Recruitment of host CD8+ T cells by tumor-infiltrating lymphocytes and recombinant interleukin-2 during adoptive immunotherapy of cancer

Recruitment of host CD8 + T cells by tumor-infiltrating lymphocytes and recombinant interleukin-2 during adoptive immunotherapy of cancer Ulrike L. Burger, MD, a Maximilian P. Chang, BA, Peter S. Goedegebuure, PhD, and Timothy J. Eberlein, MD, Boston, Mass.

Background. Previously we demonstrated that optimal doses of tumor-infiltrating lymphocytes (TIL) concomitant with recombinant interleukin-2 (rlL-2) effectively mediated complete tumor regression of murine 3-day pulmonary metastases. Methods. In the present study we have investigated the contribution of the host immune response to the effectiveness of adoptive immunotherapy with TIL in combination with low-dose rIL-2. All experiments were performed in a murine pulmonary metastases model induced by intravenous injection of methylcholanthrene-induced sarcoma (MCA- 705) cells into C57BL/6 mice. As a novel approach we used monoclonal antibody specific for CDd + or CD8 + T cells to deplete the host of its T-cell subpopulations. Results. Depletion of host CD8 + T cells 24 hours after tumor injection and d8 hours before TIL + rlL-2 treatment abrogated all antitumor activity of this type of immunotherapy and resulted in significant metastatic pulmonary disease (p < O.OOI). In contrast, depletion of host CDd + T cells did not alter the efficacy of TIL + rIL-2 treatment in tumor eradication. The loss of tumoricidal activity of TIL + rIL-2 treatment in a CD8 + T cell-depleted host could be overcome by adding back normal uneducated splenocytes 2 hours after TIL therapy (p < 0.001). In contrast, adding back CD8- CD4 + splenocytes to a CD8 + T cell-depleted host 2 hours after TIL + rlL-2 treatment resulted in significant pulmonary disease comparable to untreated animals. Conclusions. We conclude that the recruitment of host CD8 + T cells by adoptively transferred TIL + rlL-2 appears to be important for effective tumor eradication in this type of immunotherapy. (SURGERY 1995;117.'325-33.) From the Department of Surgery, Division of Surgical Oncology, Brigham ~ Women's Hospital, Harvard Medical School, Boston, Mass.

ADOPTIVE IMMUNOTHERAPY WITH tumor-infiltrating lymphocytes ( T I L ) and interleukin-2 (IL-2) has proved to be 50 to 100 times more potent in eliminating micrometastatic pulmonary disease in murine tumor models than comparable therapy with lymphokine-activated killer (LAK) cells) '2 Although it has been assumed that this increased therapeutic effect is related

Supported by the National Institutes of Health grant CA 45484 and the American Cancer Society Faculty Research Award FRA-407. Accepted for publication June 27, 1994. Reprint requests: Timothy J. Eberlein, MD, Department of Surgery, Division of Surgical Oncology,Brigham & Women's Hospital, Harvard Medical School, 75 Francis St., Boston, MA 02115. aSurgical research fellow from the Surgical Hospital, University of Heidelberg, Heidelberg, Germany. Copyright 9 1995 by Mosby-Year Book, Inc. 0039-6060/95/$3.00 + 0 11/56/59150

to the specificity of the adoptively transferred T I L , little information has been published that discriminates the contribution of adoptively transferred T I L versus the contribution of the host immune system to the observed responses. This differentiation has been difficult because most therapeutic strategies based on T I L include host immunosuppression by mild irradiation or cyclophosphamide administration),3 T h e therapeutic benefit of host immunosuppression has been explained by the elimination of host suppressor cells 4"8 such that the observed tumor eradication has generally been attributed to the adoptively transferred T I L population. However, more recent data suggest that the effect of irradiation in tumor immunotherapy models is due to direct effects on the tumor and not on the host immune system. 9 It has also been reported that L A K cell antitumor cytotoxicity is not nullified by irradiationJ ~ Taken together, these findings show that the mild im-

SURGERY

325

326 Burger et al.

munosuppression used in animal and human adoptive immunotherapy trials does not necessarily abolish host immune cell participation. Tumor regression caused entirely by the adoptively transferred T I L would appear to be unlikely. Instead, the observed therapeutic effect of T I L may be the consequence of a cooperative interaction between the adoptively transferred T I L population and an endogenous immune response rather than of the donor cells alone. We have designed a murine model to elucidate the immunologic mechanisms behind the antitumor efficacy of T I L + recombinant IL-2 (rIL-2) treatment by investigating the contribution of the host immune system. In previous murine experiments ll we showed complete tumor eradication of metastatic pulmonary disease after intravenous administration of T I L + rIL-2. With this methylcholanthrene-induced (MCA-105) sarcoma tumor system and monoclonal antibody (mAb) specific for CD4 + or CD8 + T cells, we show here that depletion of host CD8 + T cells 48 hours before T I L + rIL-2 treatment abrogated all antitumor activity usually associated with this type of immunotherapy. Adding back normal uneducated splenoeytes to a CD8 + T cell-depleted host 2 hours after T I L therapy could overcome the depletion of host CD8 + T cells and resulted in a highly significant reduction of pulmonary disease. Therefore the treatment with T I L + rIL-2 may operate by the recruitment of a host immune response for successful eradication of tumor. MATERIAL A N D M E T H O D S

Mice. Female C57BL/6 mice, 6 to 8 weeks old, obtained from The Jackson Laboratory, Bar Harbor, Maine, were used for all experiments. The animals were caged in groups of 10 or fewer. They were housed in the Dana-Farber Animal Facility under National Institutes of Health/Harvard Medical School approved animal subject conditions. All mice received animal laboratory food and water ad libitum. Tumor. The weakly immunogenic fibrosarcoma MCA-105 was provided by Dr. S. A. Rosenberg (National Cancer Institute, Bethesda, Md.). It was maintained in vivo by subcutaneous passage of 5 • 105 cryopreserved or fresh tumor cells in C57BL/6 syngeneic mice.S, lZ Use of MCA-105 was limited to the first eight passage generations. MCA-105 tumor was also used for generating T I L 2 to 3 weeks after subcutaneous injection when the tumor had reached a diameter of 1 to 2 cm.

Digestion of tumor. Tumor was dissected, minced, and digested by mechanical stirring in a solution of 0.1% collagenase type IV, 0.01% hyaluronidase type V, and 0.002% deoxyribonuclease type I (Sigma Chemical Company, St. Louis, Mo.) in 40 ml Hanks' balanced

Surgery March 1995

salt solution (HBSS) (Gibco Laboratories, Grand Island, N.Y.) for 2 to 3 hours at room temperature. The resulting single cell suspension was filtered through a sterile Nitex mesh (Lawshe Instrument Co., Rockville, Md.) and washed twice in HBSS. Recombinant IL-2. Purified human rlL-2 was supplied by Amgen (Thousand Oaks, Calif.) at a concentration of 3.0 X 106 IU/ml. Protein content was 0.36 mg/ml. Preparation of TIL. The resulting single cell suspension after MCA-105 tumor digestion was suspended at a density of 5 X 105 cells/ml in R P M I 1640 (Bio Whittaker, Walkersville, Md.) containing 10% fetal calf serum (Bio Whittaker), 2 mmol/L L-glutamine (Bio Whittaker), 1 mmol/L sodium pyruvate (Gibco Laboratories), 0.1 mmol/L nonessential amino acids (Gibco Laboratories), 100 IU/ml penicillin, 100 #g/ml streptomycin (both Bio Whittaker), 5 • 10 -s mol/L 2-mercapto-ethanol (Sigma Chemical C o m p a n y ) a n d 100 IU/ml rIL-2 (AMGEN). After activation on solidphase anti-CD3 flasks for 48 hours, 11 the cells were collected with cell scrapers (no. 3010; Costar, Cambridge, Mass.), washed, and resuspended in culture medium as described above. Every 3 to 4 days cultures were split by cell redistribution into fresh flasks with addition of new culture medium. In the first week of culture the tumor cells gradually disappeared, whereas the number of the T I L steadily increased. T I L were expanded for 4 to 5 weeks before they were used in in vivo experiments. At this point T I L cultures reproducibly showed a CD8 + phenotype of more than 95% and a CD4 + phenotype of less than 5% 11 (data not shown). Preparation of splenocytes. Spleens were removed aseptically and crushed with the hub of a 10 cc syringe in HBSS. The single cell suspension was filtered through a nylon mesh and washed once in HBSS. The cell pellet was resuspended in HBSS at a concentration of 1 • 10 7 nucleated cells/ml or less and subjected to a Lympholyte-M (Cedarlane Laboratories, Hornby, Ontario, Canada) gradient for the elimination of red blood cells, dead cells, and debris from murine spleens. After centrifugation for 20 minutes at 1200 rpm at room temperature, the lymphocyte layer was collected and washed twice in HBSS before cell count. Monoclonal antibody. Aseites-containing mAb was prepared by injecting 3 • 10 6 anti-CD8 (clone 2.43) or anti-CD4 (clone GK1.5) mAb-seereting hybridoma cells (both American Type Culture Collection, Rockville, Md.) intraperitoneally into pristane-primed athymic Swiss Nude mice (Taconic, Germantown, N.Y.). Ascites was harvested 10 to 15 days later by intraperitoneal drainage. After ultracentrifugation and sterile filtration (Nalge Company, Rochester, N.Y.) the supernatant was stored in aliquots at - 7 0 ~ C. Quantita-

Surgery

Burger et al.

327

Volume 777, Number 3 30

30 =---- CO4+ T CELLS

CO4+ T CELLS U') ,,-I ..J U.J U i.l.

9 20'

o

._f

CD8+ T CELLS

~

(wq --I .J u.l CJ

9

20

o

~J (.9

uJ ~.~

-

,<

u.J

10"

l,.u r O

\J i

I

CD8+ T CELLS

UJ O,. I

|

I

I

i

i

I

DayO Oayl Day2 Day3 Day4 Days Day6 Day7 Day8

DAYS AFTER DEPLETION

0

!

i

i

DayO Day1 Day2 Day3 Day4 Days Day6 Day7 Day8

DAYS AFTER DEPLETION

Fig. 1. Duration of host CD8 depletion. Anti-CD8 mAb ascites at a dilution of 1:150 was injected intraperitoneally into C57BL/6 mice on day 0. Phenotypic analysis of splenocytes was performed at different time intervals after host CD8 depletion. Data point day 0 represents percentage of CD4 + and CD8 + T cells in an undepleted control mouse. Data points days l, 2, 3, 4, and 8 represent T-cell percentages after CD8 depletion.

Fig. 2. Duration of host CD4 depletion. Anti-CD4 mAb ascites at dilution of 1:50 was injected intraperitoneally into C57BL/6 mice on day 0. Phenotypic analysis of splenocytes was performed at different time intervals after host CD4 depletion. Data point day 0 represents undepleted control mouse. Data points days J, 2, 3, 5, and8 represent T-cell percentages after CD4 depletion.

tive titration studies with ascites were performed in vivo in C57BL/6 mice by intraperitoneal injection to find the minimum concentration of mAb necessary for complete in vivo depletion of host T-cell subpopulations. Afterwards, the minimum concentrations of anti-CD8 and anti-CD4 mAb were each tested separately for the duration of host CD8 or CD4 depletion in vivo and for the identification of the threshold point where host T cells would begin to reappear. At this critical time point no free mAb would be expected in the circulation, but host depletion would still be complete.

experiments were performed in addition to the mAb titration studies and in addition to the duration of host depletion studies. Five • 106 T I L (more than 95% CD8 +) and 1 X 10 s normal splenocytes (5% to 10% CD8 +) were injected intravenously into a CD8 + T cell-depleted host 48 hours after the depletion had taken place. One day later animals were killed and their spleens were used for phenotypic analysis. Preparation of CD8 + T cell-depleted splenocytes. Mice were injected intraperitoneally with anti-CD8 mAb ascites diluted at 1:150. One day later the mice were killed; their spleens were sterilely dissected and subjected to the procedures described above. Phenotypic analysis of splenocyte samples was performed to verify that complete CD8 + T-cell depletion had taken place.

Cell surface staining and flow cytometric analysis. Splenocytes (1 • 106) from a CD4 + or a CD8 + T celldepleted mouse were incubated with anti-CD4/phycoerythrin and anti-CDS/fluorescein isothiocyanate mAb conjugate (Becton Dickinson Immunocytometry Systems, Mountain View, Calif.) for 30 minutes at 4 ~ C in the dark. Afterwards, cells were pelleted and washed twice in chilled HBSS plus 5% fetal calf serum and 0.1% sodium azide (Sigma Chemical Company). Cells were then fixed in 1% paraformaldehyde (Sigma Chemical Company). Flow cytometric analysis of stained lymphocytes was performed on a Coulter Epics C cytometer (Coulter Electronics, Hialeah, Fla.). Splenocytes from a nondepleted mouse were used as a positive control. Control study for isolated host CD8 depletion. To prove that no free mAb was present 48 hours after host CD8 + T-cell depletion that could interfere with the mainly CD8 + adoptively transferred T I L , control

Adoptive immunotherapy model in combination with in vivo depletion. Mice were injected intravenously with 5 • 105 MCA-105 tumor cells in 1 ml of HBSS ]3 on day 0. T I L were transferred intravenously at 5 • 10 6 cells on day 3 after tumor injection. All T I L cultures used in in vivo experiments had reached the critical number of cells necessary for adoptive transfer. The in vitro cytotoxicity of these cultures ranged from 3% to 55% against autologous tumor and showed no correlation to in vivo efficacy. Recombinant IL-2 was given intraperitoneally twice daily at 30,000 IU/injection from day 3 to day 7. In vivo depletion of host CD4 + or CD8 + T cells was performed on day 1 by intraperitoneal injection of 0.25 ml of either anti-CD4 or antiCD8 mAb. One treatment group received 1 X 108 nor-

328

Burger et al.

Surgery March 7995

BACKGROUND FLUORESCENCE

CD4/CD8

a) 0.2X

i

CD4

I

|9

9

9 5,,'4

t

/.~

: 9 9

"5

0..11.)

G ~

GRICICI4 TLUQR

FIAIOR

CD8

b)

I~O.2XO.6J~ CD4

I

-'"

9 '~.. 'o , .9

GRIDEN FI,UOR

9

CdlUEI~I I ~ U O R

CD8 Fig. 3. Two-parameter flow cytometric histograms of CD4 vs CD8 on splenocytes. CD4 + cells were labeled with anti-CD4/PE mAb conjugate (red fluorescence). CD8 + cells were stained with anti-CD8/fluorescein isothioeyanate mAb conjugate (greenfluorescence), a), Control for complete host depletion of CD8 + cells 24 hours after intraperitoneal injection of anti-CD8 mAb aseites at 1:150. b), Reappearance of CD8 + cells. One X 108 normal splenocytes had been injected together with 5 X 106 TIL 48 hours after complete host CD8 depletion to facilitate detection of CD8 + T cells. Fluorescence-activated cell sorter analysis was performed 24 hours after treatment.

mal uneducated splenocytes intravenously 2 hours after T I L injection and 50 hours after host depletion of C D 8 + T cells9 An additional treatment group received 1 • 108 CD8 + T cell-depleted splenocytes in the same experimental setup9 Control groups received rIL-2 alone or no treatment 9 Animals were killed 21 days after tumor injection. Murine lungs were insufflated with a 15% solution of India ink (Farber-Castell Co., Lewisburg, Tenn.), washed in water, and bleached in Fekete's solution,14 which made the tumor nodules appear white on the blackened surfaces of the lungs. The number of pulmonary metastases in treatment and control groups was enumerated in a blinded fashion 15 and reported as the mean _+ SEM. Metastatic tumor nodules too numerous to count were assigned a value of >200. Statistical analysis. T h e statistical significance between treatment and control groups was determined with the Wilcoxon rank sum test. 16 All p values are two-sided.

RESULTS Anti-CD8 and anti-CD4 mAb titration studies and duration of host depletion studies. Ascites-containing

anti-CD8 mAb was titrated to a dilution of 1:150, the minimum concentration for complete depletion of host C D 8 + T cells (data not shown) 9 The duration of host C D 8 depletion with anti-CD8 mAb ascites at 1:150 was determined by phenotypic analysis of splenocytes over time. This study was repeated at least twice. Fig. 1 shows a representative example of complete depletion of host CD8 + T cells in the first 48 hours after intraperitoneal injection of anti-CD8 mAb. This was the critical point after which host C D 8 + T cells reappeared, indicating the absence of residual free mAb in the circulation. By the third day after depletion about 22% and by the fourth day about 54% of the original host C D 8 + T-cell population had returned 9 After the fourth day there was only a slight increase in the C D 8 + T-cell population through day 8. As expected, anti-CD8 mAb did not cause any nonspecific depletion of C D 4 + T cells (Fig. 1). Similar studies were performed with anti-CD4 mAb containing ascites diluted to the minimum concentration of 1:50, which still showed complete depletion of host C D 4 + T cells (data not shown). Fig. 2 shows a representative example of complete depletion of host C D 4 +

Surgery Volume 177, Number 3

Burger et al.

300

329

250 u9 uJ .r

(/3

200

200 ~ X

Z z

3

150

100

100

'~ w

X Z

50

Z

0 Tumor alone rlL-2 alone* TIL alone n=13 n=13 n=14

TIL+rlL-2 n=lO

CD4 Depl.1" n=13

TREATMENT GROUPS

Fig. 4. CD4 depletion experiments. All animal groups received 5 • 10s MCA-105 tumor cells intravenously on day 0 of our treatment protocol. *This group received only rIL-2 injections from day 3 to day 7 after tumor passage on day 0. tin this group the host was depleted of its CD4 + T-cell subpopulation 24 hours after tumor injection and 48 hours before treatment with TIL -6 rIL-2.

T cells in the first 48 hours after intraperitoneal injection of anti-CD4 mAb. Interestingly, our nondepleted control mouse represented on day 0 showed high percentages of CD4 + and CD8 + T cells (normal range for CD4 + cells, 11% to 22%; for CD8 + cells, 4% to 13%; own data). By the third day after depletion 12% and by the fifth day 33% of the initial host CD4 + T-cell population had returned. A further increase up to 47% of the initial CD4 + T-cell percentage could be detected by day 8. Anti-CD4 mAb did not cause any nonspecifie depletion of CD8 + T cells.

Control study for isolated host CD8 depletion. Depleting the host of its CD8 + T cells with mAb before T I L treatment runs the risk of unwanted depletion of the mainly CD8 + donor TIL. However, depletion studies over time (Figs. 1 and 2) showed that the depleted host T-cell subsets gradually returned after 48 hours, indicating that no residual free mAb remained. In addition, Fig. 3 shows the results of the cell surface phenotyping analysis of a CD8 + T cell-depleted host that had received 5 • 106 T I L + 1 • 108 splenocytes 48 hours after depletion. Splenocytes were added to the T I L because the low number of T I L that had been transferred did not allow detection by fluorescence-activated cell sorter analysis. The presence of 7.8% CD8 + T cells in a CD8-depleted host after T I L and splenocyte treatment is a percentage approaching control levels of a nondepleted murine spleen (1 • 108 splenocytes, own data) as seen in Fig. 1 on day 0. This finding with the above data confirms that host overdepletion with

Z

0 Tumor alone n=l 5

rlL-2 alone n=l 3

TIL+rlL-2 n=14

CD8 Depl.* n=12

TREATMENT GROUPS

Fig. 5. CD8 depletion experiments. All animal groups were injected intravenously with MCA-105 tumor cells on day 0 of our protocol. *In this group the host was depleted of its CD8 + T-cell subsets 24 hours after tumor injection and 48 hours before TIL + rIL-2 treatment.

anti-CD8 mAb and interference with the adoptively transferred T I L population had not occurred. CD4 depletion experiments. We investigated the role of host CD4 depletion during immunotherapy with T I L + rIL-2. In this set of experiments one treatment group was depleted of its host CD4 + T-cell population 24 hours after tumor cell injection and 48 hours before T I L + rIL-2 treatment. This CD4-depleted treatment group showed little (2 + 1 [SEMI) to no pulmonary metastases similar to the regular T I L + rIL-2 treatment group (3 _ 2) (Fig. 4). T I L given without rIL-2 resulted in significant lung disease (121 _+ 22) similar to the rIL-2 alone control group (104 + 26). The tumor alone control group yielded severe disease (more than 200 metastases) on visual inspection of the lungs. The differences between the T I L + rIL-2 treatment group and the CD4-depleted treatment group in comparison with all of the control groups were highly significant (p < 0.001). CD8 depletion experiments. Fig. 5 illustrates the number of lung metastases found in treatment and control groups. The control groups, tumor alone and rIL-2 alone, showed 177 + t 6 and 140 _+ 22 pulmonary metastases, respectively. T I L given in combination with rIL-2 3 days after intravenous tumor cell injection reduced the number of pulmonary metastases significantly (28 _+ 15) in comparison with both of the control groups (p < 0.001). Depletion of host CD8 + T cells 24 hours after tumor cell injection and 48 hours before T I L + rIL-2 treatment led to severe lung disease (158 + 20) similar to the control group. The CD8-depleted treatment group (158 _+ 20) and the regular

330

Burger et al.

Surgery March 1995 DISCUSSION

::'OC

v~

t.9 Z

10(2

IX Z .-j Z Tumor alone n=lZ

rlL-2 alone n=16

TIL+rlL-2 n=9

CD8 De#.* n=12

Normal Splenoc, l" n=13

CD8Splenoc.l"l" n=7

TREATMENT GROUPS

Fig. 6. Splenocyte recruitment experiments. All animal groups were injected intravenously with MCA-105 tumor cells on day 0 of our protocol. *In this group the host was depleted of its CD8 + T-cell subpopulations 24 hours after tumor injection and 48 hours before TIL + rIL-2 therapy. ]-This group represents a CD8-depleted host that received 1 x 108 normal uneducated splenoeytes 2 hours after T I L transfer on protocol day 3. rIL-2 was administered from day 3 to day 7 of our protocol. ~-This group represents a CD8-depleted host that received 1 • 108 CD8-depleted splenocytes in addition to T I L + rIL-2 treatment.

T I L + rIL-2 treatment group (28 _+ 15) differed significantly (p < 0.001). These data suggest that the recruitment of host CD8 + T cells is important for effective tumor eradication in this adoptive immunotherapy model. Splenocyte recruitment experiments. To confirm that recruitment of host CD8 + T cells is required, we added back normal uneducated splenocytes to a C D 8 depleted host 2 hours after T I L + r I L - 2 treatment on day 3 of our protocol. Fig. 6 illustrates that if normal splenocytes are added back shortly after T I L + r I L - 2 therapy, tumor eradication takes place (28 + 12) despite the previous depletion of the host's own C D 8 + T cells. This is in line with our hypothesis that a recruitment of unstimulated CD8 + T cells is initiated by the adoptively transferred T I L in combination with rIL-2. To prove that the recruited cells are in fact C D 8 +, we introduced a treatment group that received CD8depleted splenocytes 2 hours after T I L + r I L - 2 therapy. In this case a recruitment of splenocytes did not take place or was ineffective as indicated by the significant number of pulmonary metastases (153 _+ 30). T h e difference between the CD8-depleted treatment group that received normal splenocytes after T I L + r I L - 2 therapy (28 + 12) and the CD8-depleted treatment group that received CD8-depleted splenocytes after T I L + r I L - 2 treatment (153 _ 30) was significant (p < 0.001).

In this study we have taken the first step in dissecting the mechanism behind adoptive immunotherapy of metastatic cancer with T I L + rIL-2. Although murine 2 and human 17"19 trials have progressed from a therapy with L A K cells to a more potent, tumor-specific therapeutic strategy with T I L , the overall response rate is still only 42% at best in patients with metastatic melanoma)9, 20 Our own L A K cell data have shown that the additional change from high-dose rIL-2 to a low-dose rIL-2 regimen resulted in similar response rates and fewer side effects for the patient. 21,22 We (data not shown) have found tumor regression in about 30% of the patients with metastatic melanoma or renal cell carcinoma who participated in our recent clinical trial of solid-phase anti-CD3 activated T I L in combination with low-dose rIL-2. Understanding the immunologic mechanisms behind the antitumor efficacy of T I L + r l L - 2 treatment would allow optimization of this type of immunotherapy. So far, little proof has been found as to whether T I L induce tumor eradication by direct cytolysis, 2~ by cytokine secretion, 23 or by a combination of both. It is also unknown to what extent the host's immune response is required for the observed tumor regression during therapy with T I L + rIL-2 and how the adoptively transferred cells may interact with the endogenous immune system. In our murine studies we approached these questions by depleting the host of its C D 4 + or C D 8 + T-cell subpopulations with mAb 24 hours after intravenous injection of M C A - 1 0 5 tumor cells and 48 hours before adoptive transfer of T I L in combination with rIL-2. In the M CA-105 pulmonary metastases model depletion of host C D 8 + T cells before T I L + r I L - 2 treatment abrogated all tumoricidal activity (Fig. 4) usually associated with this type of immunotherapy, ll In contrast, depletion of host C D 4 + helper cells did not interfere with the antitumor efficacy of T I L + rIL-2 treatment. However, T I L given without r I L - 2 resulted in a significant decrease of tumoricidal capacity (Fig. 4). By adding back large numbers (1 • 108) of normal uneducated splenocytes to a C D 8 + T cell-depleted host shortly after T I L transfer, we were able to overcome the abrogation of tumoricidal capacity seen in CD8-depleted hosts (Fig. 6). T h e underlying theory is that the added back CD8 + population of splenocytes replaced the host's depleted C D 8 + T cells as a pool for recruitment by the adoptively transferred T I L . To further verify that the cell population recruited from the added back splenocyte pool are phenotypically C D 8 + T cells, we added back CD8-depleted splenocytes to a CD8 + T cell-depleted host shortly after T I L treatment. T h e significant amount of lung metastases found in this treatment group shows that without a recruitment pool of C D 8 + lymphocytes, tumor eradication fails in this treatment protocol. These

Surgery Volume 117, Number 3

findings suggest that the recruitment of host CD8 + T cells by adoptively transferred T I L in combination with rIL-2 seems to be important for effective tumor eradication in this type of cancer immunotherapy. One potential criticism of these studies is the inadvertent depletion of the adoptively transferred T I L . However, the amount and concentration of the antiCD4 or anti-CD8 mAb administered had been chosen such that complete depletion of host T-cell subsets occurred without any free mAb present at the time of T I L transfer (Figs. 1 through 3). With our mAb dilutions, mAb injection 48 hours before T I L treatment was the optimal time point for host depletion of T-cell subsets. The add back experiments with T I L + normal uneducated splenocytes, which yielded significant tumor eradication in a CD8-depleted host, represent further evidence that the adoptively transferred T I L were not affected by the host depletion with anti-CD8 mAb. Because the T I L were injected 2 hours before the splenocytes, the T I L would have been depleted first and therefore become ineffective, provided that free mAb was left over in the host. The tumor eradication seen in this treatment group cannot be explained by therapy with uneducated splenocytes and low-dose rIL-2. Therefore the effectiveness of the T I L in these depleted hosts further verifies the absence of residual free anti-CD8 mAb. In previous experiments we had considered the possibility of giving anti-CD8 mAb intraperitoneally 1 to 2 days before tumor injection. However, all mice that underwent host depletion of CD8 + T cells before tumor injection yielded early massive pulmonary disease compared with the tumor alone control group and therefore would have altered our treatment protocol (data not shown). Other investigators have also identified the CD8 + T cell as the predominant cellular mediator in the response to tumor challenge. 24"26 Mule et al. 2s reported reduced therapeutic efficacy of high-dose rIL-2 treatment in a murine pulmonary metastases model, when host CD8 + T cells had been depleted by intravenous injection of anti-CD8 mAb. Natural killer cells were attributed an ancillary role only. Depletion of host CD4 + T cells had no effect on the rIL-2 induced tumor regression. However, in our studies for the first time, depletion of host T-cell subsets was applied to cancer immunotherapy with T I L + rIL-2, which allowed us to differentiate between the roles of host immune response and adoptively transferred T I L . The fact that CD4 + T cells played a minor role in our immunotherapy protocol might be due to the exogenously administered rIL-2. This theory is also supported by the work of Mule et al. 27 showing that regression of 3-day pulmonary metastases from weakly immunogenic solid tumors by intraperitoneal injection of rIL-6 required the partici-

Burger et al.

331

pation of both host CD4 + and CD8 + T cells, whereas intraperitoneal treatment with high-dose rIL-2 did not require host CD4 + T cells. 25 When T I L were given without rIL-2, we found significant pulmonary disease in comparison with the animal group that had received immunotherapy with T I L + rIL-2 (Fig. 4). These data suggest that host CD4 + T cells by themselves are not triggered to produce high enough levels of IL-2 to make exogenous administration of rIL-2 superfluous. Because MCA-105 is a weakly immunogenic murine sarcoma expressing major histocompatibility complex class I tumor antigen complexes only, these tumor cells are restricted to recognition by cells of the CD8 + phenotype. 28 This fact might further explain why CD4 + T cells, representing the major histocompatibility complex class II-restricted cell type, play such a subordinate role in our tumor model. Our observations show that T I L given at a concentration of 5 • 106 cells seem to require a host immune response, specifically the CD8 + T-cell population, for tumor eradication. Theoretically the intravenous application of a larger number of T I L might be just as tumoricidal without depending on host CD8 + T cells. However, in a murine model the increased size of 3- to 4-week-old, in vitro nonspecifically activated T I L does not permit intravenous injection of these cells in numbers much higher than 5 X 106, because pulmonary and cerebral embolisms would occur. The adoptive transfer of LAK cells29' 30 or short-term cultivated tumor-draining lymph node cells 31 in numbers up to 1 X 108 cells is feasible because of a smaller cell diameter, but their tumoricidal ability is not as efficient as T I L + rIL-2 treatment.l Other investigators have demonstrated significant antitumor activity of T I L + rIL-2 in treatment protocols including whole body sublethal irradiation or cyclophosphamide administration. However, these studies vary from each other in timing and dosage of immunosuppression 1, 3, 9 and are therefore responsible for variability of effect in each treatment modality. As described in our introduction , the mechanism and degree of immunosuppression during adoptive immunotherapy is still the subject of discussion. In our experiments we were able to achieve significant tumor eradication in the T I L + rIL-2 treatment group without additional variables like timing and dosage of immunosuppression. In addition, we performed depletions of host CD4 + or CD8 + T-cell subsets in some groups before T I L + rIL-2 therapy to define their role during this type of immunotherapy. Similar results in T I L + rIL-2 treatment protocols with or without immunosuppression do not exclude the possibility of different mechanisms of tumor eradication. We postulate that T I L given with rIL-2 most likely are not mediating tumor regression primarily via direct

332

Burger et al.

cytolysis but t h r o u g h a n activation a n d integration of the host i m m u n e system. T h e host i m m u n e response to intravenous t u m o r challenge is by itself insufficient to stop the development of p u l m o n a r y metastases. However, with the help of T I L + r I L - 2 t r e a t m e n t we were able to eradicate metastatic p u l m o n a r y disease significantly (p < 0.001). al O u r data suggest a r e c r u i t m e n t of host C D 8 + T cells by the adoptively transferred T I L in c o m b i n a t i o n with r I L - 2 . T h e question of how this rec r u i t m e n t proceeds still remains. G o e d e g e b u u r e et al. 32 have demonstrated that T I L isolated from M C A - 1 0 5 t u m o r release cytokines in vitro such as IL-2, interferon-T, g r a n u l o c y t e - m a c r o p h a g e colony-stimulating factor, a n d low a m o u n t s of t u m o r necrosis f a c t o r - a a n d I L - 6 by using identical culture conditions as described above. T h e r e f o r e we can a s s u m e that cytokine production by in vitro activated and adoptively transferred T I L m a y aid these cells in the r e c r u i t m e n t of host C D 8 + T cells in vivo. W e hypothesize that the cytokines produced by donor T I L initiate a cytokine cascade w i t h i n the host. T h i s involves a nonspecific i n f l a m m a t o r y response initiated by g r a n u l o c y t e - m a c r o p h a g e c o l o n y - s t i m u l a t i n g factor, i n t e r f e r o n - % a n d t u m o r necrosis f a c t o r - a , which activate maerophages a n d granulocytes to produce further cytokines such as I L - l a , IL-1B, a n d IL-6. 23' 24, z7 In r e t u r n , I L - 6 is k n o w n to stimulate cell proliferation a n d cytotoxic function in T lymphoeytes. 33 In addition, the presence of exogenous r I L - 2 together with the I L - 2 produced by donor T I L m a y provide a proliferation s t i m u l u s for resting precursor T cells. 34 W e suggest that as a result of a cytokine cascade u n e d u c a t e d precursor T cells m a y be attracted to the t u m o r site, where a n o n specific i n f l a m m a t o r y response is in progress. E x p o s u r e to cytokines and t u m o r antigen stimulates these cells to become cytotoxic and in t u r n to eradicate t u m o r cells by direct eytolysis. T a k e n together, o u r findings suggest that adoptive i m m u n o t h e r a p y with T I L + r I L - 2 m a y operate by the r e c r u i t m e n t of host C D 8 + T cells for successful t u m o r eradication. T h e m e c h a n i s m b e h i n d this r e c r u i t m e n t m a y be the initiation of a cytokine cascade at the t u m o r site by the adoptively transferred T I L . REFERENCES

1. Rosenberg SA, Spiess P, Lafreniere R. A new approach to the adoptive immunotherapy of cancer with tumor-infiltrating lymphocytes. Science 1986;233:1318-21. 2. Spiess PJ, Yang JC, Rosenberg SA. In vivo antitumor activity of tumor-infiltrating lymphocytes expanded in recombinant interleukin-2. J Natl Cancer Inst 1987;79:1067-75. 3. School DD, Massaro AF, Obando JA, Cusaek JC Jr, Eberlein TJ. The biological effects of immunosuppression on cellular immunotherapy. Surg Oneol 1992;1:27-35. 4. Berendt M J, North RJ. T-cell mediated suppression of antitumor immunity. J Exp Med 1980;151:69-80. 5. Dye ES, North RJ. T-cell mediated immunosuppression as an

Surgery March 1995 obstacle to adoptive immunotherapy of the P815 mastocytoma and its metastases. J Exp Med 1981;154:1033-42. 6. North RJ. Cyclophosphamide-facilitated adoptive immunotherapy of an established tumor depends on elimination of tumor-induced suppressor T cells. J Exp Med 1982;155:1063-74. 7. Bursuker I, North RJ. Generation and decay of the immune response to a progressivefibrosarcoma.J Exp Med 1984;159:131221. 8. Shu S, Rosenberg SA. Adoptive immunotherapy of newly induced murine sarcomas. Cancer Res 1985;45:1657-62. 9. Camerson RB, Spiess PJ, Rosenberg SA. Synergistic antitumor activity of tumor-infiltratinglymphocytes, interleukin 2, and local tumor irradiation. J Exp Med 1990;171:249-63. 10. Malkovsky M, Jira M, Madar J, Malkovska V, Loveland B, Asherson GL. Generation of lymphokine-activated killer cells does not require DNA synthesis. Immunology 1987;60:471-3. 11. Massaro AF, Schoof DD, Rubinstein A, Zuber M, LeonardVidal FJ, Eberlein TJ. Solid-phase anti-CD3 antibody activation of murine tumor-infiltrating lymphocytes. Cancer Res 1990;50:2587-92. 12. Parker GA, Rosenberg SA. Serologic identification of multiple tumor-associated antigens on murine sarcomas. J Natl Cancer Inst 1977;58:1303-9. 13. Mule J J, Shu S, Rosenberg SA. The anti-tumor efficacy of lymphokine-activated killer cells and recombinant interleukin-2 in vivo. J Immunol 1985;135:646-52. 14. Fekete E. A comparative morphological study of the mammary gland in a high and low tumor strain of mice. Am J Pathol 1938;14:557-78. 15. Wexler H. Accurate identification of experimental pulmonary metastases. J Natl Cancer Inst 1966;36:641-5. 16. Gehan EA. A generalized Wileoxon test for comparing arbitrarily singly-censored samples. Biometrica 1965;52:203-17. 17. Topalian SL, Muul LM, Solomon D, Rosenberg SA. Expansion of human tumor infiltrating lymphocytes for use in immunotherapy trials. J Immunol Methods 1987;102:127-41. 18. 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 1988;6:839-53. 19. Rosenberg SA, Packard BS, Aebersold PM, et al. Use of tumorinfiltrating lymphocytes and interleukin-2 in the immunotherapy of patients with metastatic melanoma. N Engl J Med 1988;319:1676-80. 20. Aebersold P, Hyat C, Johnson S, et al. Lysis of autologous melanoma cells by tumor-infiltratinglymphocytes: association with clinical response. J Natl Cancer Inst 1991;83:932-7. 21. Eberlein T J, School DD, Jung SE, et al. A new regimen of interlenkin 2 and lymphokine-activated killer cells. Arch Intern Med 1988;148:2571-6. 22. Eberlein T J, Rodrick ML, Massaro AF, Jung SE, Mannick JA, Schoof DD. Immunomodulatory effects of systemic lowdose recombinant interleukin-2and lymphokine-activated killer cells in humans. Cancer Immunol Immunother 1989;30:145-50. 23. Barth RJ Jr, Mule J J, Spiess PJ, Rosenberg SA. Interferon 3' and tumor necrosis factor have a role in tumor regression mediated by murine CD8 + tumor-infiltrating lymphoeytes. J Exp Med 1991;173:647-58. 24. Stoppacciaro A, Melani C, Parenza M, et al. Regression of an established tumor genetically modified to release granulocyte colony-stimulating factor requires granulocyte-T cell cooperation and T eell-produced interferon % J Exp Med 1993;178:15161. 25. Mule J J, Yang JC, Lafreniere R, Shu S, Rosenberg SA. Identification of cellular mechanisms operational in vivo during the

Surgery Volume 17 7, Number 3 regression of established pulmonary metastases by the systemic administration of high-dose recombinant interleukin 2. J Immunol 1987;139:285-94. 26. Tuttle TM, Inge TH, MeCrady CM, Bethke KP, Bear HD. Activation of CD8 + murine T cells from tumor-draining lymph nodes by phorbol dibutyrate plus calcium ionophore. J Immunother 1992;12:32-40. 27. Mule J J, Custer MC, Travis WD, Rosenberg SA. Cellular mechanisms of the antitumor activity of recombinant IL-6 in mice. J Immunol 1992;148:2622-9. 28. Restifo NP, Esquivel F, Asher AL, et al. Defective presentation of endogenous antigens by a murine sarcoma. J Immunol 1991;147:1453-9. 29. Eberlein T J, Rosenstein M, Rosenberg SA. Regression of a disseminated syngeneie solid tumor by systemic transfer of lymphoid cells expanded in interleukin-2. J Exp Med 1982; 156:38597.

Burger et al.

333

30. Lafreniere R, Rosenberg SA. Successful immunotherapy of murine experimental hepatic metastases with lymphokine-actirated killer cells and recombinant interleukin-2. Cancer Res 1985;45:3735-41. 31. Geiger JD, Wagner PD, Cameron M J, Shu S, Chang AE. Generation of T cells reactive to poorly immunogenic B 16-BL6 melanoma with efficacy in the treatment of spontaneous metastases. J Immunother 1993;13:153-65. 32. Goedegebuure PS, Zuber M, Leonard-Vidal DL, et al. (Re) Activation of murine tumor-infiltrating lymphocytes with solidphase anti-CD3 antibody: in vitro cytokine production is associated with in vivo efficacy. Surg Oneol 1994;3:79-89. 33. Kishimoto T. The biology of interleukin-6. Blood 1989;74:1-10. 34. Bich-Thuy LT, Dukovich M, Pfeffer N J, Fauci AS, Kehrl JH, Greene WC. Direct activation of human resting T cells by IL-2: the role of an IL-2 receptor distinct from the Tac protein. J Immunol 1987;139:1550-6.

BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of SURGERY are available to subscribers (only) for the 1995 issues from the publisher at a cost of $74.00 for domestic, $100.58 for Canadian, and $94.00 for international subscribers for Vol. 117 (January to June) and Vol. 118 (July to December). Shipping charges are included. Each bound volume contains a subject index and an author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular journal subscription. Payment must accompany all orders. Contact Mosby-Year Book, Inc., Subscription Services, 11830 Westline Industrial Dr., St. Louis, M O 63146-3318; phone (800)3254177, ext. 4351, or (314)453-4351.