletely resected, advanced murine neuroblastoma were evaluated. MATERIALS AND METHODS FemaleA/J mice, 8 to 10 weeks old, were obtained from Jackson

Laboratories (Bar Harbor, ME). C1300-NB has been maintained in From the Department of Pediatric Surgery, Children’s Hospital of Buflalo, University of Buffalo, and the State University of New York, Buffalo, NY. Supported in part by the Women & Children’s Health Research Foundation, Inc. Children’s Hospital, Buffalo, NY, and the James H. Cummings Foundation, Inc. Buffalo, NY. Presented at the 20th Annual Meeting of the American Pediatric Surgical Association, Baltimore. Maryland, May 28-31 I 1989. Address reprint requests to Donald R. Cooney, MD, Head, Department of Pediatric Surgery, ChildrenS Hospital of Buffnlo. 219 Bryant St, Buffalo, NY 14222. o 1990 by W.B. Saunders Company. 0022-3468/90/2502-0009$03.00/0

229

FOWUR ET AL

230 our laboratory by serial passage both in vivo and in vitro. Only in vivo passaged tumor was used in the present study. For passage in vivo, tumor cell suspensions were prepared by aseptically excising subcutaneous tumor tissue, then mechanically disaggregating it in phosphate-buffered saline (PBS) (pH, 7.2). Viability of cells was determined by trypan blue exclusion. A suspension of 1 x lo6 live tumor cells was then injected subcutaneously in the right flank.

Cyclophosphamide CY was obtained from Sigma Chemical Co (St Louis, MO). Each dose of CY was administered as a single intraperitoneal (IP) injection on the days indicated in the treatment protocol.

Surgical Procedure When tumors were 1.5 cm in diameter (10 days after the implantation), mice were anesthetized with sodium pentobarbital. Sharp resection of 85% of the tumor was performed under sterile conditions, and the wound was closed primarily.

Treatment Schedules Eighty-four mice with l&cm Cl300-NB tumors were divided into the following experimental groups (Fig 1): I, untreated controls (n = 14); II, operative controls, with 85% tumor resection (n = 18); III, sham-operated controls, with a skin incision and partial mobilization of the tumor but no resection (n = 18); IV, no tumor resection, and multiple-dose CY (n = 6); V, 85% tumor resection and single-dose CY (n = 14); VI, 85% tumor resection and multipledose CY (n = 14). CY was administered (100 mg/kg IP) to groups IV, V, and VI on day 11 (24 hours after surgery in operated groups V and VI). Mice in groups IV and VI also received low-maintenance doses of CY each week (25 mg/kg). Treatment was continued until 60 days after initial tumor implantation.

Immunologic Studies Lymphocyte phenotype studies. Postbulbar blood was collected from groups I to VI on day 10 (prior to surgery), day 13 (2 days after CY treatment, postoperative day 3 [POD 3]), then weekly until day 60. The blood collections were performed 2 days after CY treatment (each week). Lymphocyte subsets were labeled with the monoclonal antibodies antiLyt 2,3 (antisuppressor/cytolytic T cells) followed by

Day

0

10

I ,_ .

Cl300

I

I

Untreated

II

OP

I I Sham OP

IV

CY

v

OPtCYxl

VI

OPtCY/wk

11

..I ^, I f t op CY-100

FITC (second antibody), and antiIgG-FITC (anti-IgG+ B cells; all were obtained from Becton Dickinson, Mountain View, CA). Flow cytometric analysis was performed with Cytofluorograf, System 50-L (Ortho Instruments, Westwood, MA). Twelve to 24 normal mice without tumors served as controls for each time period. White blood cell (WBC) and differential counts. WBC counts by the standard Coulter counter method and differential counts from Wright-stained smears were performed concurrently with lymphocyte subset determinations in groups I to VI. Twelve to 24 mice served as normal controls. Reimplantation of tumor. All mice from groups I to VI noted to be cured of the tumor on day 60 after initial tumor implantation were observed for an additional 30 days with no further treatment. If no tumor regrowth was apparent, the animals were reimplanted with 1 x lo6 Cl300-NB cells. Mice were observed for another 60 days for reappearance of the tumor. Three comparison groups were formed, including group 1 with 1.5-cm tumors treated with very high-dose CY (300 mg/kg, n = 4). Group 2 animals, also with 1.5~cm tumors, were treated w?& 300 mg/kg CY one day after 85% tumor resection (n = 4). Group 3, with early 0.5-cm tumors, was treated by complete excision (n = 4). With no regrowth of neuroblastoma observed in 30 days, 1 x lo6 Cl 300-NB cells were reimplanted. Mice were examined daily for 60 days for signs of tumor growth. T cell Concanavalin A (Con-A) blastogenesis assay. Thirty-four mice were divided into three groups to test T cell suppressor function. Group 1 consisted of normal mice (n = 12); group 2 (n = 11) consisted of mice bearing 1.5-cm tumors; group 3 mice (n = 11) also had 1.5-cm tumors, but were treated with a single dose of CY (100 mg/kg). Suppressor function was tested 48 hours after CY adminisPooled peripheral blood (1 mL) from three mice in each group was obtained by retrobulbar puncture, diluted at a 1:2 ratio in Hanks balanced salt solution (HBSS) and applied to a gradient of Histopaque 1.083 (Sigma Chemical Co, St Louis, MO). Following centrifugation at 1,500g for 30 minutes at 25“C, the mononuclear cell layer was removed and washed once in HBSS and twice in RPM1 1640 media. The cell pellet was resuspended in 0.5-mL AIM-V serum-free media (GIBCO, Grand Island, NY) that was supplemented with 5 x 10e5 mol/L 2-mercaptoethanol. In duplicate, 3 x 10’ lymphocytes were diluted to a total volume of 0.5 mL

18

25

I

I

r

CY-25

t

CY-25

60

‘+

t

(+I

t

t

Fig 1. All groups were injected with C1300-NB (1 x 10’ cells) subcutaneously on day 0, then received various treatment regimens including partial tumor resection end cyclophosphemide starting on day 10. OP = 86% tumor resection; CY-100 = CY 100 mglkg IP: CY-25 = CY 25 mg/kg IP.

IMMUNOTHERAPY

FOR MURINE NEUROBLASTOMA

231

with supplemented AIM-V. One of the two tubes received 20 @/mL Con-A. Samples were incubated at 37OC in 5% CO, for 48 hours. Tritiated thymidine (0.05 microcuries per tube) was added to all tubes, and incubation continued for an additional 24 hours. The cells were washed three times in PBS, 0.25 mL of a 1:2 ratio Protosol (NEN-DuPont, Wilmington, DE)/ethanol solution was added, then 10 mL of Liquifluor scintillation fluid (NEN-Dupont, Wilmington, DE) was added for determination of tritiated thymidine incorporation with an LS-31331 Liquid Scintillation Counter (Beckman Instruments, Fullerton, CA). This procedure was performed in quadruplicate, and the results of stimulation indexes are expressed as group means + SE. The stimulation index (SI) was determined using the formula: SI =

mean experimental counts per minute (with Con-A) mean background counts per minute (without Con-A)

In vivo thymocyte depletion. Five mice with 0.75cm tumors were treated with CY (100 mg/kg) on day 8, then received antithymocyte serum (ATS; 0.25 mL IP) on days 10, 12, and 14, as in a protocol published by Dray and Mokyr.” Survival was compared with that of control groups that received ATS alone (n = 5) and CY alone (n = 5). In vivo inactivation of IJ’ cells. Four mice with l-cm (7-day) tumors were treated with anti-IJ’ serum (Accurate Chemical and Scientific Corp. Westbury, NY). Mice received 0.2 mL of antiserum daily by a tail vein injection on days 7.8, 10, and 13. Tumor growth was determined daily as the product of greatest width times length, and reported in square centimeters. The growth curves of untreated tumors were used for comparison.

Statistical

Analysis

of Data

Survival was followed for 60 days after tumor implantation, and groups were compared with untreated control group I using the log-rank test. P values less than or equal to .05 were considered statistically significant. Additionally, survival of groups was expressed as means + SD, but was included for gross comparison only. Other data were compared by ANOVA and Neuman-Keuls multiple comparison tests. RESULTS

Survival

There was no statistical difference in survival among the three control groups (I, no treatment; II, 85% tumor resection; III, sham operation) (Fig 2, Table 1). Mean survival in these groups was 23 to 24 days. Survival was significantly increased in all three groups receiving CY treatment (IV to VI). When resection was followed by a single postoperative dose of 100 mg/kg CY (group V), survival was significantly increased over that of controls (P c .OOl), and one (7%) tumor cure was noted. Survival was improved further with the addition of weekly 25mg/kg maintenance doses of CY (group VI, P < .OOl). Mean survival of this group increased to 51 f 9.9 days, and seven mice (50%) lived longer than 60 days. Four of these seven mice were cured of the tumor, and three were alive at 60 days with the tumor present. Groups IV and VI received identical multi-dose CY treatment; however, the tumor was not resected in

l

I

0 II

Untreated

A IV CY

OP

c v

6 III Sham OP

.

OP+CYx1

VI OP+CY/wk

Kaplan-Meier survival curves of treatment groups I to VI Fig 2. are shown. Pvalues (log-rank test) are shown in Table 1.

group IV. Surgical resection in group VI increased survival significantly (P < .05), with mean survival changing from 33 -I 2.6 days to 51 * 9.9 days. Cell Populations WBC and differential counts. WBC counts increased directly with tumor growth, from 6,152 2 1,714/mm3 in normal mice, up to 20,000 to 70,000/ mm3 in mice with very advanced tumors. As demonstrated by the differential counts, the leukocytosis reflected a proliferation of polymorphonuclear cells (PMNs), which increased from 23.5 rt 10.3% in normal mice to 80 to 95% of the leukocytes in mice with very advanced tumors (Fig 3). Elevation of the PMN count, as determined by either differential or absolute count, was a sensitive indicator of the presence of tumor and preceded the gross appearance of the tumor. The WBC and PMN responses were not accompanied by signs of infection or acute inflammation in the mice. Inversely related to the increasing number of PMNs in tumor-bearing mice was a simultaneous decrease in lymphocyte counts (Fig 3). The normal range for mice was 69.8 + 10.9%, but this diminished to about 5% in mice with very advanced tumors. Both B cell and Tsupp (Lyt 2,3+) cell counts were affected by the decreasing number of lymphocytes in mice with advanced tumors. For purposes of discussion, Lyt 2,3+ cells will be referred to as Tsupp cells, although it is

FOWLER ET AL

232

Table 1. Therapeutic

Groups

”

I Untreated II Operated Ill Sham-operated IV CY (100

+ 25/wk)

V Operation + CY ( 100) VI Operation + CY (100

+ 25/wk)

Results No.

Surviving at 60 Days

Survivalin Days (mean * SD)

No. of ClWS

14

23 f 2.3

0

0

18

24 ” 4.4, NS

0

0

18

23 +- 4.5, NS

6

33 f 2.6, P<

.05

0

0

0

0

14

35 f 9.1, P-e

,001

1 (7%)

1 (7%)

14

51 f 9.9, PC

,001

7 (60%)

4 (29%)

NOTE. P values were determined by log-rank test from survival data shown in Fig 2.

recognized that lymphocytes with other functions are also identified with this monoclonal antibody. To assess the effect of the various treatment regimens on WBC, B cell, and Tsupp cell counts, data were analyzed only up to day 20 (POD 10). After this time, the influence of morbid animals made data difficult to interpret. White blood cell counts steadily increased from day 10 (7,063 + 2,468.3/mm3) through day 20 (13,842 + 5,789.7/mm3) in untreated tumor-bearing mice (group I) and sham-operated mice (group III); both groups showed increasing tumor burdens (Fig 4). On day 13 (POD 3) WBC counts were within the range of normal control mice in groups treated by surgery alone (II) or CY alone (IV). Effects of combined surgery and CY treatment (groups V and VI) appeared to be additive in that a significant WBC depression (1,679 + 1,094.0/mm3, P < .OOl) was apparent on day 13. One week later, on day 20, WBC counts of groups I (untreated), III (sham-operated), and IV (CY alone) were all significantly elevated (P < ,001) as compared to group II (surgery alone) and groups V and VI (surgery plus CY regimens).

Absolute lymphocyte counts were depressed significantly in mice with lo-day tumors (Fig 5). The counts remained depressed in untreated mice (groups I and II), while partial resection (III) reverted counts to normal on day 17 (POD 3). CY (100 mg/kg) brielly depressed the counts further on POD 3, whether combined with surgery (V and VI) or not (IV). By day 20, lymphocyte counts were depressed in animals with growing tumors (I to V), but were returning to normal in group VI. Normal values were reached in group VI on day 28. Tsupp (Lyt 2,3’) and B cell counts. Tsupp cell counts were significantly depressed from the normal control values of 2,159 f 754.8/mm3 down to 782 + 16000 1600014000.

m 12000. g 10000~ Qp

100

5 0

I

8000. 6000 -

PMN’s 4000 * 2000

I 10 13 Days

0 Lymphocytes

b

io

Fig 3. increased decreased

= I

Untreated

A IV

CY

0 II

OP

0 v

OP+CYxl

+ III Sham

i3

Days Polymorphonuclear cell counts from peripheral blood directly with tumor growth, while lymphocyte counts proportionately.

I 20

OP

l

VI OP+CYlwk

Fig 4. WBC counts increased with tumor growth 8 and 1111. WBC counts briefly stabilized or were reduced by treatments (II. IV, and VII on day 13. By day 20. WBC counts again increased in groups having en increasing tumor mass. On day 20, groups Ill. V, and VI were near-normal. while I. II. and IV were significantly elevated (P -z .QOl).

IMMUNOTHERAPY FOR MURINE NEUROBLASTOMA

233

V and four in group VI. Cl 300-NB did not grow in any of these five surgery-plus-CY-treated animals after reimplantation. In contrast, in all mice cured of early tumor by complete surgical excision (nonimmunologic treatment), tumor thrived after reimplantation. Early tumors were used for excision because a cure was not assured in mice with more advanced tumors. Advanced 1.5-cm tumors could not be eradicated by treatment with high-dose CY (300 mg/kg), whether combined with 85% surgical resection or not; therefore, reimplantation experiments could not be performed in this group.

1 I

1000

‘

I-

I

10 13 Days

O

20

mI

Untreated

A IV CY

0 II

OP

0 v

l

Ill Sham OP

oP+cYxl

. VI OP+CYlwk

Fig 5. Total lymphocyte counts were depressed in the presence of tumor tissue. Partial tumor reseotion (III) briefly normelized the count in the immediate postoperative period, but counts diminished with tumor regrowth. Normal counts were reached on day 28 in cured mice from groups V and VI.

333.2/mm3 on day 10 of tumor growth (P < .OOl), and they remained depressed in all groups through day 20 (Fig 6). This reflected an overall depression of total lymphocyte counts that resulted from residual tumor. Despite this underlying lymphopenia, changes resulting from the various treatments were apparent. On day 13, combined surgery plus CY groups (V and VI) demonstrated further depression of Tsupp counts that were significantly lower than untreated group I (P = .006). On day 20, the Tsupp cell counts of all groups that received CY (IV,V,VI) were still significantly lower than those of the groups that did not receive CY (I,II,III). Like Tsupp cells, IgG+ B cell counts were affected by the depression in total lymphocyte counts. B cell counts decreased from 1,075 f 444.4/mm3 on day 0 to 443 i 190.7/mm3 on day 10 of tumor growth. No further change was noted in any group on day 13 or day 20. In mice that were cured of tumor (from V and VI), the WBC, B cell, and Tsupp cell counts reverted to normal control levels by day 38 (POD 28).

T Cell Con-A Blastogenesis

Compared with the SI of T cells from normal mice, a significant suppression of T cell blastogenesis was demonstrated in tumor-bearing animals (Fig 7). Treatment of tumor-bearing mice with low-dose CY eradicated this tumor-associated suppression. Thymocyte Depletion

To evaluate the role of T cells in the therapeutic effect observed with CY, T cells were eliminated with antithymocyte serum. Four of five mice (80%) with 0.75-cm C1300-NB tumors treated with a single dose of CY 100 mg/kg were cured of the tumor. In

b

io

Days . I

Untreated

A IV CY

0 II

OP

0 v

+ III Sham OP Reimplantation

To evaluate whether host resistance to tumor had developed as a result of treatment, all mice with complete eradication of tumor were reimplanted with tumor. There was one animal cured of tumor in group

i3

l

OP+CYxl

VI OP+CY/wk

Fig 8. Tsupp cell counts were depressed in the presence of tumor. as were total lymphocyte counts. Counts were further depressed by CY, especially by the combination of OP + CY (V and VI) on day 13. At 20 days. Tsupp counts of group5 receiving CY (IV to VI) were significantly lower than those of group5 not receiving CY (I to Ill).

FOWLER ET AL

234

80 70 : u _c

60 50

ig

40

z

20

+ t

10 z

30 I

T I

Untreated

I f I Cl306

Cl3gQlcY

Fig 7. Compared with the 81 of T cells from normal mice, a significant suppression of T cell blastogenesis was noted in tumor-bearing mice W = .0061. This tumor-associated suppression was eradicated by treatment with CY. Values are expressed as group meens + SE.

comparison, this curative effect of CY was abrogated by the administration of antithymocyte serum. The tumor continued to enlarge in all of these mice, and mean survival was similar to that of untreated animals. Antithymocyte serum had no affect on survival of tumor-bearing mice that did not receive CY. In Vivo Inactivation of ZJK Cells The role of IJK cells in tumor growth was investigated by inactivation with anti-IJK serum. Compared with the growth rate of untreated tumors, significant suppression of the normal tumor growth rate was observed on days 10 to 12 following treatment with anti-IJK serum (Fig 8). When treatment was interrupted, normal tumor growth resumed by 48 hours, showing the sensitivity of tumor growth to IJK cell activity. One additional dose of anti-IJK serum on day 13 resulted in an immediate suppression of tumor growth for another 24 hours before normal tumor growth resumed.

phenomenon of spontaneous resolution that is occasionally observed in neuroblastoma may be an immunologic event.6*16 In the present study the therapeutic effects of postoperative immunomodulating doses of CY were evaluated in a murine model of partially resected, advanced neuroblastoma. This model was chosen to simulate surgical debulking of advanced, unresectable human neuroblastoma. It is now well recognized that CY produces immunomodulating effects that are timeand dose-dependent. High doses of CY are used therapeutically for direct drug-induced tumoricidal effects; however, immunosuppression of both cellular and antibody responses is a toxic side effect, and response of advanced tumors is poor. Lower doses of CY are more selective in their effects. Doses of CY lower than 100 mg/kg have minimal effects on B cells; however, doses greater than 100 mg/kg have been found to severely deplete B cells in experimental mice.17 T lymphocyte-mediated suppressor function,2*9 and possibly the number of Tsupp cell precursors,‘* are depressed by low-dose CY. This property of low-dose CY is important in tumor immunology because some tumors avoid destruction by the host antitumor response by increasing a population of Tsupp cells that down-regulates the immune response before sufficient numbers of effector T cells are generated to mediate tumor regression.13 During tumor growth, the downregulation of tumor-induced Tsupp cell activity with appropriate therapy, such as with low-dose CY, could contribute towards reestablishing the effectiveness of

1

12

control

*IJ

DISCUSSION

Standard therapy for neuroblastoma using highdose chemotherapy has not improved survival significantly in patients with advanced tumors. However, because neuroblastoma is an immunogenic tumor, immunotherapy may provide an alternate, more successful mode of therapy. In previous work we demonstrated the ability of C1300-NB to evoke an immune respqnse by way of prolonging host survival in mice preimmunized with Cl 300-NB membrane preparations.3 Similarly, human neuroblastoma has been described as immunogenic, and it is believed that the

I , , , , , , , , , , ~ 0

10

14

12

16

18

day Fig 6. Anti-IJK serum IO.2 mL per dose) was injected via the tail vein on days 7, 6. 10. and 13 into mica with l-cm tumors. Compared with the growth rate of untreated C1300-NE. growth of treated tumors was significantly delayed on days 10 to 12 and 13 to 14 (P = .02).

IMMUNOTHERAPY

FOR MURINE

NEUROBLASTOMA

host antitumor immunity.” Regression of established experimental tumors has been shown to follow the down-regulation of Tsupp cell function.10”4*‘5 In the current study, the best therapeutic regimen for C 1300-NB consisted of a surgical debulking procedure followed by multi-dose CY therapy (group VI). Surgery alone (group II) did not prolong survival over that of control animals. However, the value of resection in multimodal therapy is evident by the increased survival rate of animals receiving both multidose CY therapy plus surgery (group IV) as compared to the lower survival rate of animals that received multidose CY therapy without surgery (group IV). Naito et al” also found that C 1300-NB tumor growth was impaired and host survival prolonged with a combination of surgical debulking and postoperative low-dose CY. Clinical experience supports operative reduction of tumor mass in patients with neuroblastoma.21T22 In our study, surgical debulking of the tumor may have been beneficial because it briefly stabilized tumorassociated leukocytosis and Tsupp counts, and normalized total lymphocyte counts. During the short postoperative period in which the immune status was optimized, immunomodulating doses of CY were administered. Radov et a123showed that the effect of CY is dependent on the status of the host’s immunologic responsiveness at the time of treatment. In determining the optimal time to start postoperative chemotherapy, Naito et al” analyzed cell kinetics in a C1300-NB model after 50% cytoreductive surgery, and found that tumor growth was impaired and animal survival improved if CY (75 mg/kg) were given in the early postoperative period. No difference in outcome was observed whether CY was administered 24 or 72 hours postoperatively. In the present experiment, CY therapy was initiated 24 hours after surgery when tumor proliferation has been shown to be greatest and should be most susceptible to chemotherapy.20 In addition, low doses of CY were administered weekly to maintain Tsupp cell suppression, because regeneration of the T cells following CY therapy has been shown to begin at about 7 days and be complete by 30 days.24,25The value of these maintenance doses is evident when the improved outcome of group V (surgery plus single-dose CY) is compared with that of group VI (surgery plus multiple-dose CY). There is also support for using two doses of CY because different ranges of low-dose CY have been shown to inactivate different T supp cell populations that are generated during early and late stages of tumor growth.‘3*26 We demonstrated increased suppressor activity in mice bearing C1300-NB, which should render the tumor susceptible to the ability of CY to downregulate enhanced tumor-associated suppression. This

235

effect was evident in the present neuroblatoma model, and the increased suppressor activity in tumor-bearing mice was eradicated by treatment with one dose of CY 100 mgfkg. As in other studies,“q2’ we found that treatment with antithymocyte serum abolished the therapeutic effect of low-dose CY, which indicates that low-dose CY is effective through T cell mechanisms. C1300-NB tumor growth was delayed by in vivo inhibition of IJK cells with anti-IJK serum. Because cells bearing the IJ locus of the major histocompatibility complex are restricted to cells in the Tsupp cell network,28 these data suggest that C1300-NB tumor growth depends on enhanced Tsupp function. Furthermore, because the IJK Tsupp cell network can be stimulated by IJK-compatible macrophages acting as antigen-presenting cells, a tumor-associated antigen may be responsible for increased Tsupp cell function in the present model. Tumor debulking may have contributed to the success of immunotherapy in this model because it decreased the antigen load. Because the normal rate of tumor growth immediately returned after cessation of anti-IJK serum, it is suggested that there is a functional interference and not a depletion of an IJK subset of cells. The effect of in vivo depletion would have been slower in onset and longer in duration. The possibility that tumor-associated Tsupp cells have a role in human neuroblastoma is suggested in a study by Okabe et al,’ who found that 35% of the differential counts from peripheral blood samples in children with neuroblastoma showed an increase in the percentage of Tsupp cells. In these children, delayedtype hypersensitivity reactions were enhanced. Tsupp cell counts and delayed-type hypersensitivity appeared to be markers of disease that returned to normal when the tumor was eradicated or in remission, and increased when the tumor reappeared. Although most children in the study had normal Tsupp counts and decreased delayed-type hypersensitivity reactions, they usually manifested advanced disease with generalized immunosuppression. In contrast to the immunomodulating effect of low-dose CY, high-dose CY has been shown to act primarily by direct tumoricidal-tumoristatic action.19927 In the present model, advanced C1300-NB was not sensitive to high-dose CY and was not eradicated by this treatment, whether or not the tumor burden was reduced by surgical resection. This is a characteristic shared by human neuroblastoma. C1300-NB in mice responded to low-dose CY, but it remains to be seen whether human neuroblastoma will also respond. There are reports of other experimental and human tumors that are as responsive or more responsive to low-dose than high-dose CY. 27*29 One preliminary clinical report of patients with multiple myeloma refractory to L-

FOWLER ET AL

236

PAM and prednisone showed that weekly low-dose CY therapy (150 to 300 mg/m2) in combination with alternate-day prednisone resulted in successful palliation in three of five patients.30 Furthermore, palliation was achieved with little or no drug toxicity. In another clinical trial, complete regression of metastatic melanoma was noted in two of nine patients treated once a month with CY (300 mg/m2) 3 days before an autologous melanoma vaccines31 Seven of these CYpretreated patients developed delayed-type hypersensitivity responses to autologous melanoma cells compared with only two of seven melanoma-vaccinated controls (P = .034). This CY-induced augmentation of cell-mediated immunity together with the therapeutic response observed suggested to the authors that tumor-directed Tsupp cell activity is important in the particular tumor-host interaction. In clinical cases in which low- and high-dose chemotherapy may be equally effective in tumor therapy, the decision to use one dose over the other may depend on other variables, such as resulting host resistance to subsequent challenge. We demonstrated that mice cured with low, immunomodulating doses of CY were resistant to growth of reimplanted tumors. Although in the present model, advanced C1300-NB could not be eradicated by the nonimmunologically mediated tumoricidal action of very high-dose CY, in other tumor models that can be cured with high doses of CY, it has been documented that reimplanted tumor thrives in the hosts.” Because complete surgical excision and cure of an early tumor was possible with C1300-NB, this was used as a model of nonimmunologically mediated cure. Growth of subsequently implanted tumor in these mice was prompt-similar to the response of the other tumor models cured by high-dose CY. Because immunomodulating doses of CY evoke an antitumor immune response that can be recalled in

the presence of repeated exposure to the tumor, primary tumor recurrence or progression of tumor implants and metastases might be prevented.lg Furthermore, low-dose immunotherapy may be beneficial to human tumors that behave clinically like C1300-NB; that is, they are resistant to the tumoricidal action of high-dose chemotherapy but responsive to lower doses that recruit host antitumor immunity to aid in tumor eradication. Although it appears that Tsupp cells (Lyt 2,3+) and IJ+ cells have important roles in Cl 300-NB tumor growth, the identification of possible host effector cells is not complete. Significant tumor-associated neutrophilia was observed in A/J mice bearing C1300-NB tumors. A similar phenomenon has been reported in C3H/HeJ mice having C3HBA breast adenocarcinoma. In the latter study, it was suggested that a neutrophil-stimulating factor was secreted by the tumor. It is possible that neutrophils may play a role in the host immune response to these tumors. The development of successful immunotherapies for immunogenic tumors depends on identification of both the tumor-induced mechanisms contributing to tumor growth, as well as the host effector cell populations responsible for the antitumor response. Because evidence suggests that tumor-associated Tsupp cell populations are important to experimental neuroblastoma growth, a trial of clinical treatment of neuroblastoma patients with low-dose CY to inactivate these populations and support the host antitumor response may be justified, especially in the face of treatment failures with high-dose CY, ACKNOWLEDGMENT We would like to acknowledge the excellent technical assistance of Gary Rich and Bela Viszt, as well as Miles Pharmaceuticals for providing the “mousometer.”

REFERENCES 1. Okabe I, Kuroso Y, Morita K: Cell-mediated immune reactions to clinical neuroblastoma. Jpn J Surg 15:368-374,198s 2. Berd D, Maguire HC Jr, Mastrangelo MJ: Potentiation of human cell-mediated and humoral immunity by low-dose cyclophosphamide. Cancer Res 44:5439-5443.1984 3. Bruce J, Brooks SP, Rich GA, et al: Effects of immunostimulation on host survival in A/J mice with transplantable Cl300 neuroblastoma. Curr Surg Jan-Feb: 17-19, 1988 4. Kawai K, Sakatoku H, Matsuda T, et al: Development of autologous specific reactivity to human neuroblastoma. Pediatr Res 20~915-919, 1986 5. Everson TC, Cole WH: Spontaneous Regression of Cancer. Philadelphia, PA; WB Saunders, 1966, pp 88-163 6. Squire R, Fowler CL, Rich GA, et al: The relationship of class I MHC antigen expression to stage IV-S disease and survival in neuroblastoma. J Pediatr Surg (in press) 7. Sigal RK, Reynolds JV, Markmann JF, et al: Upregulation of

MHC class I: Effect on growth and LAK sensitivity of neuroblastoma. Surg Forum 39:572-575, 1988 8. 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 9. Berd D, Maguire HC Jr, Mastrangelo MJ: Impairment of Concanavalin A-inducible suppressor activity following administration of cyclophosphamide to patients with advanced cancer. Cancer Res44:1275-1280, 1984 10. Greene MI, Perry LL, Benacerraf B: Regulation of the immune response to tumor antigen. Am J Pathol95:159-169, 1979 11. Dray S, Mokyr MB: Immunomodulation by cyclophosphamide and its effect on eradication of established tumors, in Fundenberg HH, Whitner HD, Ambrog F (eds): Immunomodulation: New Frontiers and Advances. New York, NY, Plenum 1984, pp 349-361 12. North RJ, Awwad M: T cell suppression as an obstacle to

IMMUNOTHERAPY

FOR MURINE NEUROBLASTOMA

237

immunologically-mediated tumor regression: Elimination of suppression results in regression, in Truitt RL, Gale RP, Bortin MM (eds): Cell Immunother of Cancer. New York, NY, Liss, 1987, pp 345-358 13. DiGiacomo A, North RJ: T cell suppressors of antitumor immunity. J Exp Med 164:1179-l 192, 1986 14. Dye ES, North RJ: Adoptive immunization against an established tumor with cytolytic versus memory T cells. Transplantation 37:600-605.1984 15. North RJ: Cyclophosphamide-facilitated adoptive immunotherapy of an established tumor depends on elimination of tumorinduced suppressor T cells. J Exp Med 55:1063-1074, 1982 16. Ziegler MM, Vega A, Koop CE: Electrocoagulation induced immunity-An explanation for regression of neuroblastoma. J Pediatr Surg 15:34-37, 1980 17. Marbrook J, Baguley BC: The immunological status of mice during the generation of cyclophosphamide-induced tolerance. Cell Immunol25:217-227,1976 18. Iwaguchi T, Nakamura M, Kitagawa H: Host-mediated inhibition of rat bladder cancer growth by cyclophosphamide and purine salvage pathway-related enzyme activity of lymphocytes. Jpn J Exp Med 54:201-206.1984 19. Mokyr MB, Dray S: Interplay between the toxic effects of anticancer drugs and host antitumor immunity in cancer therapy. Cancer Invest 5:31-38, 1987 20. Naito H, Ziegler MM, Tsou KC: Rational selection of adjuvant chemotherapy after cytoreduction surgery for murine neuroblastoma. Cancer Res 45:3554-3560, 1985 21. Block JB, Isacoff WH: Adjuvant chemotherapy in cancer. Semin Oncol4:109-115, 1977 22. Fortner J, Nicastri A, Murphy ML: Neuroblastoma: Natural history and results of treating 133 cases. Ann Surg 167:132-142, 1968

23. Radov LA, Haskill JS, Korn JH: Host immune potentiation of drug responses to a murine mammary adenocarcinoma. Int J Cancer 17~773-779, 1976 24. Sy MS, Miller SD, Claman HN: Immune suppression with supraoptimal doses of antigen in contact sensitivity. I. Demonstration of suppressor cells and their sensitivity to cyclophosphamide. J Immunol 119:240-244, 1977 25. Dumont F: Destruction and regeneration of lymphocyte populatins in the mouse spleen after cyclophosphamide treatment. IntArchAllergy47:110-123, 1974 26. North RJ, Bursuker I: Generation and decay of the immune response to a progressive fibrosarcoma. J Exp Med 159:1295- 13 11, 1984 27. Mokyr MB, Dray S: Some advantages of curing mice bearing a large subcutaneous MOPC-315 tumor with a low dose rather than a high dose of cyclophosphamide. Cancer Res 43:3 112-3 119,1983 28. Gershon RK, Eardley DD, Durum S, et al: Contrasuppression: A novel immunoregulatory activity. J Exp Med 153:1533-1546, 1981 29. Mathe G, Halle-Pannenko 0, Bourut C: Effectiveness of murine leukemia chemotherapy according to the immune state. Cancer Immunol Immunother 2: 139- 141, 1977 30. Brandes LJ, Israels LG: Treatment of advanced plasma cell myeloma with weekly cyclophosphamide and alternate-day prednisone. Cancer Treat Rep 66:1413-1415, 1982 31. Berd D, Maguire HC Jr, Mastrangelo MJ: Induction of cell-mediated immunity to autologous melanoma cells and regression of metastases after treatment with a melanoma cell vaccine preceded by cyclophosphamide. Cancer Res 46:2572-2577, 1986 32. Lan S, Rettura G, Levenson SM, et al: Granulopoiesis associated with the C3HBA tumor in mice. JNCI 67:1135-l 138, 1981

Discussion M. Ziegler (Philadelphia, PA): We concur completely with the type of therapeutic approach put forward in this report. That is, tumor bulk-reducing surgery followed by adjuvant chemotherapy or immunotherapy. In 1985, Dr Naito, working in our laboratory, made the same observation in the C1300-NB model. He reported that the antitumor activity of adjuvant CY was demonstrable only in hosts that had an intact immune system, when immunocompetence was impaired by thymectomy and irradiation, CY was no longer effective as an antitumor agent. Similarly, we subsequently showed that for CY to be an effective antitumor therapy in a low-dose antisuppressor cell format that the host must bear an immunogenic tumor. We have two questions for you, Dr Fowler. (1) How would the investigators protect or enhance the host’s own antitumor mechanism when using CY so that effective antitumor immune activity can be achieved? (2) How would the investigators suggest that the antigenicity of the putatively nonimmunogenic human neuroblastoma be augmented so that such immunotherapy might be expected to work in patients?

C. Fowler (response): To answer your first question, low-dose CY is being used as an immunotherapeutic agent because, at the doses we tried, it specifically down-regulated only Tsupp cell function. It does not eliminate T ceil activity or affect other immunologic functions. Tsupp cell function is specifically eliminated. An advantage is the fact that whether CY is opposed to high-dose CY, which is generally immunosuppressive, low-dose cyclophosphamide is suppressive in its effects. Human neuroblastoma is a heterogeneous tumor that has low antigen expression. We now have studies ongoing to increase class I expression so that immunotherapy will be more effective. As alluded to in a previous report, retinoic acid is one of the agents that occasionally increases class I expression. Certainly, interferon has been shown many times to increase class I expression, but now we are engaged in several experiments to increase class expression with combination immunotherapy.