- Email: [email protected]
Mechanisms of cellular cytotoxicity mediated by a recombinant antibody-IL2 fusion protein against human melanoma cells
Immunology Letters, 39 (1994) 91-99 © 1994 Elsevier Science B.V. All rights reserved IMLET 02067
Mechanisms of cellular cytotoxicity mediated by a recombinant antibody-IL2 fusion protein against human melanoma cells M a y u m i N a r a m u r a a, Stephen D. Gillies b, J o h n M e n d e l s o h n c, R a l p h A. Reisfeld a a n d B a r b a r a M. Mueller a'* ~The Scripps Research Institute, La Jolla, CA 92037, USA; bFuji ImrnunoPharmaceuticals Corporation, Lexington, MA 02173, USA; CMemorial Sloan-Kettering Cancer Center, New York, N Y 10021, USA (Received 1 July 1993; accepted 12 November 1993)
1.
Summary
Functional characteristics were established for a genetically engineered fusion protein between human IL2 and mouse/human chimeric mAb 225 directed against human epidermal growth factor receptor (EGFR), aberrantly expressed on human melanoma cells. The emphasis of these studies was on the mechanism(s) of action by which the ch225-IL2 fusion protein mediated cytotoxic killing of human melanoma cells by different human immune effector cells. Ch225-IL2 fusion protein bound to human EGFR with the high affinity of the parental antibody, and was as active as the equivalent amount of rhlL2. Ch225-IL2 enhanced cellular cytotoxicity mediated by freshly separated PBMC, isolated natural killer (NK) cells and activated T cells against melanoma cell lines. NK cells, which constitutively express both Fc~RIII and IL2R, interacted with ch225-IL2, mainly through FcTRIII, while the involvement of IL2R was secondary. However, the effect of ch225-IL2 on activated T cells was most likely mediated through IL2R. These results sug-
Key words: Recombinant antibody-IL2 fusion protein; Cellular cytotoxicity; Fc),RIII; IL2R; Melanoma; Immunotherapy "Corresponding author: Dr. Barbara M. Mueller, The Scripps Research Institute, Dept. of Immunology, IMM13, 10666 N. Torrey Pines Road, La Jolla, CA 92037, USA. Tel.: (619) 5548124; Fax: (619) 554-6705.
gest that the genetically engineered ch225-IL2 fusion protein may become a potent immunotherapeutic agent capable of stimulating various immune effector populations to effectively kill human melanoma cells. 2.
Introduction
Interleukin-2 (IL2) is a potent immunostimulator for a broad range of cell types, including both T and B cells, natural killer (NK) cells, monocytes and macrophages [1]. The use of IL2 led to some success in cancer immunotherapy [2], but systemic administration of high-dose IL2, initially employed in the clinic, was frequently accompanied by severe toxicity. However, doses of IL2 lower than those given originally showed some efficacy in modifying the host immune response without systemic toxicity [1]. An alternative to the systemic administration of IL2 may be the use of recombinant anti-tumor antibody-IL2 fusion proteins. Such constructs are designed to specifically target and concentrate IL2 at tumor sites while minimizing systemic toxicities and optimizing interactions between tumor cells and IL2-activated immune effector cells [3,4]. Mechanisms by which immune effector cells destroy tumor cells can be divided into two categories based on the involvement of receptors for IgG, i.e., Fc7 receptors (FcTRs). Thus, killing of tumor cells that occurs independent from Fc~Rs may involve cytotoxic T lymphocytes (CTLs), 91
NK cells and lymphokine-activated killer (LAK) cells. In contrast, monoclonal antibodies (mAbs) directed against tumor antigens effectively mediate antibody-dependent cellular cytotoxicity (ADCC) by immune effector cells bearing FcyRs such as NK cells, monocytes and macrophages. The importance of IL2 in both mechanisms has been well demonstrated, since this cytokine is not only indispensable for the generation of CTLs and LAK cells but also is capable of enhancing ADCC activities against tumor cells [5,6]. We previously reported the production of a recombinant anti-ganglioside GD2 antibody IL2 fusion protein (chl4.18-IL2) with full IL2 and antigen binding activity, capable of enhancing in vitro killing of autologous GD2-positive human melanoma cells by a tumor-infiltrating lymphocyte (TIL) cell line [4]. However, the mechanisms of action by which antibody-IL2 fusion proteins potentiate melanoma cell killing and the range of immune effector cell populations that can be activated remain to be clarified. Here we report on the functional characteristics of a recombinant fusion protein of human IL2 and a chimeric mouse/human monoclonal antibody (ch225) against the epidermal growth factor receptor (EGFR), aberrantly expressed on amelanotic human melanomas. The emphasis of these studies is on the mechanism of action by which this fusion protein (ch225-IL2) potentiates the killing of human melanoma cells by different immune effector cell populations. 3.
3.1.
Materials and Methods
Cytokines, antibodies and antibody-cytokine fusion protein
Recombinant human IL2 (rhlL2) derived from Escherichia coli was purchased from R&D Systems (Minneapolis, MN). Mouse/human chimeric anti-human EGFR mAb 225 (ch225) and its parental mouse anti-human EGFR mAb 225 (m225) were described previously [7,8]. The ch225-IL2 fusion protein was constructed as already reported for the ch14.18-IL2 fusion protein [4] by fusing a synthetic human IL2 sequence to the carboxyl end of the human C~I gene. Anti-human FcTRIII mAb 3G8 [9], anti-human IL2R p55 ct92
chain mAb anti-Tac [10], and anti-human IL2R p75 fl-chain mAb Mik-fll [11] were generously provided by Drs. J.C. Unkeless (Mount Sinai School of Medicine, New York, NY), R.P. Junghans (New England Deaconess Hospital, Boston, MA), and T.A. Waldmann (NIH, Bethesda, MD), respectively. Leu series mAbs were purchased from Becton Dickinson (San Jose, CA) and anti-CD3 mAb OKT3 was prepared from hybridoma culture supernatant in our laboratory.
3.2.
Tumor cell lines
Both M24met and C8161 are non-pigmented human melanoma cell lines that spontaneously metastasize in immunodeficient laboratory animals [12,13]. M24met was maintained in RPMI1640 medium supplemented with 10% fetal calf serum (FCS) and 2 mM L-glutamine. C8161 was kindly provided by Dr. M.J.C. Hendrix (University of Arizona, Tucson, AZ) and was maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FCS and 2 mM L-glutamine. Both cell lines were treated with 0.5 mM EDTA to detach from tissue culture flasks and were used as single cell suspensions.
3.3.
Indirect immunofluorescence
Melanoma cells were incubated with 1 /~g of either m225, ch225 or ch225-IL2. After removal of excess primary antibody by washing, cells were incubated with either goat anti-mouse IgG-FITC (for m225) or with goat anti-human IgG-FITC (for ch225 and ch225-IL2). Stained cells were analyzed in a FACScan flow cytometer (Becton Dickinson).
3.4.
Direct binding assays with radioiodinated proteins
Ch225 and ch225-IL2 were radiolabeled with Na125I (Amersham, Arlington Heights, IL) by the IODO-BEADS method according to the manufacturer's instructions (Pierce, Rockford, IL). Binding assays were performed and the data were analyzed, as described previously [7].
3.5.
IL2 assay
IL2 activity of the ch225-IL2 fusion protein was determined with a standard T-cell proliferation assay by measuring the [3H]thymidine incorporation of the IL2-dependent mouse cell line CTLL-2 [14].
3.6.
Effector cells
The source of effector cells was heparinized venous blood from healthy volunteer donors. Each of the individual experiments was performed with effector cells from a single donor. However, since we did not have access to the same donor repeatedly, the donor varied from one experiment to the other. Peripheral blood mononuclear cells (PBMC), isolated by centrifuging the whole blood through Ficoll-Paque (Pharmacia LKB, Piscataway, NJ), were suspended at 2 x 106/ml in RPMI-1640 and incubated in T150 tissue culture flasks (Corning, Corning, NY) at 37°C for 1 h. This procedure removes the majority of monocytes. Unstimulated T cells were prepared from the nonadherent population by negative immunomagnetic selection. Cells were coated with mAbs directed against CD14 (Leu-M3), CD19 (Leu-12) and CD56 (Leu-19), and incubated with beads coated with goat anti-mouse IgG (Dynabeads; Dynal, Great Neck, NY). Monocytes, N K cells and B cells were removed in a magnetic field. The resulting cell population was typically > 9 0 % CD3 +, < 3 % CD 16 +, when analyzed in a FACScan flow cytometer. Activated T cells were obtained by stimulating T cells with anti-CD3 mAb and rhlL2. To this end, T cells were positively selected from plastic-nonadherent cells in flasks coated with anti-CD3 mAb (MicroCELLector; AIS, Santa Clara, CA), followed by incubation for 3-5 days in RPMI-1640/10% FCS with 25 ng/ml rhlL2. The cells ( > 9 5 % CD3+, > 8 0 % CD25 ÷) were further separated into CD4 + and CD8 + populations in anti-CD4 and anti-CD8 coated flasks, respectively. To isolate N K cells, T cells, B cells and monocytes, PBMCs were depleted by adhesion to anti-CD3 mAb-coated flasks and negative immunomagnetic selection with anti-CD3 (OKT3), anti-CD14 (Leu-M3) and anti-CD19 (Leu-12) mAbs. The resulting popula-
tion contained > 7 0 % CD16*, > 8 0 % CD56" cells.
3.7.
Cytotoxicity assays
Cytotoxic activity of various effector cell populations was measured in 4-h SlCr-release assays against melanoma cell lines, as described previously [7]. All the effector cells were tested in triplicate at target to effector (T/E) ratios of 1:100, 1:50, 1:25, and 1:12.5, except for N K cells, where T/E ratios were 1:40, 1:20, 1:10, and 1:5, respectively. In some cases, cytotoxicity assays were done in the presence of ~2-antiplasmin. To investigate the involvement of FcTRIII and IL2R, effector cells were first incubated with either antiFcTRIII mAb 3G8 or anti-IL2R mAbs anti-Tac and Mik-fll at 37°C for 30 rain, prior to adding tumor cells and anti-EGFR mAbs. Statistical analysis of the data was performed by Student's t test. A probability of less than 5% (P<0.05) for no difference between the experimental group and the control group was considered significant [15]. 4.
4.1.
Results
Antigen-binding activity of ch225-IL2 fusion protein
Both human melanoma cell lines, M24met and C8161, were shown to express a comparable number of E G F R molecules on the surface by indirect flow cytometry (Fig. 1). The parental murine mAb and its genetically engineered derivatives all showed essentially identical binding patterns. The dissociation constants (KDS) for 125I-labeled ch225 and ch225-IL2 were calculated from saturation binding curves obtained with the M24met cell line. The average KDS in three independent assays were 2.62 x 10 -1° M for ch225 and 3.31 x 10 -1° M for ch225-IL2, indicating that fusing the cytokine IL2 to this chimeric antibody did not change its binding characteristics.
4.2.
IL-2 activity
The IL-2 activity of ch225-IL2 was measured as the proliferation of the IL2-dependent cell line CTLL-2 and compared against that of rhlL2. 93
a)
b) 80
80 M24met =.60-
c.~5~,~ 1
E ;~ 400
C8161
ch225..IL2
60 . Go~HIgG
40
ch225-1L2
J~
m225
20
200
/
,
m~5
i
J
i lll.~
i
i
101
10 o
i
ill.'~
i
102
i
i
Illll
t
103
0
i
J
i
ill.
100
I
101
i
i
i
ill.
I
10 ~
I
I
I
ll.I
103
Fluorescence Intensity Fig. 1. Flow cytometric analysis of human melanoma cell lines M24met (a) and C8161 (b) incubated with either murine anti-human EGFR mAb m225 and goat anti-mouse [gG-FITC; mouse/human chimeric mAb ch225 and goat anti-human IgG-FITC; or ch225 IL2 fusion protein and goat anti-human IgG-FITC. Cells incubated only with goat anti-human IgG-FITC are shown as control.
Cells were incubated with serial dilutions of rhlL2, ch225 or ch225-IL2 covering the range of 0.01-10000 pM. The concentration of IL2 in the ch225-IL2 fusion protein was calculated as two molar equivalents of IL2 per mole of fusion protein. As shown in Fig. 2, IL2 activity of ch225IL2 was essentially identical to that of a comparable amount of rhlL2. The anti-human E G F R antibody ch225 alone had no effect on CTLL-2 proliferation.
30.
i10. t.
7,
¢, z
0
.01
t
1
10
IL2 (pM)
100
1000
10000
Fig. 2. IL2 activity of ch225-IL2 fusion protein (O) as measured by [3H]thymidine incorporation of the CTLL-2 cell line. Recombinant IL2, expressed in E. coli, was used as control (C)). Results are normalized to the IL2 concentration. Bars denote the standard deviation.
94
4.3.
Cytotoxicity mediated by fresh PBMC and NK cells
We have previously reported that ch225 enhances cytotoxic activity of human PBMC against a melanoma cell line bearing EGFR, but that it does not mediate complement-dependent cytolysis [7]. Thus, in the first series of experiments, the ability of ch225-IL2 to potentiate cellular cytotoxicity was compared to that of ch225. PBMC freshly separated from the blood of healthy volunteer donors were better able to mediate cytotoxicity against M24met and C8161 cells with ch225-IL2 than with ch225 (Fig. 3). The carboxyl terminal lysine residue on the chl4.18-IL2 fusion protein heavy chain is accessible to cleavage by plasmin [4]. Since FCS contains small amounts of plasminogen and tumor cells are capable of producing plasminogen activators [16], it was necessary to determine whether the intact fusion protein molecule is capable of stimulating effector cells or whether it needs to be cleaved into its ch225 and IL2 components to exhibit biological activities. To clarify this point, cytotoxicity assays were done in the presence or absence of 50 /~g/ml ~2-antiplasmin, a potent inhibitor of plasmin. Addition of ~2-antiplasmin did not change the ability of ch225-IL2 to enhance cellular cytotoxicity mediated by PBMC, indicating that the intact fusion protein is fully capable to interact with effector cells (data not shown).
(a)
(b) 6O
70
5O
60
-lOd
"r
Control
r
eh 995
"r
a
0
~ 225-11.2
,
,
,
Control
e~ 225
eh ~5-11.2
Fig. 3. Enhancement of cellular cytotoxicity mediated by P B M C against h u m a n melanoma cells M24met (a) and C8161 (b) coated with 1 /~g/ml ch225 or 1 /~g/ml ch225-IL2. Cytotoxicity was measured in a 4-h SlCr release assay at target-to-effector cell ratios of 1:100 ([]), 1:50 (I--1), 1:25 (11) and 1:12.5 ([]). Standard deviations are indicated. The probability of no difference between controls and ch225 or ch225-IL2 coated tumor cells, respectively, as determined by Student's t test was P < 0.005.
To further investigate the respective mechanisms of action of ch225-IL2 and ch225, we tried to identify the effector population(s) that interact with the ch225-IL2 fusion protein. Since preliminary experiments indicated that purified granulocytes did not have any cytotoxic activity against our melanoma cell lines, we mainly focused on lymphocytes. In this regard, the data in Table 1 show that ch225 and ch225-IL2 can mediate NK cells cytotoxicity, while resting T cells are not stimulated by either chimeric antibody or the fusion protein. Unlike fresh PBMC, purified NK cells coated with ch225 or ch225-IL2 killed melanoma target cells with the same efficacy.
4.4.
Involvement of FcTRIII and 1L2R in NKmediated cytotoxicity
NK cells are unique in that they express solely FcvRIII and not the other two types of Fc),Rs [17], and that they express IL2R constitutively [18]. The majority of IL2Rs on the surface of NK cells are of intermediate affinity, being mostly composed of p75 r-chain and p70 y-chain [1]. The number of p55 a-chains is quite limited in NK cells. We hypothesized that cytotoxicity by IL2-antibody fusion proteins is mediated through Fc~,R and/or IL2R. To this end, we dissected the interaction of NK cells with ch225 and
TABLE 1 C Y T O T O X I C I T Y O F R E S T I N G T A N D N K CELLS M E D I A T E D BY T H E ch225-IL2 F U S I O N P R O T E I N A G A I N S T H U M A N M E L A N O M A CELLS Percent o f specific lysis of 51Cr-labeled target ceils in 4-h assays at target-to-effector cell ratios of 1:100 (T cells) and 1:40 ( N K ceils), expressed as average + standard deviation (n = 3). Target cells
M24met
C8161
Effector cells
T cells
N K cells
T cells
N K cells
Control ch225 (1 /~g/ml) ch225-1L2 (1 /tg/ml)
2.98 + 2.93* 3.98 + 1.46 --0.48 ___ 2.62
2.02 ___ 4.88 16.38 + 2.79 13.70 _ 1.26
0.79 ___ 0.89 4.88 + 0.64 7.23 ___ 2.20
2 1 . 8 0 _ 0.78 43.22 _ 4.47 37.71 + 1.51
95
TABLE 2 E F F E C T OF ANT1-Fc~RIII and ANTI-IL2R ANTIBODIES O N N K CELL C Y T O T O X I C I T Y M E D I A T E D BY THE ch225-IL2 FUSION PROTEIN N K cells were coated with either 20/ag/ml 3G8 anti-FcyRIII mAb or 10/~g/ml each of anti-Tac and Mik-fll anti-IL2R mAbs. Figures in the table denote the average percent specific lysis _ standard deviation determined in triplicate in 4-h 51Cr release assays at targetto-effector cell ratios of 1:40. M24met NK
C8161 N K + 3G8
N K + ~tlL2R
NK
N K + 3G8
1.48 _ 0.69 - 0 . 4 4 ___ 1.65 7.85 + 2.031
6.06 + 1.59 n.d." 32.76 + 3.512
3.43 + 0.13 38.40 ___ 2.57 34.25 _ 1.56
1.78 _ 1.71 2.15 + 1.57 5.36 + 1.672 n.d. n.d.
0.87 ___ 0.94 n.d. 8.37 ___ 0.892 -0.11 _ 2.86 n.d.
NK+~IL2R
Experiment 1 Control ch225 (1 #g/ml) ch225-IL2 (1 #g/ml)
7.67 + 1.26 46.10 _ 1.54 45.72 _+ 2.01
- 3 . 1 0 ___ 0.38 - 1 . 9 8 + 1.27 4.24 + 1.421
7.46 ___ 1.55 n.d. 30.99 + 2.713
1 1 . 8 3 _+ 1.50 14.24 _ 1.18 1 5 . 4 5 _ 0.791 n.d. n.d.
18.21 _ 1.86 n.d. 32.72 + 3.002 18.84 + 2.54 n.d.
Experiment 2 Control ch225 (0.1 #g/ml) ch225-IL2 (0.1 #g/ml) rhlL2 (20 ng/ml) ch225 (0.1 /zg/ml) + rhlL2 (20 ng/ml)
-1.83 19.28 16.90 0.70 19.61
___ 1.91 _+ 2.38 _ 1.29 _ 2.80 + 0.00
17.63 _ 39.68 + 48.95 + 33.20 + 47.77 _
0.77 1.40 1.55 0.43 1.81
Probability of no difference to cytotoxicity mediated by N K ceils and ch225-IL2 as determined by Student's t test: 1p<0.005; 2 P < 0.05; 3P> 0.05. "n.d.: not determined.
ch225-IL2 by blocking either FcTRIII or IL2R with their specific antibodies. When N K cells were coated with 20 /~g/ml anti-human FcTRIII mAb 3G8, neither 1 #g/ml ch225 nor ch225-IL2 could enhance N K cytotoxicity against melanoma cells (Table 2, Exp. 1). However, blocking IL2R with 10 #g/ml each of anti-Tac anti-human p55 IL2R ~-chain and Mik-fll anti-human p75 IL2R fl-chain mAbs resulted in only a slight reduction in the activity of ch225-IL2. The concentration of IL2 in 1 #g/ml ch225-IL2 is approximately 160 ng/ml (11 nM), which may not have been blocked completely by 10 #g/ml each of anti-IL2R mAbs [18]. Therefore, in the next set of experiments, 10-fold smaller concentrations of ch225 and ch225-IL2 were tested. In this case, both ch225 and ch225-IL2 were effective in potentiating N K cells at 0.1 #g/ml. However, while anti-FcyRIII m A b almost completely abrogated the potentiating effect of ch225 and ch225-IL2, anti-IL2R mAbs were less effective in reducing the activity of ch225-IL2 (Table 2, Exp. 2). The 96
blocking of IL2 b~nding to its receptor by 10/~g/ ml each of anti-IL2R mAbs was confirmed in a control experiment where the effect of an equivalent amount of rhlL2 (20 ng/ml) was abolished. To demonstrate that this partial reduction in cytotoxic activity of the fusion protein is due to the blocking of the IL2R and not the presence of a large amount of an irrelevant Ab, N K cells were coated with the same amount of isotype-matched (IgG2a) murine mAb OKT3. This maneuver did not affect the ability of either ch225 or ch225IL2 to mediate cellular cytotoxicity (data not shown). Therefore, we concluded that N K cells interact with ch225-IL2 mainly through FcyRIII and that the involvement of IL2R is secondary. The modification of the Fc region of ch225 in the ch225-IL2 fusion protein did not block its binding to the Fc~RIII. 4.5.
Cytotoxicity mediated by activated T cells
It has already been reported that recombinant
TABLE 3 C Y T O T O X I C I T Y O F A C T I V A T E D T CELLS M E D I A T E D BY T H E ch225-IL2 F U S I O N P R O T E I N A G A I N S T H U M A N M E L A N O M A CELLS H u m a n T cells activated by anti-CD3 antibodies and rhIL2 were separated into CD4 + and CD8 + subpopulations. Cytotoxicity was evaluated in 4-h 51Cr release assays. Figures in the table denote average percent of specific lysis ___ standard deviation obtained in triplicate assays at a target-to-effector cell ratio of h l00. M 24met
C8161
CD4
CD8
CD4
CD8
4.45 + 0.64 5.55 _+ 0.20* 24.22 __ 1.40"*
8.07 _+ 0.52 8.46 + 0.29" 19.90 _+ 0.83**
17.29 _+ 0.32 23.37 ___ 1.28" 42.00 _+ 1.37""
Experiment 1 Control ch225 (1 /~g/ml) ch225-IL2 (1 ~g/ml)
--0.60 _+ 0.47 7.86 + 0.88* 12.55 ___ 0.51"* CD8
CD8
Experiment 2 Control ch225 (1 #g/ml) ch225-IL2 (1 /~g/ml) rhlL2 (200 ng/ml) ch225 (1 #g/ml) + rhlL2 (200 ng/ml)
35.87 35.92 51.85 52.01 43.90
+ 2.88 _+ 1.35" + 2.66** + 2.15"* + 0.67 *°
1.37 3.47 37.35 23.46 26.87
+ _+ + + +
3.11 3.11" 2.45"* 1.99"* 2.10"*
17.50 18.49 18.82 19.26 20.63
_+ 0.69 + 1.01 ° ___ 0.02" + 1.95" + 1.00""
41.47 33.79 34.27 36.06 37.48
_+ 4.89 + 1.12" ___ 3.79* ___ 2.64* + 0.58*
Experiment 3 Control ch225 (0.1 #g/ml) ch225-IL2 (0.1 /~g/ml) rhlL2 (20 ng/ml) ch225 (0.1 #g/ml) + rhlL2 (20 ng/ml)
Probability of no difference to untreated controls by Student's t test: "P>0.05; **P<0.05.
m A b - I L 2 fusion proteins enhance melanoma cell killing by activated T cells from peripheral blood [3] or by a TIL cell line [4]. We tested the cytotoxicity of activated T cells against two melanoma cell lines, M24met and C8161, established from different patients, by activating T cells nonspecifically through CD3 and by stimulation with exogenous rhIL2. In this case, the activated T cells themselves were capable of killing melanoma cells, and this effect was further enhanced in the presence of ch225-IL2. Since we found that CD8 ÷ cell populations are much more potent effectors than CD4 + T cells (Table 3, Exp. 1), we tested only activated CD8 + cells in the experiments described in Table 3. An equivalent amount of rhIL2 was also found capable of stimulating activated T cells to lyse melanoma targets (Table 3,
Exp. 2), implying that the effect of ch225-IL2 on activated T cells was most likely mediated through IL2R. This is in contrast to observations made with N K cells where 0.1 #g/ml ch225-IL2 was sufficiently effective to stimulate these effector cells, whereas equivalent amounts of ch225IL2 or rhlL2 did not potentiate activated T cells (Table 3, Exp. 3). 5.
Discussion
We report here on the functional characteristics of a fusion protein between mouse/human chimeric anti-EGFR mAb and human IL2, ch225IL2, with emphasis on the mechanism of action by which different immune effector cells potentiate cytotoxic killing of human melanoma cells. 97
This fusion protein retains the high antigen binding affinity of the parental mAb ch225 and is biologically as active as an equivalent amount of rhlL2 in supporting proliferation of an IL2-de'pendent cell line. Furthermore, ch225-IL2 is capable of stimulating immune effector cells to mediate cellular cytotoxicity against M24met and C8161 melanoma cell lines, that express comparable numbers of E G F R s on their surface. These cell lines were established from melanoma metastases of two different individuals and vary in their susceptibility to cellular cytotoxicity. In 4-h cytotoxicity assays, ch225-IL2 interacted with effector cells as an intact molecule, and does not need to be enzymatically cleaved into its ch225 and IL2 components to be active. Fell et al. [3] previously reported the production of a fusion protein, L6/IL2, consisting of a mouse/human chimeric F(ab') with specificity for carcinomas and human IL2. This fusion protein potentiated the cytotoxicity of CD3-activated blast cells, but not that of resting PBL. Since the major difference between L6/IL2 and our ch225IL2 fusion protein is the presence of the immunoglobulin Fc portion in the latter, we postulated that the effect of ch225-IL2 on fresh PBMC is mediated through FcTRs that are expressed on a variety of leukocytes, including monocytes, macrophages, neutrophils, N K cells, as well as B and T cells. When testing the capacity of neutrophils, N K and T cells from peripheral blood to mediate ADCC, we found N K cells more effective in mediating cytotoxicity against melanoma cells coated with ch225-IL2 than T cells. In view of the fact that fresh PBMC mediated more potent cytotoxicity when stimulated with ch225-IL2 than with ch225, and that ch225-IL2 could potentiate the purified N K cells only equally as well as ch225, the contribution of other populations in PBMCs, such as monocytes/macrophages, will have to be further investigated. We were able to show in this report that ch225-IL2 interacts with N K cells mainly through Fc~RIII and that the involvement of IL2R is secondary. It is to be noted that binding of ch225-IL2 to Fc),RIII is retained regardless of addition of IL2 molecules to the Fc portion of this fusion protein. This characteristic should allow ch225-IL2 to stimulate a wide variety of im98
mune effector cells bearing FcTRs and thus hopefully translate into a strong anti-tumor effect, possibly better than that mediated by F(ab') fusion proteins [3]. We also tested the capacity of activated T cells as effector cells in order to compare the efficacy of ch225-IL2 with other antibody-IL2 fusion proteins previously reported. Although resting T cells are not capable of mediating cytotoxicity with ch225 or ch225-IL2, activated CD8 ÷ cells can be further potentiated by ch225-IL2. The discrepancy in cytotoxic activity between ch225 and ch225-IL2 implies that in this case, the effect of ch225-IL2 is not mediated through FcTRIII. The fact that excess amounts of heat-inactivated human serum (1 mg/ml human IgG) could not abrogate the effect of ch225-IL2 on activated T cells, also supports the contention that this phenomenon is independent of FcTR (data not shown). We also tried to prove that ch225-IL2 potentiates activated T cells through IL2R. However, in our hands, anti-IL2R mAbs did not abrogate the effect of ch225-IL2. It is possible that the inability to block the cytotoxic activity of ch225-IL2 with anti-IL2R mAbs does not necessarily contradict our hypothesis, because 1 /~g/ml of ch225-IL2 fusion protein contains the equivalent amount of about 150 to 200 ng/ml IL2 (10-15 nM), and anti-Tac is known to be unable to inhibit the binding of such a large amount of IL2 to the receptor [19]. For this reason, we repeated the same experiment with a 10-fold lower dose of ch225 IL2; however, neither 0.1 #g/ml of ch225-IL2, which was effective to abrogate the effect of N K cells, nor 20 ng/ml rhlL2 were sufficient to enhance the killing of human melanoma cells by activated CD8 + cells. This requirement for relatively high concentrations of IL2 for the cytotoxicity of activated T cells is in agreement with the observation of Fell et al. [3]. One of the rationales behind fusing IL2 with an antibody against a tumor-associated antigen is to deliver relatively larger amounts of this immunostimulator directly to tumor sites and with hopefully fewer of the toxicities observed with large amounts of systemically applied rhlL2. In this regard, preliminary experiments with radiolabeled ch225-IL2 indicated that ch225-IL2 can effectively target to tumor xenografts in scid mice.
These data suggest that c h 2 2 5 - I L 2 m a y be a potent i m m u n o t h e r a p e u t i c agent that c a n recruit effector cells expressing FcvRs a n d that also m i g h t be capable o f delivering relatively large a m o u n t s o f IL2 to t u m o r sites, to locally activate T I L s / CTLs.
Acknowledgements This w o r k was s u p p o r t e d by grants from the N a t i o n a l C a n c e r Institute 1R35CA42508 a n d 1U01CA51946. This is The Scripps Research Institute's m a n u s c r i p t n u m b e r 8 0 1 7 - I M M .
References [1] Smith, K.A. (1993) Blood 81, 1414. [2] Maas, R.A., Dullens, H.F.J. and Den Otter, W. (1993) Cancer Immunol. Immunother. 36, 141. [3] Fell, H.P., Gayle, M.A., Grosmaire, L. and Ledbetter, J.A. (1991) J. Immunol. 146, 2446. [4] Gillies, S.D., Reilly, E.B., Lo, K.-M. and Reisfeld, R.A. (1992) Proc. Natl. Acad. Sci. USA 89, 1428. [5] Honsik, C.J., Jung, G. and Reisfeld, R.A. (1986) Proc. Natl. Acad. Sci. USA 83, 7893. [6] Hank, J.A., Robinson, R.R., Surfus, J., Mueller, B.M., Reisfeld, R.A., Cheung, N.-K. and Sondel, P.M. (1990) Cancer Res. 50, 5234.
[7] Naramura, M., Gillies, S.D., Mendelsohn, J., Reisfeld, R.A. and Mueller, B.M. (1993) Cancer Immunol. Immunother., in press. [8] Masui, H., Kawamoto, T., Sato, J.D., Wolf, B., Sato, G. and Mendelsohn, J. (1984) Cancer Res. 44, 1002. [9] Fleit, H.B., Wright, S.D. and Unkeless, J.C. (1982) Proc. Natl. Acad. Sei. USA 79, 3275. [10] Uchiyama, T., Broder, S. and Waldmann, T.A. (1981) J. Immunol. 126, 1393. [11] Tsudo, M., Kitamura, F. and Miyasaka, M. (1989) Proc. Natl. Acad. Sci. USA 86, 1982. [12] Mueller, B.M., Romerdahl, C.A., Trent, J.M. and Reisfeld, R.A. (1991) Cancer Res. 51, 2193. [13] Welch, D.R., Bisi, J.E., Miller, B.E., Conaway, D., Seftor, E.A., Yohem, K.H., Gilmore, L.B., Seftor, R.E.B., Nakajima, M. and Hendrix, M.J.C. (1991) Int. J. Cancer 47, 227. [14] Gillis, S., Ferrn, M.M., Ou, W. and Smith, K.A. (1978) J. Immunol. 120, 2027. [15] Herlyn, D., Herlyn, M., Steplewski, Z. and Koprowski, H. (1985) Cellular Immunol. 92, 105. [16] Duffy, M.J. (1992) Clin. Exp. Metastasis 10, 145. [17] Ravetch, J.V. and Kinet, J.-P. (1991) Annu. Rev. Immunol. 9, 457. [18] Caligiuri, M.A., Zmuidzinas, A., Manley, T.J., Levine, H., Smith, K.A. and Ritz, J. (1990) J. Exp. Med. 171, 1509. [19] Leonard, W.J., Depper, J.M., Robb, R.J., Waldmann, T.A. and Greene, W.C. (1983) Proc. Natl. Acad. Sci. USA 80, 6957.
99
Mechanisms of cellular cytotoxicity mediated by a recombinant antibody-IL2 fusion protein against human melanoma cells M a y u m i N a r a m u r a a, Stephen D. Gillies b, J o h n M e n d e l s o h n c, R a l p h A. Reisfeld a a n d B a r b a r a M. Mueller a'* ~The Scripps Research Institute, La Jolla, CA 92037, USA; bFuji ImrnunoPharmaceuticals Corporation, Lexington, MA 02173, USA; CMemorial Sloan-Kettering Cancer Center, New York, N Y 10021, USA (Received 1 July 1993; accepted 12 November 1993)
1.
Summary
Functional characteristics were established for a genetically engineered fusion protein between human IL2 and mouse/human chimeric mAb 225 directed against human epidermal growth factor receptor (EGFR), aberrantly expressed on human melanoma cells. The emphasis of these studies was on the mechanism(s) of action by which the ch225-IL2 fusion protein mediated cytotoxic killing of human melanoma cells by different human immune effector cells. Ch225-IL2 fusion protein bound to human EGFR with the high affinity of the parental antibody, and was as active as the equivalent amount of rhlL2. Ch225-IL2 enhanced cellular cytotoxicity mediated by freshly separated PBMC, isolated natural killer (NK) cells and activated T cells against melanoma cell lines. NK cells, which constitutively express both Fc~RIII and IL2R, interacted with ch225-IL2, mainly through FcTRIII, while the involvement of IL2R was secondary. However, the effect of ch225-IL2 on activated T cells was most likely mediated through IL2R. These results sug-
Key words: Recombinant antibody-IL2 fusion protein; Cellular cytotoxicity; Fc),RIII; IL2R; Melanoma; Immunotherapy "Corresponding author: Dr. Barbara M. Mueller, The Scripps Research Institute, Dept. of Immunology, IMM13, 10666 N. Torrey Pines Road, La Jolla, CA 92037, USA. Tel.: (619) 5548124; Fax: (619) 554-6705.
gest that the genetically engineered ch225-IL2 fusion protein may become a potent immunotherapeutic agent capable of stimulating various immune effector populations to effectively kill human melanoma cells. 2.
Introduction
Interleukin-2 (IL2) is a potent immunostimulator for a broad range of cell types, including both T and B cells, natural killer (NK) cells, monocytes and macrophages [1]. The use of IL2 led to some success in cancer immunotherapy [2], but systemic administration of high-dose IL2, initially employed in the clinic, was frequently accompanied by severe toxicity. However, doses of IL2 lower than those given originally showed some efficacy in modifying the host immune response without systemic toxicity [1]. An alternative to the systemic administration of IL2 may be the use of recombinant anti-tumor antibody-IL2 fusion proteins. Such constructs are designed to specifically target and concentrate IL2 at tumor sites while minimizing systemic toxicities and optimizing interactions between tumor cells and IL2-activated immune effector cells [3,4]. Mechanisms by which immune effector cells destroy tumor cells can be divided into two categories based on the involvement of receptors for IgG, i.e., Fc7 receptors (FcTRs). Thus, killing of tumor cells that occurs independent from Fc~Rs may involve cytotoxic T lymphocytes (CTLs), 91
NK cells and lymphokine-activated killer (LAK) cells. In contrast, monoclonal antibodies (mAbs) directed against tumor antigens effectively mediate antibody-dependent cellular cytotoxicity (ADCC) by immune effector cells bearing FcyRs such as NK cells, monocytes and macrophages. The importance of IL2 in both mechanisms has been well demonstrated, since this cytokine is not only indispensable for the generation of CTLs and LAK cells but also is capable of enhancing ADCC activities against tumor cells [5,6]. We previously reported the production of a recombinant anti-ganglioside GD2 antibody IL2 fusion protein (chl4.18-IL2) with full IL2 and antigen binding activity, capable of enhancing in vitro killing of autologous GD2-positive human melanoma cells by a tumor-infiltrating lymphocyte (TIL) cell line [4]. However, the mechanisms of action by which antibody-IL2 fusion proteins potentiate melanoma cell killing and the range of immune effector cell populations that can be activated remain to be clarified. Here we report on the functional characteristics of a recombinant fusion protein of human IL2 and a chimeric mouse/human monoclonal antibody (ch225) against the epidermal growth factor receptor (EGFR), aberrantly expressed on amelanotic human melanomas. The emphasis of these studies is on the mechanism of action by which this fusion protein (ch225-IL2) potentiates the killing of human melanoma cells by different immune effector cell populations. 3.
3.1.
Materials and Methods
Cytokines, antibodies and antibody-cytokine fusion protein
Recombinant human IL2 (rhlL2) derived from Escherichia coli was purchased from R&D Systems (Minneapolis, MN). Mouse/human chimeric anti-human EGFR mAb 225 (ch225) and its parental mouse anti-human EGFR mAb 225 (m225) were described previously [7,8]. The ch225-IL2 fusion protein was constructed as already reported for the ch14.18-IL2 fusion protein [4] by fusing a synthetic human IL2 sequence to the carboxyl end of the human C~I gene. Anti-human FcTRIII mAb 3G8 [9], anti-human IL2R p55 ct92
chain mAb anti-Tac [10], and anti-human IL2R p75 fl-chain mAb Mik-fll [11] were generously provided by Drs. J.C. Unkeless (Mount Sinai School of Medicine, New York, NY), R.P. Junghans (New England Deaconess Hospital, Boston, MA), and T.A. Waldmann (NIH, Bethesda, MD), respectively. Leu series mAbs were purchased from Becton Dickinson (San Jose, CA) and anti-CD3 mAb OKT3 was prepared from hybridoma culture supernatant in our laboratory.
3.2.
Tumor cell lines
Both M24met and C8161 are non-pigmented human melanoma cell lines that spontaneously metastasize in immunodeficient laboratory animals [12,13]. M24met was maintained in RPMI1640 medium supplemented with 10% fetal calf serum (FCS) and 2 mM L-glutamine. C8161 was kindly provided by Dr. M.J.C. Hendrix (University of Arizona, Tucson, AZ) and was maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FCS and 2 mM L-glutamine. Both cell lines were treated with 0.5 mM EDTA to detach from tissue culture flasks and were used as single cell suspensions.
3.3.
Indirect immunofluorescence
Melanoma cells were incubated with 1 /~g of either m225, ch225 or ch225-IL2. After removal of excess primary antibody by washing, cells were incubated with either goat anti-mouse IgG-FITC (for m225) or with goat anti-human IgG-FITC (for ch225 and ch225-IL2). Stained cells were analyzed in a FACScan flow cytometer (Becton Dickinson).
3.4.
Direct binding assays with radioiodinated proteins
Ch225 and ch225-IL2 were radiolabeled with Na125I (Amersham, Arlington Heights, IL) by the IODO-BEADS method according to the manufacturer's instructions (Pierce, Rockford, IL). Binding assays were performed and the data were analyzed, as described previously [7].
3.5.
IL2 assay
IL2 activity of the ch225-IL2 fusion protein was determined with a standard T-cell proliferation assay by measuring the [3H]thymidine incorporation of the IL2-dependent mouse cell line CTLL-2 [14].
3.6.
Effector cells
The source of effector cells was heparinized venous blood from healthy volunteer donors. Each of the individual experiments was performed with effector cells from a single donor. However, since we did not have access to the same donor repeatedly, the donor varied from one experiment to the other. Peripheral blood mononuclear cells (PBMC), isolated by centrifuging the whole blood through Ficoll-Paque (Pharmacia LKB, Piscataway, NJ), were suspended at 2 x 106/ml in RPMI-1640 and incubated in T150 tissue culture flasks (Corning, Corning, NY) at 37°C for 1 h. This procedure removes the majority of monocytes. Unstimulated T cells were prepared from the nonadherent population by negative immunomagnetic selection. Cells were coated with mAbs directed against CD14 (Leu-M3), CD19 (Leu-12) and CD56 (Leu-19), and incubated with beads coated with goat anti-mouse IgG (Dynabeads; Dynal, Great Neck, NY). Monocytes, N K cells and B cells were removed in a magnetic field. The resulting cell population was typically > 9 0 % CD3 +, < 3 % CD 16 +, when analyzed in a FACScan flow cytometer. Activated T cells were obtained by stimulating T cells with anti-CD3 mAb and rhlL2. To this end, T cells were positively selected from plastic-nonadherent cells in flasks coated with anti-CD3 mAb (MicroCELLector; AIS, Santa Clara, CA), followed by incubation for 3-5 days in RPMI-1640/10% FCS with 25 ng/ml rhlL2. The cells ( > 9 5 % CD3+, > 8 0 % CD25 ÷) were further separated into CD4 + and CD8 + populations in anti-CD4 and anti-CD8 coated flasks, respectively. To isolate N K cells, T cells, B cells and monocytes, PBMCs were depleted by adhesion to anti-CD3 mAb-coated flasks and negative immunomagnetic selection with anti-CD3 (OKT3), anti-CD14 (Leu-M3) and anti-CD19 (Leu-12) mAbs. The resulting popula-
tion contained > 7 0 % CD16*, > 8 0 % CD56" cells.
3.7.
Cytotoxicity assays
Cytotoxic activity of various effector cell populations was measured in 4-h SlCr-release assays against melanoma cell lines, as described previously [7]. All the effector cells were tested in triplicate at target to effector (T/E) ratios of 1:100, 1:50, 1:25, and 1:12.5, except for N K cells, where T/E ratios were 1:40, 1:20, 1:10, and 1:5, respectively. In some cases, cytotoxicity assays were done in the presence of ~2-antiplasmin. To investigate the involvement of FcTRIII and IL2R, effector cells were first incubated with either antiFcTRIII mAb 3G8 or anti-IL2R mAbs anti-Tac and Mik-fll at 37°C for 30 rain, prior to adding tumor cells and anti-EGFR mAbs. Statistical analysis of the data was performed by Student's t test. A probability of less than 5% (P<0.05) for no difference between the experimental group and the control group was considered significant [15]. 4.
4.1.
Results
Antigen-binding activity of ch225-IL2 fusion protein
Both human melanoma cell lines, M24met and C8161, were shown to express a comparable number of E G F R molecules on the surface by indirect flow cytometry (Fig. 1). The parental murine mAb and its genetically engineered derivatives all showed essentially identical binding patterns. The dissociation constants (KDS) for 125I-labeled ch225 and ch225-IL2 were calculated from saturation binding curves obtained with the M24met cell line. The average KDS in three independent assays were 2.62 x 10 -1° M for ch225 and 3.31 x 10 -1° M for ch225-IL2, indicating that fusing the cytokine IL2 to this chimeric antibody did not change its binding characteristics.
4.2.
IL-2 activity
The IL-2 activity of ch225-IL2 was measured as the proliferation of the IL2-dependent cell line CTLL-2 and compared against that of rhlL2. 93
a)
b) 80
80 M24met =.60-
c.~5~,~ 1
E ;~ 400
C8161
ch225..IL2
60 . Go~HIgG
40
ch225-1L2
J~
m225
20
200
/
,
m~5
i
J
i lll.~
i
i
101
10 o
i
ill.'~
i
102
i
i
Illll
t
103
0
i
J
i
ill.
100
I
101
i
i
i
ill.
I
10 ~
I
I
I
ll.I
103
Fluorescence Intensity Fig. 1. Flow cytometric analysis of human melanoma cell lines M24met (a) and C8161 (b) incubated with either murine anti-human EGFR mAb m225 and goat anti-mouse [gG-FITC; mouse/human chimeric mAb ch225 and goat anti-human IgG-FITC; or ch225 IL2 fusion protein and goat anti-human IgG-FITC. Cells incubated only with goat anti-human IgG-FITC are shown as control.
Cells were incubated with serial dilutions of rhlL2, ch225 or ch225-IL2 covering the range of 0.01-10000 pM. The concentration of IL2 in the ch225-IL2 fusion protein was calculated as two molar equivalents of IL2 per mole of fusion protein. As shown in Fig. 2, IL2 activity of ch225IL2 was essentially identical to that of a comparable amount of rhlL2. The anti-human E G F R antibody ch225 alone had no effect on CTLL-2 proliferation.
30.
i10. t.
7,
¢, z
0
.01
t
1
10
IL2 (pM)
100
1000
10000
Fig. 2. IL2 activity of ch225-IL2 fusion protein (O) as measured by [3H]thymidine incorporation of the CTLL-2 cell line. Recombinant IL2, expressed in E. coli, was used as control (C)). Results are normalized to the IL2 concentration. Bars denote the standard deviation.
94
4.3.
Cytotoxicity mediated by fresh PBMC and NK cells
We have previously reported that ch225 enhances cytotoxic activity of human PBMC against a melanoma cell line bearing EGFR, but that it does not mediate complement-dependent cytolysis [7]. Thus, in the first series of experiments, the ability of ch225-IL2 to potentiate cellular cytotoxicity was compared to that of ch225. PBMC freshly separated from the blood of healthy volunteer donors were better able to mediate cytotoxicity against M24met and C8161 cells with ch225-IL2 than with ch225 (Fig. 3). The carboxyl terminal lysine residue on the chl4.18-IL2 fusion protein heavy chain is accessible to cleavage by plasmin [4]. Since FCS contains small amounts of plasminogen and tumor cells are capable of producing plasminogen activators [16], it was necessary to determine whether the intact fusion protein molecule is capable of stimulating effector cells or whether it needs to be cleaved into its ch225 and IL2 components to exhibit biological activities. To clarify this point, cytotoxicity assays were done in the presence or absence of 50 /~g/ml ~2-antiplasmin, a potent inhibitor of plasmin. Addition of ~2-antiplasmin did not change the ability of ch225-IL2 to enhance cellular cytotoxicity mediated by PBMC, indicating that the intact fusion protein is fully capable to interact with effector cells (data not shown).
(a)
(b) 6O
70
5O
60
-lOd
"r
Control
r
eh 995
"r
a
0
~ 225-11.2
,
,
,
Control
e~ 225
eh ~5-11.2
Fig. 3. Enhancement of cellular cytotoxicity mediated by P B M C against h u m a n melanoma cells M24met (a) and C8161 (b) coated with 1 /~g/ml ch225 or 1 /~g/ml ch225-IL2. Cytotoxicity was measured in a 4-h SlCr release assay at target-to-effector cell ratios of 1:100 ([]), 1:50 (I--1), 1:25 (11) and 1:12.5 ([]). Standard deviations are indicated. The probability of no difference between controls and ch225 or ch225-IL2 coated tumor cells, respectively, as determined by Student's t test was P < 0.005.
To further investigate the respective mechanisms of action of ch225-IL2 and ch225, we tried to identify the effector population(s) that interact with the ch225-IL2 fusion protein. Since preliminary experiments indicated that purified granulocytes did not have any cytotoxic activity against our melanoma cell lines, we mainly focused on lymphocytes. In this regard, the data in Table 1 show that ch225 and ch225-IL2 can mediate NK cells cytotoxicity, while resting T cells are not stimulated by either chimeric antibody or the fusion protein. Unlike fresh PBMC, purified NK cells coated with ch225 or ch225-IL2 killed melanoma target cells with the same efficacy.
4.4.
Involvement of FcTRIII and 1L2R in NKmediated cytotoxicity
NK cells are unique in that they express solely FcvRIII and not the other two types of Fc),Rs [17], and that they express IL2R constitutively [18]. The majority of IL2Rs on the surface of NK cells are of intermediate affinity, being mostly composed of p75 r-chain and p70 y-chain [1]. The number of p55 a-chains is quite limited in NK cells. We hypothesized that cytotoxicity by IL2-antibody fusion proteins is mediated through Fc~,R and/or IL2R. To this end, we dissected the interaction of NK cells with ch225 and
TABLE 1 C Y T O T O X I C I T Y O F R E S T I N G T A N D N K CELLS M E D I A T E D BY T H E ch225-IL2 F U S I O N P R O T E I N A G A I N S T H U M A N M E L A N O M A CELLS Percent o f specific lysis of 51Cr-labeled target ceils in 4-h assays at target-to-effector cell ratios of 1:100 (T cells) and 1:40 ( N K ceils), expressed as average + standard deviation (n = 3). Target cells
M24met
C8161
Effector cells
T cells
N K cells
T cells
N K cells
Control ch225 (1 /~g/ml) ch225-1L2 (1 /tg/ml)
2.98 + 2.93* 3.98 + 1.46 --0.48 ___ 2.62
2.02 ___ 4.88 16.38 + 2.79 13.70 _ 1.26
0.79 ___ 0.89 4.88 + 0.64 7.23 ___ 2.20
2 1 . 8 0 _ 0.78 43.22 _ 4.47 37.71 + 1.51
95
TABLE 2 E F F E C T OF ANT1-Fc~RIII and ANTI-IL2R ANTIBODIES O N N K CELL C Y T O T O X I C I T Y M E D I A T E D BY THE ch225-IL2 FUSION PROTEIN N K cells were coated with either 20/ag/ml 3G8 anti-FcyRIII mAb or 10/~g/ml each of anti-Tac and Mik-fll anti-IL2R mAbs. Figures in the table denote the average percent specific lysis _ standard deviation determined in triplicate in 4-h 51Cr release assays at targetto-effector cell ratios of 1:40. M24met NK
C8161 N K + 3G8
N K + ~tlL2R
NK
N K + 3G8
1.48 _ 0.69 - 0 . 4 4 ___ 1.65 7.85 + 2.031
6.06 + 1.59 n.d." 32.76 + 3.512
3.43 + 0.13 38.40 ___ 2.57 34.25 _ 1.56
1.78 _ 1.71 2.15 + 1.57 5.36 + 1.672 n.d. n.d.
0.87 ___ 0.94 n.d. 8.37 ___ 0.892 -0.11 _ 2.86 n.d.
NK+~IL2R
Experiment 1 Control ch225 (1 #g/ml) ch225-IL2 (1 #g/ml)
7.67 + 1.26 46.10 _ 1.54 45.72 _+ 2.01
- 3 . 1 0 ___ 0.38 - 1 . 9 8 + 1.27 4.24 + 1.421
7.46 ___ 1.55 n.d. 30.99 + 2.713
1 1 . 8 3 _+ 1.50 14.24 _ 1.18 1 5 . 4 5 _ 0.791 n.d. n.d.
18.21 _ 1.86 n.d. 32.72 + 3.002 18.84 + 2.54 n.d.
Experiment 2 Control ch225 (0.1 #g/ml) ch225-IL2 (0.1 #g/ml) rhlL2 (20 ng/ml) ch225 (0.1 /zg/ml) + rhlL2 (20 ng/ml)
-1.83 19.28 16.90 0.70 19.61
___ 1.91 _+ 2.38 _ 1.29 _ 2.80 + 0.00
17.63 _ 39.68 + 48.95 + 33.20 + 47.77 _
0.77 1.40 1.55 0.43 1.81
Probability of no difference to cytotoxicity mediated by N K ceils and ch225-IL2 as determined by Student's t test: 1p<0.005; 2 P < 0.05; 3P> 0.05. "n.d.: not determined.
ch225-IL2 by blocking either FcTRIII or IL2R with their specific antibodies. When N K cells were coated with 20 /~g/ml anti-human FcTRIII mAb 3G8, neither 1 #g/ml ch225 nor ch225-IL2 could enhance N K cytotoxicity against melanoma cells (Table 2, Exp. 1). However, blocking IL2R with 10 #g/ml each of anti-Tac anti-human p55 IL2R ~-chain and Mik-fll anti-human p75 IL2R fl-chain mAbs resulted in only a slight reduction in the activity of ch225-IL2. The concentration of IL2 in 1 #g/ml ch225-IL2 is approximately 160 ng/ml (11 nM), which may not have been blocked completely by 10 #g/ml each of anti-IL2R mAbs [18]. Therefore, in the next set of experiments, 10-fold smaller concentrations of ch225 and ch225-IL2 were tested. In this case, both ch225 and ch225-IL2 were effective in potentiating N K cells at 0.1 #g/ml. However, while anti-FcyRIII m A b almost completely abrogated the potentiating effect of ch225 and ch225-IL2, anti-IL2R mAbs were less effective in reducing the activity of ch225-IL2 (Table 2, Exp. 2). The 96
blocking of IL2 b~nding to its receptor by 10/~g/ ml each of anti-IL2R mAbs was confirmed in a control experiment where the effect of an equivalent amount of rhlL2 (20 ng/ml) was abolished. To demonstrate that this partial reduction in cytotoxic activity of the fusion protein is due to the blocking of the IL2R and not the presence of a large amount of an irrelevant Ab, N K cells were coated with the same amount of isotype-matched (IgG2a) murine mAb OKT3. This maneuver did not affect the ability of either ch225 or ch225IL2 to mediate cellular cytotoxicity (data not shown). Therefore, we concluded that N K cells interact with ch225-IL2 mainly through FcyRIII and that the involvement of IL2R is secondary. The modification of the Fc region of ch225 in the ch225-IL2 fusion protein did not block its binding to the Fc~RIII. 4.5.
Cytotoxicity mediated by activated T cells
It has already been reported that recombinant
TABLE 3 C Y T O T O X I C I T Y O F A C T I V A T E D T CELLS M E D I A T E D BY T H E ch225-IL2 F U S I O N P R O T E I N A G A I N S T H U M A N M E L A N O M A CELLS H u m a n T cells activated by anti-CD3 antibodies and rhIL2 were separated into CD4 + and CD8 + subpopulations. Cytotoxicity was evaluated in 4-h 51Cr release assays. Figures in the table denote average percent of specific lysis ___ standard deviation obtained in triplicate assays at a target-to-effector cell ratio of h l00. M 24met
C8161
CD4
CD8
CD4
CD8
4.45 + 0.64 5.55 _+ 0.20* 24.22 __ 1.40"*
8.07 _+ 0.52 8.46 + 0.29" 19.90 _+ 0.83**
17.29 _+ 0.32 23.37 ___ 1.28" 42.00 _+ 1.37""
Experiment 1 Control ch225 (1 /~g/ml) ch225-IL2 (1 ~g/ml)
--0.60 _+ 0.47 7.86 + 0.88* 12.55 ___ 0.51"* CD8
CD8
Experiment 2 Control ch225 (1 #g/ml) ch225-IL2 (1 /~g/ml) rhlL2 (200 ng/ml) ch225 (1 #g/ml) + rhlL2 (200 ng/ml)
35.87 35.92 51.85 52.01 43.90
+ 2.88 _+ 1.35" + 2.66** + 2.15"* + 0.67 *°
1.37 3.47 37.35 23.46 26.87
+ _+ + + +
3.11 3.11" 2.45"* 1.99"* 2.10"*
17.50 18.49 18.82 19.26 20.63
_+ 0.69 + 1.01 ° ___ 0.02" + 1.95" + 1.00""
41.47 33.79 34.27 36.06 37.48
_+ 4.89 + 1.12" ___ 3.79* ___ 2.64* + 0.58*
Experiment 3 Control ch225 (0.1 #g/ml) ch225-IL2 (0.1 /~g/ml) rhlL2 (20 ng/ml) ch225 (0.1 #g/ml) + rhlL2 (20 ng/ml)
Probability of no difference to untreated controls by Student's t test: "P>0.05; **P<0.05.
m A b - I L 2 fusion proteins enhance melanoma cell killing by activated T cells from peripheral blood [3] or by a TIL cell line [4]. We tested the cytotoxicity of activated T cells against two melanoma cell lines, M24met and C8161, established from different patients, by activating T cells nonspecifically through CD3 and by stimulation with exogenous rhIL2. In this case, the activated T cells themselves were capable of killing melanoma cells, and this effect was further enhanced in the presence of ch225-IL2. Since we found that CD8 ÷ cell populations are much more potent effectors than CD4 + T cells (Table 3, Exp. 1), we tested only activated CD8 + cells in the experiments described in Table 3. An equivalent amount of rhIL2 was also found capable of stimulating activated T cells to lyse melanoma targets (Table 3,
Exp. 2), implying that the effect of ch225-IL2 on activated T cells was most likely mediated through IL2R. This is in contrast to observations made with N K cells where 0.1 #g/ml ch225-IL2 was sufficiently effective to stimulate these effector cells, whereas equivalent amounts of ch225IL2 or rhlL2 did not potentiate activated T cells (Table 3, Exp. 3). 5.
Discussion
We report here on the functional characteristics of a fusion protein between mouse/human chimeric anti-EGFR mAb and human IL2, ch225IL2, with emphasis on the mechanism of action by which different immune effector cells potentiate cytotoxic killing of human melanoma cells. 97
This fusion protein retains the high antigen binding affinity of the parental mAb ch225 and is biologically as active as an equivalent amount of rhlL2 in supporting proliferation of an IL2-de'pendent cell line. Furthermore, ch225-IL2 is capable of stimulating immune effector cells to mediate cellular cytotoxicity against M24met and C8161 melanoma cell lines, that express comparable numbers of E G F R s on their surface. These cell lines were established from melanoma metastases of two different individuals and vary in their susceptibility to cellular cytotoxicity. In 4-h cytotoxicity assays, ch225-IL2 interacted with effector cells as an intact molecule, and does not need to be enzymatically cleaved into its ch225 and IL2 components to be active. Fell et al. [3] previously reported the production of a fusion protein, L6/IL2, consisting of a mouse/human chimeric F(ab') with specificity for carcinomas and human IL2. This fusion protein potentiated the cytotoxicity of CD3-activated blast cells, but not that of resting PBL. Since the major difference between L6/IL2 and our ch225IL2 fusion protein is the presence of the immunoglobulin Fc portion in the latter, we postulated that the effect of ch225-IL2 on fresh PBMC is mediated through FcTRs that are expressed on a variety of leukocytes, including monocytes, macrophages, neutrophils, N K cells, as well as B and T cells. When testing the capacity of neutrophils, N K and T cells from peripheral blood to mediate ADCC, we found N K cells more effective in mediating cytotoxicity against melanoma cells coated with ch225-IL2 than T cells. In view of the fact that fresh PBMC mediated more potent cytotoxicity when stimulated with ch225-IL2 than with ch225, and that ch225-IL2 could potentiate the purified N K cells only equally as well as ch225, the contribution of other populations in PBMCs, such as monocytes/macrophages, will have to be further investigated. We were able to show in this report that ch225-IL2 interacts with N K cells mainly through Fc~RIII and that the involvement of IL2R is secondary. It is to be noted that binding of ch225-IL2 to Fc),RIII is retained regardless of addition of IL2 molecules to the Fc portion of this fusion protein. This characteristic should allow ch225-IL2 to stimulate a wide variety of im98
mune effector cells bearing FcTRs and thus hopefully translate into a strong anti-tumor effect, possibly better than that mediated by F(ab') fusion proteins [3]. We also tested the capacity of activated T cells as effector cells in order to compare the efficacy of ch225-IL2 with other antibody-IL2 fusion proteins previously reported. Although resting T cells are not capable of mediating cytotoxicity with ch225 or ch225-IL2, activated CD8 ÷ cells can be further potentiated by ch225-IL2. The discrepancy in cytotoxic activity between ch225 and ch225-IL2 implies that in this case, the effect of ch225-IL2 is not mediated through FcTRIII. The fact that excess amounts of heat-inactivated human serum (1 mg/ml human IgG) could not abrogate the effect of ch225-IL2 on activated T cells, also supports the contention that this phenomenon is independent of FcTR (data not shown). We also tried to prove that ch225-IL2 potentiates activated T cells through IL2R. However, in our hands, anti-IL2R mAbs did not abrogate the effect of ch225-IL2. It is possible that the inability to block the cytotoxic activity of ch225-IL2 with anti-IL2R mAbs does not necessarily contradict our hypothesis, because 1 /~g/ml of ch225-IL2 fusion protein contains the equivalent amount of about 150 to 200 ng/ml IL2 (10-15 nM), and anti-Tac is known to be unable to inhibit the binding of such a large amount of IL2 to the receptor [19]. For this reason, we repeated the same experiment with a 10-fold lower dose of ch225 IL2; however, neither 0.1 #g/ml of ch225-IL2, which was effective to abrogate the effect of N K cells, nor 20 ng/ml rhlL2 were sufficient to enhance the killing of human melanoma cells by activated CD8 + cells. This requirement for relatively high concentrations of IL2 for the cytotoxicity of activated T cells is in agreement with the observation of Fell et al. [3]. One of the rationales behind fusing IL2 with an antibody against a tumor-associated antigen is to deliver relatively larger amounts of this immunostimulator directly to tumor sites and with hopefully fewer of the toxicities observed with large amounts of systemically applied rhlL2. In this regard, preliminary experiments with radiolabeled ch225-IL2 indicated that ch225-IL2 can effectively target to tumor xenografts in scid mice.
These data suggest that c h 2 2 5 - I L 2 m a y be a potent i m m u n o t h e r a p e u t i c agent that c a n recruit effector cells expressing FcvRs a n d that also m i g h t be capable o f delivering relatively large a m o u n t s o f IL2 to t u m o r sites, to locally activate T I L s / CTLs.
Acknowledgements This w o r k was s u p p o r t e d by grants from the N a t i o n a l C a n c e r Institute 1R35CA42508 a n d 1U01CA51946. This is The Scripps Research Institute's m a n u s c r i p t n u m b e r 8 0 1 7 - I M M .
References [1] Smith, K.A. (1993) Blood 81, 1414. [2] Maas, R.A., Dullens, H.F.J. and Den Otter, W. (1993) Cancer Immunol. Immunother. 36, 141. [3] Fell, H.P., Gayle, M.A., Grosmaire, L. and Ledbetter, J.A. (1991) J. Immunol. 146, 2446. [4] Gillies, S.D., Reilly, E.B., Lo, K.-M. and Reisfeld, R.A. (1992) Proc. Natl. Acad. Sci. USA 89, 1428. [5] Honsik, C.J., Jung, G. and Reisfeld, R.A. (1986) Proc. Natl. Acad. Sci. USA 83, 7893. [6] Hank, J.A., Robinson, R.R., Surfus, J., Mueller, B.M., Reisfeld, R.A., Cheung, N.-K. and Sondel, P.M. (1990) Cancer Res. 50, 5234.
[7] Naramura, M., Gillies, S.D., Mendelsohn, J., Reisfeld, R.A. and Mueller, B.M. (1993) Cancer Immunol. Immunother., in press. [8] Masui, H., Kawamoto, T., Sato, J.D., Wolf, B., Sato, G. and Mendelsohn, J. (1984) Cancer Res. 44, 1002. [9] Fleit, H.B., Wright, S.D. and Unkeless, J.C. (1982) Proc. Natl. Acad. Sei. USA 79, 3275. [10] Uchiyama, T., Broder, S. and Waldmann, T.A. (1981) J. Immunol. 126, 1393. [11] Tsudo, M., Kitamura, F. and Miyasaka, M. (1989) Proc. Natl. Acad. Sci. USA 86, 1982. [12] Mueller, B.M., Romerdahl, C.A., Trent, J.M. and Reisfeld, R.A. (1991) Cancer Res. 51, 2193. [13] Welch, D.R., Bisi, J.E., Miller, B.E., Conaway, D., Seftor, E.A., Yohem, K.H., Gilmore, L.B., Seftor, R.E.B., Nakajima, M. and Hendrix, M.J.C. (1991) Int. J. Cancer 47, 227. [14] Gillis, S., Ferrn, M.M., Ou, W. and Smith, K.A. (1978) J. Immunol. 120, 2027. [15] Herlyn, D., Herlyn, M., Steplewski, Z. and Koprowski, H. (1985) Cellular Immunol. 92, 105. [16] Duffy, M.J. (1992) Clin. Exp. Metastasis 10, 145. [17] Ravetch, J.V. and Kinet, J.-P. (1991) Annu. Rev. Immunol. 9, 457. [18] Caligiuri, M.A., Zmuidzinas, A., Manley, T.J., Levine, H., Smith, K.A. and Ritz, J. (1990) J. Exp. Med. 171, 1509. [19] Leonard, W.J., Depper, J.M., Robb, R.J., Waldmann, T.A. and Greene, W.C. (1983) Proc. Natl. Acad. Sci. USA 80, 6957.
99