ARTICLE
Received 8 Apr 2016 | Accepted 24 Mar 2017 | Published 12 May 2017
Rodrigo Vazquez-Lombardi1,2, Claudia Loetsch1,2, Daniela Zinkl1, Jennifer Jackson1, Peter Schoeld1,Elissa K. Deenick1,2, Cecile King1,2, Tri Giang Phan1,2, Kylie E. Webster1,2, Jonathan Sprent1,2 & Daniel Christ1,2
Interleukin-2 (IL-2) is an established therapeutic agent used for cancer immunotherapy. Since treatment efcacy is mediated by CD8 and NK cell activity at the tumour site, considerable efforts have focused on generating variants that expand these subsets systemically, as exemplied by IL-2/antibody complexes and superkines. Here we describe a novel determinant of antitumour activity using fusion proteins consisting of IL-2 and the antibody fragment crystallizable (Fc) region. Generation of long-lived IL-2-Fc variants in which CD25 binding is abolished through mutation effectively prevents unwanted activation of CD25 regulatory T-cells (Tregs) and results in strong expansion of CD25 cytotoxic subsets. Surprisingly, however, such variants are less effective than wild-type IL-2-Fc in mediating tumour rejection. Instead, we report that efcacy is crucially dependent on depletion of Tregs through Fc-mediated immune effector functions. Our results underpin an unexpected mechanism of action and provide important guidance for the development of next generation IL-2 therapeutics.
DOI: 10.1038/ncomms15373 OPEN
Potent antitumour activity of interleukin-2-Fc fusion proteins requires Fc-mediated depletion of regulatory T-cells
1 Immunology Division, Garvan Institute of Medical Research, Sydney, New South Wales 2010, Australia. 2 St Vincents Clinical School, University of New South Wales, Sydney, New South Wales 2010, Australia. Correspondence and requests for materials should be addressed to J.S.(email: mailto:[email protected]
Web End [email protected] ) or to D.C. (email: mailto:[email protected]
Web End [email protected] ).
NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 1
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373
Interleukin-2 (IL-2) is a pleiotropic cytokine essential for the development, activation and homoeostasis of multiple lymphocyte subsets1. Initially identied as a potent T-cell
proliferating factor present in mixed leukocyte cultures2, IL-2 was rst cloned and synthesized in Escherichia coli in 1983 (refs 3,4) and underwent initial clinical evaluation for cancer indications in 1985 (ref. 5). Despite severe toxicity, the potent antitumour activity observed in a subset of patients led to the regulatory approval of a high-dose recombinant IL-2 formulation for cancer immunotherapy of metastatic renal cancer in 1992 and for metastatic melanoma in 1998 (ref. 6).
IL-2 is a member of the common gamma-chain (gc) cytokine family and shares the gc receptor subunit with IL-4, IL-7, IL-9, IL-15 and IL-21. Importantly, immune cells express dimeric or trimeric IL-2 receptors (IL-2R), with the former composed of IL-2Rb (CD122) and gc (IL-2Rbgc intermediate-afnity receptor, KDB1 nM), and the latter composed of IL-2Ra (CD25), IL-2Rb and gc (IL-2Rabgc high-afnity receptor, KDB10 pM)7,8. Dimeric receptors are expressed on cytotoxic CD8 T-cells and natural killer (NK) cells, while trimeric receptors are predominantly displayed on activated lymphocytes and CD4 CD25 FoxP3 regulatory T-cells (Tregs)1. Due to its increased afnity for trimeric receptors, IL-2 induces preferential stimulation of Tregs, which are crucial for maintaining immune tolerance and display pro-tumourigenic activity9. In addition to the undesired promotion of Treg proliferation, clinical use of IL-2 is further complicated by a short serum half-life (B7 min) and dose-limiting toxicities10,11.
These shortcomings are greatly improved when IL-2 is complexed with anti-IL-2 antibodies. IL-2/mAb complexes display a substantial prolongation of serum half-life and modulate IL-2 selectivity for specic immune cell subsets12. In particular, IL-2/mAb complexes that target IL-2 to cells expressing CD122, but not CD25 (such as Tregs), induce preferential expansion of CD122high populations (that is, memory-phenotype (MP) CD8
T-cells and NK cells), leading to increased antitumour activity at low doses and reduced toxicity in animal models of malignancy13. The favourable properties of IL-2/mAb complexes result from a combination of factors including reduced renal clearance, FcRn recycling, and steric blockade of the CD25-binding site14. In addition to IL-2 immunocomplexing, modulation of activity has been reported for engineered IL-2 superkine variants with altered binding to components of the IL-2 receptor, namely binding to IL-2Ra (refs 1517), IL-2Rb (refs 16,18,19) and gc (refs 16,19).
Although cytokineantibody complexes and variants are considerably more potent than unmodied cytokines12,18,20,21, development into validated human therapeutics has so far not been demonstrated. While this is likely a reection of their relative recent discovery, the need for humanization of the antibody component and the requirement for formulating multiple proteins complicate development and regulatory approval of complexes, whereas IL-2 superkines suffer from short serum half-lives, due to their low molecular mass and absence of half-life extension. By contrast, Fc-fusion proteins generated through the genetic linkage of antibody Fc regions with an effector moiety (such as a cytokine or cytokine receptor) have a well-established track record as human therapeutics, as exemplied by the TNFR2-Fc-fusion protein etanercept (Enbrel)22. Indeed, a large proportion of new biologic drugs contain antibody Fc regions due to the commercial requirement for half-life extension23, further highlighting the relevance of this format for drug development applications. As such, a number of IL-2-Fc-fusion proteins with therapeutic potential for induction of transplantation tolerance19,2426 and prevention of autoimmunity27,28 have been reported. Furthermore, recent
studies highlight the synergistic nature of combination therapy strategies consisting of antitumour antigen antibodies and IL-2-IgG fusions in models of malignancy29,30. Here we apply the Fc-fusion protein concept to the IL-2 system and systematically investigate the contribution of cytokine and Fc components to antitumour activity.
ResultsAbolition of CD25 binding enhances IL-2-Fc selectivity. We rst generated genetic fusions, by linking human IL-2 with the Fc region of murine IgG2c by means of a short glycine-serine linker (see Methods). The fusion proteins display a molecular weight of B80 kDa (Supplementary Fig. 1A) compared to about 15 kDa for human IL-2. As observed for IL-2/mAb complexes, the presence of the Fc component results in a molecular mass well above the glomerular ltration cut-off (B60 kDa), and increased serum half-life through reduced renal clearance and FcRn recycling31.
To reduce binding of our IL-2-Fc construct to CD25 cells (and Tregs in particular), we introduced mutations directed at disrupting the IL-2/CD25 interaction (Fig. 1a, Supplementary Fig. 1). Inspection of the quaternary IL-2/IL-2R complex structure32 revealed that many of the residues in the IL-2/CD25 interface are charged and participate in electrostatic interactions with the receptor. This motivated us to introduce charge-reversal mutations at contact positions in order to considerably reduce binding afnity and activation of CD25 cells (see Supplementary Discussion). Our strategy relied on the introduction of charge-reversal substitutions directly into bivalent IL-2-Fc constructs, thus differing from previous approaches that have targeted both charged and aromatic residues in monovalent unfused IL-2 (refs 15,33).
Although binding was considerably reduced, residual afnity to CD25 was observed for all designed single mutations, as well as for a previously reported F42A mutant34, particularly when expressed bivalently in an Fc-fusion format (Supplementary Fig. 1). To further reduce CD25 interactions, we next combined single mutations in a step-wise manner, rst into double (Supplementary Figs 13), and then into triple mutants (Supplementary Fig. 4). Scanning mutagenesis using up to 14 different amino acid substitutions at targeted residues revealed strong positional effects and the requirement for at least three mutations in the interface to abolish activation of CD25 cells both in vitro and in vivo (Supplementary Figs 4 and 5).
We then benchmarked the activity of a novel IL-23XFc triple mutant (R38D, K43E, E61R; Fig. 1a) against IL-2WTFc and IL-2/mAb immune complexes consisting of human IL-2 and the mouse anti-human antibody MAB602 (see Methods). Single-dose IL-23XFc induced robust expansion of MP CD8 and NK cell subsets in the spleens of C57BL/6 mice, substantially higher than what was observed not only for IL-2WTFc, but also for treatment with IL-2/mAb immune complexes (Fig. 1b). The superior activity of mutant IL-2 fusion protein in a single-dose setting was consistent with a prolonged serum half-life relative to IL-2/mAb complexes and IL-2WTFc (Supplementary Figs 3E and 8). Notably, IL-23XFc administration induced minimal expansion of Tregs conrming its high level of selectivity for CD8 and NK subsets (Fig. 1b).
IL-23XFc also drove potent expansion of cytotoxic subsets in multiple low-dose treatments, similar to what was observed for IL-2/mAb, but substantially higher than the parental IL-2WTFc protein (Fig. 1c). However, while IL-2/mAb complexes caused considerable Treg expansion (B5-fold), treatment with multiple low doses of IL-23XFc failed to induce Treg expansion (Fig. 1c).
2 NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373 ARTICLE
a
IL-23XFc
K43E
CD25
E61R
R38D
CD122
b Single high dose c Multiple low dose
5
35
MP CD8 NK cells Tregs
35
MP CD8 NK cells Tregs
Cell expansion relative
to PBS controls
****
25
25
****
****
Cell expansion relative
to PBS controls
****
****
15
15
****
****
****
*
***
5
NS****
****
****
****
*
NS
NS
PBS
IL-2/mAb
IL-2WT Fc
IL-23X Fc
PBS
IL-2/mAb
IL-2WT Fc
IL-23X Fc
Figure 1 | Abolition of CD25 binding results in potent and selective expansion of cytotoxic lymphocyte subsets. (a) Selected charge-reversal mutations were introduced into human IL-2 based on inspection of the IL-2/IL-2R co-crystal structure32 (PDB: 2B51), with the aim of disrupting CD25 binding. Highlighted are the mutations introduced to generate the IL-23XFc triple mutant. R38D and E61R were selected after binding kinetics and cell-based assays (Supplementary Figs 1 and 2), while K43E was incorporated after screening of in vivo activity (Supplementary Fig. 4AG). (b,c) Lymphocyte expansion proles in the spleens of C57BL/6 mice receiving IL-2/mAb, IL-2WTFc or IL-23XFc treatment. (b) Fold expansion in the total numbers of memory-phenotype CD8 T-cells (CD8 CD44high CD122high), NK cells (CD3 NK1.1 CD122high) and Tregs (CD4 FoxP3 CD25) after a single IL-2/mAb (3 mg IL-2 15 mg mAb) or IL-2-Fc (16.8 mg) i.p. injection was determined by ow cytometry on day 5. Shown is pooled data from seven independent experiments,
each normalized to the average subset numbers of two to three PBS-treated mice (PBS, n 16; treatment groups, n 48). (c) Fold expansion in the total
numbers of MP CD8, NK cells and Tregs after ve IL-2/mAb (1 mg IL-2 5 mg mAb) or IL-2-Fc (5.6 mg) i.p. injections (days 04) was determined by ow
cytometry on day 5. Shown is pooled data from ve independent experiments, each normalized to the average subset numbers of two to three PBS-treated mice (PBS, n 18; treatment groups, n 8). Data are displayed as means.e.m. Asterisks indicate signicant differences relative to PBS controls
(*Po0.05, ***Po0.001, ****Po0.0001) as determined by one-way analysis of variance with Bonferroni post hoc test for multiple comparisons.
Taken together, these experiments demonstrated that our design objectives had been achieved; with IL-23XFc treatment causing prominent expansion of CD25 MP CD8 and NK cells but no expansion of CD25 Tregs.
Low toxicity and potent antitumour activity of IL-2WTFc. Having successfully designed a highly active and selective IL-23XFc triple mutant we proceeded to evaluate its therapeutic potential. First, we examined mice for signs of treatment-associated toxicity. Notably, multiple injections of IL-23XFc resulted in weight loss, suggesting that this variant induces systemic toxicity at the administered dose (Fig. 2a). Next, we assessed mice for pulmonary oedema and compromised hepatic function as a measure of experimentally induced vascular leak syndrome, a hallmark side effect of IL-2 therapy13,35. Treatment with IL-2/mAb or IL-23XFc induced pulmonary oedema, as evidenced by increases in lung water content (Fig. 2b; Supplementary Fig. 6A). By contrast, lung water weight in mice treated with IL-2WTFc remained largely unchanged relative to PBS controls, either as absolute weight or as percentage of total body weight (Fig. 2b; Supplementary Fig. 6A). Assessment of liver weight and aspartate aminotransferase activity in serum revealed no differences to PBS controls across all treatment groups (Fig. 2c,e; Supplementary Fig. 6B). However, a signicant elevation in alanine aminotransferase activity in serum
was observed in the IL-2/mAb group only (Fig. 2d; one-way analysis of variance, P 0.0065). These low levels of hepatic
toxicity may reect the lower maximal serum concentrations of administered cytokine relative to experimental high-dose IL-2 (refs 14,35). Taken together, these results suggest a broad correlation between the magnitude of immune cell expansion and the development of treatment-associated toxicities, with low-dose IL-2/mAb or IL-23XFc treatments displaying higher levels of toxicity, particularly pulmonary oedema, compared to IL-2WTFc.
We next evaluated the therapeutic efcacy of IL-2/mAb, IL-2WTFc and IL-23XFc in the B16F10 melanoma model. For this purpose, we utilized a dosing regime consisting of ve consecutive daily intraperitoneal (i.p.) injections of 1 mg antibody-complexed
IL-2 or IL-2-Fc molar equivalent starting 1 day after subcutaneous injection of tumour cells. Strikingly, treatment with IL-2WTFc resulted in a substantial reduction of tumour growth compared to either IL-2/mAb immune complex or the IL-23XFc triple mutant (Fig. 2f). This apparent superior therapeutic efcacy was observed despite the lower potential of IL-2WTFc to induce expansion of cytotoxic immune cell subsets (Fig. 1c).
IL-2WTFc targets Tregs for FccR-mediated depletion. The observation that both IL-2/mAb immune complexes and the IL-23XFc triple mutant displayed greater toxicity and less efcient
NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 3
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373
a
b
c
NS** **
NS
*
110
200
1,500
NS
NS
Water in lungs (mg)
150
100
1,000
% Initial body weight
100
Liver weight (mg)
90
500
50
80
0
PBS
IL-2/mAb
IL-2WT Fc
IL-23X Fc
PBS
IL-2/mAb
IL-2WT Fc
IL-23X Fc
PBS
IL-2/mAb
IL-2WT Fc
IL-23X Fc
d
NS
e
NS
f
125
NS
NS
**
NS
75
180
100
ALT activity (IU l1 )
AST activity (IU l1 )
Tumour area (mm2 )
60
135
75
****
45
90
50
30
****
**
45
25
15
0
0
PBS
0 0 4 8 12 16 20
IL-2/mAb
IL-2WT Fc
IL-23X Fc
PBS
IL-2/mAb
IL-2WT Fc
IL-23X Fc
Days post tumour injection
Figure 2 | Toxicity prole and antitumour activity of IL-2-Fc variants. (ae) Symptoms of experimental VLS were assessed in mice receiving ve consecutive doses (days 04) of IL-2/mAb (1 mg IL-2 5 mg mAb per dose), IL-2WTFc (5.6 mg per dose), IL-23XFc (5.6 mg per dose) or PBS control. Lungs,
livers and blood were collected on day 6 for analysis (n 4 mice per group). (a) Body weight on day 6, represented as percentage of initial body weight on
day 0. (b) Pulmonary oedema was assessed by measurement of lung water content, with signicantly higher accumulation of uid observed in the IL-2/mAb (P 0.0042) and IL-23XFc (P 0.0018) groups, compared to PBS controls. (ce) Assessment of liver toxicity as measured by total liver
weight (c) and serum levels of liver enzymes alanine aminotransferase (ALT, d) and aspartate aminotransferase (AST, e). (f) Tumour growth after subcutaneous inoculation of B16F10 melanoma cells into the anks of mice treated with ve consecutive doses of IL-2/mAb (1 mg IL-2 5 mg mAb per
dose), IL-2WTFc (5.6 mg per dose), IL-23XFc (5.6 mg per dose) or PBS control on days 15 (n 6). Data are displayed as means.e.m. Asterisks indicate
signicant differences between specied groups (*Po0.05, **Po0.01, ****Po0.0001) as determined by one-way analysis of variance (ANOVA) (ae) or two-way ANOVA (f) with Bonferroni post hoc test for multiple comparisons.
protection against tumour growth than wild-type IL-2-Fc-fusion protein was unexpected and led to further investigation. We focused our attention on the role of the Fc part of the molecule, and in particular its interaction with FcgRs. For this purpose, we investigated mutations that abolish antibody-dependent cell-mediated cytotoxicity/phagocytosis (ADCC/ADCP). More specically, we mutated conserved residues in the Fc region of IL-2-Fc mediating binding to FcgRs (L234A, L235E and
G237A)36,37. Accordingly, we assessed the activity of IL-2WTFcnil (no effector functions) and IL-2WTFc (normal FcgR binding but disrupted C1q interaction) constructs (Fig. 3a; Supplementary Fig. 7A). Since the Fc region utilized in this study was decient for binding to the C1q complement component, we also generated a mutant (IL-2WTFcC1q ) in which C1q binding was restored through mutation (see Methods). Disruption of the
FcgR and/or C1q binding sites was validated through macrophage and C1q binding assays, respectively (Supplementary Fig. 7C,D). Importantly, mutation of the Fc region did not affect interaction with the CD25high CTLL-2 cell line (Supplementary Fig. 7B), indicating that the IL-2 component of IL-2-Fc remained intact. This allowed us to assign any differences in IL-2-Fc in vivo activity to the mutations introduced into the Fc component.
In vivo assessments in C57BL/6 mice receiving multiple low-dose IL-2-Fc treatment revealed similar increases in spleen lymphoid cellularity for all designed variants (Fig. 3b). Notably, the effector-less IL-2WTFcnil fusion protein readily expanded not only CD122high MP CD8 and NK cells but also CD4 CD25
Tregs (Fig. 3c,d). By contrast, treatment with IL-2WTFc and IL-2WTFcC1q led to comparable increases in MP CD8 and
NK cells, but no increase in the numbers or percentages of Tregs (Fig. 3c,d). In view of the similar levels of MP CD8 and NK cell expansion across all constructs and the lack of Treg expansion in the presence of Fc-mediated effector functions, we concluded that IL-2WTFc and IL-2WTFcC1q were able to selectively deplete
Tregs. We should emphasize, however, that this depletive effect was quite limited relative to PBS controls (Fig. 3e,f) and only became prominent when compared with the marked Treg expansion induced by IL-2WTFcnil. Hence the Fc-mediated depletion seemed to be largely restricted to IL-2-activated Tregs. Furthermore, Treg depletion was predominantly FcgR-mediated rather than C1q-mediated since the ability to bind C1q (IL-2WTFcC1q ) did not result in increased elimination of Tregs (Fig. 3c,d). Notably, a large proportion of Tregs in mice treated with IL-2WTFc or IL-2WTFcC1q displayed high levels of the cell proliferation factor Ki-67 (Supplementary Fig. 7E). This nding
4 NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373 ARTICLE
a
b
***
2.0108
Splenocyte number
*******
1.5108
1.0108
IL-2WTFcnil IL-2WTFc IL-2WTFcC1q+
IL-2WTFcnil IL-2WTFcC1q+
5.0107
0.0
IL-2WT Fc
IL-2WT FcC1q+
c
d
PBS
IL-2WT Fcnil
MP CD8 NK cells Tregs
7.5106
45
MP CD8 (% of CD8+)
NK cells (% of CD3)
Tregs (% of CD4+)
***
35
Cell number
Percent of parent
population
5.0106
**
** ****
****
*** **
** *** ***
****
25
2.5106
***
15
NS NS
** NS NS
NS
5
0.0
PBS
IL-2WT Fcnil
IL-2WT Fc
IL-2WT FcC1q+
PBS
IL-2WT Fcnil
IL-2WT Fc
IL-2WT FcC1q+
e f
P = 0.03
PBS
7.54 32.4 6.43
IL-2WTFc
8
P = 0.06
6.68
****
CD4+ FoxP3+ cells
(% of lymphocytes)
6
FoxP3 FoxP3
CD4
4
CD25
2
1.92 5.51 0.82
0.96
0 PBS
IL-2WT Fcnil
IL-2WT FcIL-2WT FcC1q+
Figure 3 | IL-2WTFc selectively depletes Tregs in a FccR-dependent manner. (a) Diagram of produced IL-2-Fc variants showing FcgR and C1q-binding site status. Mutated residues within the Fc region are illustrated in Supplementary Fig. 7A. (bf) Analysis of spleens (day 5) collected from C57BL/6 mice treated with ve consecutive 5.6 mg i.p. injections of IL-2WTFcnil, IL-2WTFc, IL-2WTFcC1q or PBS control on days 04 (n 4). (b) Total live splenocytes
counts showing increased lymphoid cellularity after IL-2-Fc treatment. (cf) Flow cytometric analysis of collected splenocytes. (c) Total cell numbers of MP CD8 (CD8 CD44high CD122high), NK cell (CD3 NK1.1 CD122high) and Treg (CD4 FoxP3 CD25) cell subsets. (d) Frequencies of MP CD8 (shown as proportion of CD8 ), NK cells (proportion of CD3 ) and Tregs (proportion of CD4 ). (e) Frequencies of CD4 FoxP3 cells within the lymphocyte compartment showing depletion of this subset after treatment with FcgR-binding IL-2-Fc constructs. (f) Representative ow cytometry dot plots displaying the frequency of regulatory T-cells in treated mice as dened by the co-expression of CD4, FoxP3 and the IL-2-inducible CD25 surface marker (top row, shown as proportion of CD4 ) or by expression of CD4 and FoxP3 (bottom row, shown as proportion of total lymphocytes). Data are displayed as means.e.m. Asterisks indicate signicant differences relative to PBS controls (c,d) or between specied groups (b,e) as determined by oneway analysis of variance with Bonferroni post hoc test for multiple comparisons (**Po0.01, ***Po0.001, ****Po0.0001).
supports the notion that these FcgR-binding constructs did stimulate Tregs to divide but also eliminated a large proportion of these cells via Fc-mediated killing, resulting in little or no change in Treg numbers.
We further investigated the observed preferential depletion of Tregs by assessing the interaction of IL-2-Fc with different immune cell subsets. Evaluation in an ex vivo binding assay revealed that IL-2WTFc preferentially bound CD4 CD25
Tregs over the CD25 MP CD8 or NK cell subsets (Fig. 4a). To explore differential targeting of lymphocyte subsets in vivo, we compared the cellular biodistribution proles of uorescently labelled IL-2WTFc and IL-2WTFcnil fusion proteins (Fig. 4b). We found that IL-2-Fc proteins targeted a low proportion of total CD8 T-cells, which is consistent with only a small fraction of
this compartment expressing high levels of the dimeric IL-2Rbgc (that is, MP CD8 cells). Similarly, B12% of total CD4 T-cells bound to IL-2-Fc proteins, in agreement with typical proportions of Tregs expressing the trimeric IL-2Rabgc. Accordingly, IL-2-Fc fusions were found to efciently target CD25 FoxP3 Tregs (485%, Fig. 4b), despite reductions in IL-2-Fc uorescence intensity after sample xation for FoxP3 immunostaining (Supplementary Fig. 5C). We observed that labelled IL-2-Fc proteins, regardless of their ability to bind FcgR, targeted B95% of NK cells, thus suggesting that these constructs bind to this subset predominantly through the IL-2R (IL-2Rbgc) rather than via FcgR. By contrast, disruption of FcgR binding substantially reduced the proportion of macrophages and neutrophils associated with IL-2WTFcnil (Fig. 4c), suggesting that
NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 5
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373
a
b
2,400
MFI (IL-2-Fc)
1,600
75
0
IL-2WTFc
IL-2WTFcnil
CD8 T-cells
Tregs
CD4 T-cells
NKT cells
B-cells
50
800
%IL-2-Fc+
100
25
NK cells
DCs
Macrophages
Neutrophils
0 1 0 1 2 3
Log10 [conc(nM)]
c d
100
NS
IL-2WTFc IL-2WTFcnil
CD4+ T-cells
NK cells
MFI: 2414
MFI: 11356
80
**
100
% IL-2-Fc+
60
80
Max (%)
60
40
****
40
20
20
0
0
102 103 104 105
NK cells
Macrophages
Neutrophils
101
IL-2-Fc
Tregs
e
f g
IL-2WTFc IL-2WTFcnil
15,000
15,000
***
Tregs Tregs
MFI (IL-2-Fc)
MFI (IL-2-Fc)
12,000
MFI (IL-2-Fc)
10,000
10,000
9,000
***
****
6,000
5,000
5,000
3,000
****
0
0
0 NK cells
CD4+ IL-2-Fc+
(Tregs)
CD4 IL-2-Fc+
CD8 IL-2-Fc+
NK IL-2-Fc+
NKT IL-2-Fc+
CD4 IL-2-Fc+
CD8 IL-2-Fc+
NK IL-2-Fc+
NKT IL-2-Fc+
Figure 4 | Depletive IL-2WTFc activity relies on high-afnity targeting of Tregs and interaction with myeloid effector subsets. (a) Ex vivo ow cytometric detection of labelled IL-2WTFc on the surface of MP CD8, NK cell and Treg subsets after incubation with Fc-blocked FoxP3DTR/GFP splenocytes (n 2 technical replicates). (bg) Cellular biodistribution proles of uorescently labelled IL-2WTFc and IL-2WTFcnil in the spleens of treated C56BL/6 mice
as determined by ow cytometry 12 h post injection (16.8 mg IL-2-Fc i.p., n 23 mice per group). (b) Heat-map representation of the percentages of
lymphoid and myeloid subsets bound by IL-2-Fc-fusion proteins (c) Abolition of FcgR binding causes a signicant reduction in the percentages of
macrophages (P 0.0041) and neutrophils (Po0.0001) bound by uorescent IL-2-Fc, as determined by two-tailed unpaired Students t-test.
(d) Representative histograms displaying the levels of IL-2WTFcnil present on CD4 T-cells and NK cells after i.p. injection. Box indicates that the majority of CD4 IL-2-Fc cells are Tregs, as previously shown in Supplementary Fig. 5C. (e) Quantication of d, showing signicantly higher IL-2-Fc MFI values in
IL-2-Fc CD4 T-cells (boxed cells in d) relative to IL-2-Fc NK cells (n 2 mice, two-tailed unpaired Students t-test, P 0.0004). (f,g) Injected
IL-2WTFc (f) and IL-2WTFcnil (g) proteins accumulate to higher levels on the surface of Tregs compared to any other analysed subset. Asterisks indicate signicant differences relative to CD4 IL-2-Fc Tregs, as determined by one-way analysis of variance with Bonferroni post hoc test for multiple comparisons (**Po0.01, ***Po0.001, ****Po0.0001). All data are displayed as means.e.m. MFI, mean uorescence intensity.
these myeloid subsets may function as effector cells in FcgR-dependent Treg depletion. Finally, similar frequencies of
NKT cells (B50%), dendritic cells (B3%) and B-cells (B1%) were found to be targeted by both IL-2WTFc and IL-2WTFcnil proteins (Fig. 4b).
Having examined the frequencies of IL-2-Fc cells, we next compared the intensity of IL-2-Fc uorescent signals, specically on the NK cell and Treg subsets. To exclude any potential for FcgR-mediated binding on NK cells, uorescently labelled
IL-2WTFcnil was used for this comparison. Interestingly, the mean uorescence intensity levels observed on IL-2-Fc CD4
T-cells, of which the vast majority are Tregs (see Supplementary Fig. 5C), were nearly vefold higher compared to IL-2-Fc
NK cells (Fig. 4d,e). Likewise, uorescently labelled IL-2WTFc and IL-2WTFcnil proteins were both observed at higher amounts on Tregs than on CD8 T-cells and NKT cells in addition to
NK cells (Fig. 4f,g).
Collectively, our ex vivo (Fig. 4a) and in vivo (Fig. 4cg) analyses demonstrate substantially stronger binding of IL-2-Fc proteins to high-afnity IL-2Rabgc on Tregs than to CD25
subsets expressing intermediate-afnity IL-2Rbgc, thus explaining
the selective opsonization of Tregs by IL-2WTFc constructs.
IL-2-Fc antitumour effects require CD25 and FccR interaction. To further investigate the inuence of Fc-mediated effector functions on antitumour effects, we compared the efcacy of IL-2WT and IL-23X proteins expressed as fusions to Fc (able to bind FcgR) or Fcnil (abolished effector functions). As previously observed (Fig. 2f), treatment with IL-2WTFc resulted in superior antitumour activity against B16F10 melanoma in comparison to IL-23XFc (Fig. 5a), despite the latter variant mediating considerably higher peripheral expansion of cytotoxic subsets (Fig. 1b,c). Remarkably, the efcacy of IL-2WTFc was critically
6 NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373 ARTICLE
a
b
Proportion survival
150
1.0
0.8
125
0.6
*
**
**
PBS IL-2 Fc IL-2 Fc IL-2 Fc IL-2 Fc
Spleen dLN Tumour
0.4
Tumour area (mm2 )
0.2
50
100
30 NS **
100 NS
NS
0.0 0 15 30 45 60
NS NS
****
****
40
75
75
20
120
% initial body
weight
50
Tregs (% of CD4+ )
Day 7
**
30
50
110
20
10
25
100
***
PBS IL-2 Fc IL-2 Fc IL-2 Fc IL-2 Fc
10
25
90
0
80 0 4 8 12 16
0 PBS Fc Fcnil
0 PBS Fc Fcnil
0 PBS Fc Fcnil
0 4 8 12 16 20
Days post tumour injection
Days post tumour injection
c Tumour
0 PBS Fc Fcnil
d
e
f
g
100
****
2.0107
CD45.2+ per g of tumour
6106
CD8+ per g of tumour
NS
45
10.0
CD8 : Tregs (tumour)
NS
***
NS
CD8+ T-cells
(% of tumour CD45.2+ )
****
**
NS
*
NS
CTLA-4+ Ki-67high
(% of Tregs)
75
1.5107
7.5
4106
30
50
1.0107
5.0
2106
15
25
5.0106
2.5
0.0 PBS Fc Fcnil
0 PBS Fc Fcnil
0 PBS Fc Fc
0.0 PBS Fc Fcnil
Figure 5 | Potent antitumour activity of IL-2WTFc is dependent on both FccR and CD25 binding. (a) The antitumour activity of IL-2WTand IL-23X fused to either Fc or Fcnil was assessed in the B16F10 melanoma tumour model (n 6). C57BL/6 mice were injected subcutaneously (s.c.) in their left anks with
1 105 B16F10 cells (day 0) and treated with ve consecutive doses of IL-2-Fc variants (5.6 mg per dose, i.p.) on days 15. Following treatment, mice were
monitored for tumour growth (left), survival (top right) and body weight (bottom right). (bg) Flow cytometric analysis of spleens, dLNs (inguinal) and B16F10 tumours collected from mice receiving IL-2WTFc, IL-2WTFcnil or PBS control treatment (n 5). Following s.c. tumour inoculation (day 0), mice
received a total of ten doses of IL-2-Fc on days 15 and days 1418 (5.6 mg per dose, i.p.), followed by analysis 48 h after the last dose (day 20).(b) Frequency of FoxP3 cells (percentage of CD4) in the spleen, dLNs and tumours. (c) Frequency of CTLA-4 Ki-67high intratumoural Tregs (percentage of CD4 FoxP3 cells). (de) Numbers of inltrating CD45.2 leukocytes (d) and CD8 T-cells (e) per gram of tumour. (f) Frequency of tumour-inltrating CD8 T-cells, shown as percentage of CD45.2 cells. (g) Intratumoural ratio of CD8 to regulatory T-cells, as calculated from their relative proportions in the CD45.2 compartment. dLNs, draining lymph nodes; one mouse in the IL-2WTFc group did not develop a tumour. Data are displayed as means.e.m. Asterisks indicate signicant differences between specied groups (*Po0.05, **Po0.01, ***Po0.001, ****Po0.0001) as determined by one-way analysis of variance (ANOVA) (bg) or two-way ANOVA (a) with Bonferroni post hoc test for multiple comparisons. Survival (a, top right) is displayed using KaplanMeier plots and compared by the GehanBreslowWilcoxon test.
dependent on FcgR binding, since treatment with IL-2WTFcnil resulted in no detectable differences in tumour growth or survival compared to PBS controls (Fig. 5a). Furthermore, as observed in non-tumour bearing mice (Fig. 2a), treatment with the IL-23XFc variant caused notable reductions in body weight (Fig. 5a, bottom right). This effect was accentuated in mice receiving IL-23XFcnil treatment, possibly due to the elevated levels of fusion protein detected in serum compared to IL-23XFc (Supplementary Fig. 8).
Flow cytometric analyses of spleen, draining lymph node (dLN) and B16F10 tumour tissue were performed in order to gain insights into mechanisms underpinning the efcacy of IL-2WTFc treatment. For this purpose, we compared the levels of lymphocyte expansion in mice treated with either IL-2WTFc or IL-2WTFcnil relative to PBS control treatment 48 h after the last IL-2-Fc dose (Fig. 5bg). In agreement with our previous results (Fig. 3), we found that IL-2WTFcnil treatment readily expanded splenic Tregs, while IL-2WTFc failed to expand this subset (Fig. 5b, left). While these results reinforce the previously observed Treg-depletive activity of IL-2WTFc in the spleen (Fig. 3), this effect was less pronounced in dLN (Fig. 5b, middle) and not evident in tumour lesions (Fig. 5b, right) at this time point. Notably, however, ow cytometric analysis performed 24 h after the last administered dose of fusion protein revealed that IL-2WTFc is indeed able to mediate Treg depletion in the spleen, dLN and, crucially, at the tumour site itself (Supplementary Fig. 9A,B).
We observed an increased proportion of CTLA-4 Ki-67high Tregs in the tumours of mice treated with IL-2WTFcnil but not
with IL-2WTFc (Fig. 5c). Interestingly, this difference was tumour-specic and was not observed in the spleen or dLN (Supplementary Fig. 9C). Thus, the reduced frequency of highly activated CTLA-4 Ki-67high Tregs in the tumours of
IL-2WTFc-treated mice may provide a basis for improved antitumour responses. In line with this observation, B16F10 tumours collected from mice receiving IL-2WTFc treatment displayed a pronounced increase in the numbers of inltrating leukocytes (Fig. 5d) with a clear enrichment of cytotoxic CD8
T-cells, both in terms of numbers (Fig. 5e) and frequency (Fig. 5f). This enrichment translated into an increased intratumoural ratio of CD8 to Tregs in the IL-2WTFc treatment group only (Fig. 5g). In support of this observation, antibody-mediated depletion of CD8 T-cells severely compromised
IL-2WTFc antitumour activity (Supplementary Fig. 9D), while depletion of total CD4 T-cells improved treatment efcacy (Supplementary Fig. 9E), presumably by further depletion of
CD4 CD25 Tregs.
We next investigated the potential of the IL-2-Fc-fusion proteins for therapy in the B16F10 model in combination with tumour-targeting antibodies30. For this purpose IL-2/mAb, IL-2WTFc or IL-23XFc treatment was administered in combination with a monoclonal antibody targeting the B16F10 tumour antigen tyrosinase-related protein 1 (TRP-1) (Fig. 6a). In this setting, IL-2WTFc again provided the largest reduction in tumour growth and was the only treatment to signicantly improve survival compared to anti-TRP-1 monotherapy (GehanBreslowWilcoxon test, P 0.0446).
NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 7
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373
a
B16F10
PBS TRP-1 TRP-1 +
IL-2/mAb
TRP-1 +
IL-2WTFc
TRP-1 +
IL-23XFc
*
Proportion survival
PBSTRP-1TRP-1 + IL-2/mAb TRP-1 + IL-2WTFc
TRP-1 + IL-23XFc
1.0
175
0.8
*
0.6
NS
150
0.4
NS
*
Tumour area (mm2 )
125
0.2
0.0 0 15 30 45 60
100
75
120
% initial body
weight
NS ****
PBSTRP-1TRP-1 + IL-2/mAb TRP-1 + IL-2WTFc
TRP-1 + IL-23XFc
110
50
100
25
90
0 0 4 8 12 16 20 24
80 0 4 8 12 16 20 24
Days post tumour injection
Days post tumour injection
b
CT26
100
Tumour area (mm2 )
NS
175
Tumour area (mm2 )
175
IL-2WTFc
75
125
125
50
Rejected
tumours (2/5)
75
75
****
PBS IL-2/mAb IL-2WTFc
IL-23XFc
IL-2WTFcnil
25
25
25
0 0 8 16 24 32
0 8 16 24 32 0 8 16 24 32
Days post tumour injection
Days post tumour injection
Figure 6 | IL-2WTFc treatment synergizes with targeted antibody therapy and is highly efcacious against murine colorectal carcinoma. (a) The antitumour activity of IL-2/mAb, IL-2WTFc and IL-23XFc in combination with the anti-TRP-1 tumour-targeting monoclonal antibody was assessed in the B16F10 melanoma tumour model. C57BL/6 mice were injected subcutaneously (s.c.) in their left anks with 1 105 B16F10 cells on day 0 (n 6).
Combination therapy consisted of three doses of anti-TRP-1 (200 mg per dose given i.p. on days 3, 6 and 9) plus low-dose IL-2/mAb (0.5 mg IL-2 2.5 mg
mAb per dose) or molar IL-2-Fc equivalent (2.8 mg per dose) given i.p. on days 4, 7 and 10. Following treatment, mice were monitored for tumour growth (left), body weight (bottom right) and survival (top right). (b) The antitumour activity of IL-2/mAb, IL-2WTFc, IL-23XFc and IL-2WTFcnil was assessed in the
CT26 murine colorectal carcinoma tumour model. Balb/c mice were injected s.c. in their left anks with 1 105 CT26 cells on day 0 (n 5), followed by
treatment with a total of nine doses of IL-2/mAb (1 mg IL-2 5 mg mAb per dose), IL-2-Fc molar equivalent (5.6 mg per dose) or PBS control given i.p. on
days 15 and days 13, 15, 17 and 19. Mean tumour size (left) and tumour growth in individual mice in the PBS and IL-2WTFc groups (right) are shown. Data are displayed as means.e.m. Asterisks indicate signicant differences between specied groups (*Po0.05, ****Po0.0001) as determined by two-way analysis of variance with Bonferroni post hoc test for multiple comparisons. Survival (a, top right) is displayed using KaplanMeier plots and compared by the GehanBreslowWilcoxon test.
In addition to B16F10 melanoma, the efcacy of IL-2-Fc-fusion proteins was also investigated in the CT26 colorectal carcinoma syngeneic tumour model (Fig. 6b). Results in this model closely resembled what had been observed for B16F10, with IL-2WTFc providing substantial reductions in tumour growth, while all other treatments displayed no detectable effects. Remarkably, some of the mice treated with IL-2WTFc displayed complete rejection of subcutaneous tumours (Fig. 6b, right), further highlighting the potential of IL-2WTFc therapy.
DiscussionConsidered the rst effective cancer immunotherapy, high-dose IL-2 is able to mediate durable responses in a small subset of metastatic melanoma and metastatic renal cancer patients6. However, IL-2 suffers from sub-optimal therapeutic properties arising from a short serum half-life and pleiotropic biological actions, which lead to toxicity and the unwanted expansion of immunosuppressive Tregs.
Here we have generated highly selective IL-2-Fc-fusion proteins through introduction of CD25-disrupting mutations into the cytokine component. Using a combination of structural design,
in vitro biosensor measurements and in vivo screening, we developed a novel IL-2-Fc triple mutant (IL-23XFc). Disruption of CD25 binding resulted in prolongation of serum half-life, large increases in biological activity and preferential expansion of CD25 CD122high
MP CD8 and NK cells. Notably, IL-23XFc displayed enhanced activity and selectivity compared to not only wild-type IL-2-Fc, but also previously reported IL-2/mAb complexes.
Further analyses revealed that the antitumour effects and toxicity prole of the IL-23XFc triple mutant were comparable to that of IL-2/mAb complexes, indicating that our initial design objectives had been achieved. However, intriguingly, these experiments also revealed that both IL-23XFc and IL-2/mAb were in fact less effective than the parental IL-2WTFc protein in the B16F10 melanoma model. Furthermore, unlike IL-23XFc and IL-2/mAb treatment, IL-2WTFc did not induce pulmonary oedema or reductions in body weight. This indicated that toxicities arising from these low-dose treatments might be largely immune-related, likely due to excessive activation of CD8
T-cells and NK cells (although residual effects on the lung endothelium could not be excluded)13.
Subsequent analyses revealed that selective Treg depletion was critically dependent on the high afnity of IL-2WTFc towards
8 NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373 ARTICLE
CD25 cells, as well as the interaction of the Fc region with FcgR. Crucially, the potent antitumour activity displayed by IL-2WTFc was also dependent on these two properties, with mutation of either the CD25- or FcgR-binding sites resulting in reduced treatment efcacy in the B16F10 melanoma and CT26 colorectal carcinoma models. Analysis of B16F10 tumours revealed reduced frequencies of CTLA-4 Ki-67high Tregs after
IL-2WTFc treatment, as well as a large increase in total leukocyte inltration dominated by cytotoxic CD8 T-cells. Our results therefore suggest that IL-2WTFc administration contributes to the establishment of a highly immunogenic tumour microenvironment that favours the inltration of immune cells in general, and of CD8 T-cells in particular, leading to enhanced antitumour responses. This is consistent with previously reported increases in
CD8 T-cell tumour inltration and subsequent tumour rejection following in vivo Treg depletion in transgenic mouse models38. The data are also in agreement with the nding that, in the context of immune checkpoint blockade, effective antitumour immunity correlates with a high ratio of CD8 T-cells to Tregs in tumour-inltrating lymphocytes39,40.
In addition to the depletion of Tregs, IL-2WTFc treatment also resulted in expansion of splenic MP CD8 and NK cells, although to a lesser extent than what was observed for IL-23XFc or IL-2/mAb complexes. The provision of dual activities, which can be either depleting or proliferative depending on the cellular subset, differentiates IL-2WTFc from existing therapeutic modalities, such as the anti-CD25 monoclonal antibody daclizumab (Zenapax)41. This property may also provide advantages over IL-2-toxin fusions, which target Tregs for depletion but also mediate collateral elimination of CD8 T-cells and NK cells42,43. Furthermore, pre-administration of IL-2WTFc could be used to improve the tumour-targeting properties of IL-2 immunocytokines, as recently reported by Wittrup and colleagues29. Finally, based on our ndings, ultra-low doses of the highly active and selective IL-23XFc variant may prove effective in combination with Treg-depleting antibodies, a strategy that we are currently evaluating.
Taken together, our results on the engineering of IL-2-Fc-fusion proteins outline an effective strategy to enhance the therapeutic properties of this key immunomodulatory cytokine. Recent clinical successes of immune checkpoint inhibitors highlight the potential of novel immune modulating agents in cancer immunotherapy, either as stand-alone therapeutics or in combination with anti-PD-1 and anti-CTLA-4 therapy44. The IL-2WTFc fusion protein described here displays favourable properties, as characterized by a long serum half-life, low levels of toxicity, increased antitumour activity and the ability to selectively deplete Tregs. The role of Treg depletion in the IL-2 system had so far remained elusive: indeed, IL-2/mAb complexes have been reported to act independently of Fc-mediated effector functions14 and previous reports of IL-2-Fc-fusion proteins have failed to identify any evidence of Treg depletion25,2729. By contrast, such a mechanism has been apparent for other targets and therapeutic modalities. In particular, FcgR-mediated Treg depletion has been recently described as a critical component in the antitumour activity of antibodies targeting CTLA-4, OX40 and GITR4547.
In summary, here we describe the development and characterization of IL-2-Fc-fusion proteins that exceed the specicity (IL-23XFc) and efcacy (IL-2WTFc) of IL-2/mAb complexes. Moreover, we identify the depletion of Tregs, rather than the expansion of cytotoxic CD8 T-cells and NK cells, as a major determinant for antitumour activity, providing important guidance for the future development of IL-2 reagents for cancer immunotherapy.
Methods
Mutagenesis and production of recombinant proteins. For the generation of IL-2-Fc-fusion proteins, regions encoding human IL-2 (residues Ala1Thr133)
were genetically fused to murine Fc (IgG2c Glu216Lys447, Eu antibody numbering48) containing a mutated C1q-binding site (E318A, K320A and K322A)49 by means of a short glycine-serine linker (GSGS). The construct was generated by gene synthesis (GeneArt) and cloned into the mammalian expression vector pCEP4 (Thermo Fisher). Disruption of the CD25 and FcgR binding sites and restoration of the C1q binding site were performed using the Q5 site-directed mutagenesis kit (NEB) using custom-designed primers, followed by validation of mutations by Sanger sequencing. IL-2-Fc constructs were used to transfect suspension-adapted HEK293 cells using the Expi293 expression system (Thermo Fisher). Protein purication was performed with protein G agarose beads (ACROBiosystems) using disposable columns (Thermo Fisher). All protein preparations were quality controlled for endotoxin levels, as measured by chromogenic LAL assay (Lonza). Genetic constructs coding for the ectodomains of hCD25 and hCD122 (in pCEP4, C-terminal His-tagged) were purchasedfrom Genscript, expressed as above and puried using the TALON metal afnity resin (Clontech).
Mice. C57BL/6 and BALB/c female mice were purchased from the Animal Resources Centre (Canning Vale, WA, Australia) and used at 810 weeks of age. IL-7 transgenic mice (IL-7 Tg: B6 background, Thy-1.1-congenic) and transgenic mice expressing DTR-GFP under the FoxP3 promoter50 (C57BL/6 background) were bred at the Australian BioResources facilities (Moss Vale, NSW, Australia). Animals were housed under conventional barrier protection and handled in accordance with protocols approved by the Garvan Institute of Medical Research and St Vincents Hospital Animal Experimentation and Ethics Committee, which comply with the Australian code of practice for the care and use of animals for scientic purposes.
Flow cytometry and antibodies. Flow cytometric analysis of mouse spleens and lymph nodes was performed according to standard protocols. The following antibodies were used for staining (all purchased from eBioscience unless stated otherwise): PE- or eFluor450-conjugated anti-CD8a (clone 53-6.7, used at1.25 mg ml 1); FITC- or eFluor450-conjugated anti-CD122 (clone TM-b1,1 mg ml 1); APC-conjugated anti-CD44 (clone IM7, 0.4 mg ml 1); FITC- or eFuor450-conjugated anti-CD45.2 (clone 104, 2.5 mg ml 1); PE-Cy7- or eFluor450-conjugated anti-CD3e (clone145-2C11, 1 mg ml 1); PE- or
APC-conjugated anti-NK1.1 (clone PK136, 1 mg ml 1); BV421- or BV605-conjugated anti-CD4 (clone GK1.5, 1 mg ml 1, BioLegend); PerCP/Cy5.5-conjugated anti-Thy-1.1 (clone HIS51, 0.5 mg ml 1); PerCP/Cy5.5- or eFluor450-conjugated anti-CD25 (clone PC61.5, 1 mg ml 1); PerCP-conjugated anti-B220 (clone RA3-6B2, 1 mg ml 1, BioLegend); PE/Cy7-conjugated anti-Ly6G (clone 1A8, 1 mg ml 1, BD); PE-conjugated anti-CD11b (clone M1/70, 1 mg ml 1);
PerCP-conjugated anti-CD11c (clone N418, 1 mg ml 1, BioLegend); FITC-labelled anti-F4/80 (clone BM8, 5 mg ml 1, BioLegend); PE-, APC- or eFluor450-conjugated anti-FoxP3 (clone FJK-16 s, 5 mg ml 1); FITC- or PE-conjugated anti-Ki-67 (clone SolA15, 0.4 mg ml 1) and APC-conjugated anti-CTLA-4 (clone UC10-4F10-11, 1 mg ml 1, BD). Intracellular staining with anti-FoxP3, anti-Ki-67 and anti-CTLA-4 antibodies was performed after xation/permeabilization with the FoxP3 buffer set (eBioscience).
Cytokine treatments. Mice received recombinant hIL-2, IL-2/mAb or IL-2-Fc treatments at specic quantities and dosing schedules, as described in the gure legends. Recombinant hIL-2 was obtained from Peprotech. IL-2/mAb complexes were prepared by mixing hIL-2 and the anti-hIL-2 MAB602 (mouse IgG2a, clone 5355, R&D systems) at a 2:1 molar ratio (for example, 1 mg IL-2 5 mg mAb),
followed by incubation at 37 C for 25 min prior to injection. Puried IL-2-Fc variants were stored at 80 C, thawed, ltered (0.22 mm) and re-assessed for
protein concentration prior to injection.
Assessment of treatment-associated toxicity. Mice were injected i.p. once per day with PBS, 1 mg IL-2 5 mg mAb, 5.6 mg IL-2WTFc or 5.6 mg IL-23XFc on days
04. On day 6, mice were killed for collection of blood (cardiac puncture), lungs, livers and spleens. Blood samples were allowed to clot for 2 h at room temperature (RT), followed by separation of serum by centrifugation. Aspartate aminotransferase and alanine aminotransferase activity assays (Teco Diagnostics) were performed on non-hemolyzed serum samples as per manufacturers instructions. Tissue wet weights were recorded using an analytical balance. To determine lung water content, lungs were dehydrated overnight at 42 C using a SpeedVac instrument (Savant) and the difference between lung wet weight and lung dry weight was calculated.
Assessment of IL-2-Fc binding to mouse splenocytes ex vivo. Red blood cell-depleted splenocytes from a FoxP3-DTR/GFP transgenic mouse were incubated with Fc-Block (anti-mouse CD16/CD32, BD), followed by washing in ow cytometry buffer and staining with uorophore-conjugated anti-CD3e, anti-CD8a, anti-CD4, anti-CD44 and anti-NK1.1 antibodies. Stained cells were washed, seeded in a 96-well plate (1 106 cells per well) and re-suspended in
serially diluted IL-2WTFc-biotin. After incubation for 30 min on ice, cells were
NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 9
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373
stained with APC-conjugated streptavidin (eBioscience). The levels of IL-2WTFc bound to the surface of MP CD8 (CD3 CD8 CD44high), NK cells(CD3 NK1.1 ) and Tregs (CD3 CD4 FoxP3 ) were determined by ow cytometry.
IL-2-Fc uorescent labelling and cellular biodistribution. Puried IL-2-Fc variants in PBS were ltered through a syringe-driven 0.22 mm lter prior to labelling with Alexa Fluor 647-NHS (Thermo Fisher). IL-2-Fc concentration was adjusted to 300 mg ml 1 in 500 ml, followed by addition of 50 ml 1 M sodium bicarbonate to increase pH. Alexa Fluor 647-NHS was added at a 15:1 molar ratio and labelling was allowed to take place for 2 h at RT with gentle rotation in the dark. After the labelling reaction, IL-2-Fc variants were buffer-exchanged into PBS using two sequential de-salting steps with Zeba spin columns (Thermo Fisher). Final protein concentration and labelling efciency were measured using a Nanodrop instrument (Thermo Fisher). Fluorescently labelled IL-2-Fc variants were administered to mice via i.p. as a single 16.8 mg dose and spleens were collected 12 h post injection. Flow cytometric analysis of red blood cell-depleted splenocytes was performed after staining with antibodies conjugated to uorophores with different emission spectra to that of Alexa Fluor 647. The frequencies and mean uorescence intensity levels of IL-2-Fc cells were determined for the following populations: CD8 T-cells (CD3 CD8 ), CD4 T-cells (CD3 CD4 ), Tregs (CD4 FoxP3 CD25 ), NK cells (CD3 NK1.1 ),
NKT cells (CD3 NK1.1 ), B-cells (CD3 B220 ), dendritic cells (CD3 CD11chigh), macrophages (CD3 CD11bmid-high CD11clow-mid F4/80 SSClow)
and neutrophils (CD3 CD11bmid-high CD11clow-mid Ly6G ).
Tumour models. B16F10 melanoma cells (ATCC CRL-6475 Lot 60508145) were cultured at 37 C, 5% CO2 in high-glucose DMEM containing 10% FBS, 2 mM L-glutamine, 50 U ml 1 penicillin and 50 mg ml 1 streptomycin. B16F10 cells (7080% conuent) were reconstituted in PBS and injected subcutaneously into the left anks of C57BL/6 mice (1 105 cells per mouse). CT26 colorectal carcinoma
cells (ATCC CRL-2638 Lot 63226308) were cultured in ATCC-formulated RPMI 1640 supplemented with 10% FBS, 50 U ml 1 penicillin and 50 mg ml 1 streptomycin. CT26 cells (7080% conuent) were reconstituted in PBS and injected subcutaneously into the left anks of BALB/c mice (1 105 cells per
mouse). After tumour injection, mice were treated with specic IL-2/mAb, IL-2-Fc and anti-TRP-1 (clone TA99, BioXCell) dosing schedules, as described in the gure legends. Following treatment, tumour area was monitored every 23 days using callipers. Mice were killed if showing considerable weight loss (420% of initial body weight), displayed obvious signs of systemic illness or if tumours grew larger than 144 mm2. Mycoplasma-tested B16F10 and CT26 cancer cell lines were purchased from the American Type Culture Collection (ATCC), expanded in culture once and aliquoted for storage in liquid nitrogen. Aliquots were then thawed, expanded for use in a single tumour model experiment and discarded.
Flow cytometric analysis of B16F10 tumours. B16F10 tumours were dissected and carefully separated from skin tissue. Collected tumours were weighted and collected into serum-free RPMI 1640 (Thermo Fisher) containing 2 mg ml 1 collagenase D (Roche) plus 0.1 mg ml 1 DNAse I (grade I, Roche) and digested for 1 h at 37 C. Tissue was then dispersed through 70 mm cell strainers (BD) to obtain single cell suspensions. The volume of cell suspension utilized for immunostaining was recorded for each sample (volume equivalent to B20 mg), followed by aliquoting into a 96-well plate and addition of 5 104 counting beads (FITC
Calibrite, BD) per well. Samples were then incubated with Fc-Block (anti-mouse CD16/CD32, BD), surface antibody stain and xable viability dye (eFluor780, eBioscience) in that order. This was followed by xation/permeabilization (FoxP3 buffer set, eBioscience) and incubation with intracellular stain. Samples were then analysed by ow cytometry, as described above. Total numbers of immune subsets per gram of tumour were calculated utilizing the following formula: [(no. of acquired cells no. of acquired beads) no. of added beads]
[(volume stained total volume) tumour weight in grams].
Afnity measurements. Biolayer interferometry (BLI) measurements were performed using the BLItz instrument (ForteBio). hCD25-Fc (R&D systems) was biotinylated with EZ-Link NHS-PEG4-Biotin (Thermo Fisher) and loaded onto streptavidin biosensors. Data were obtained using IL-2-Fc variants at 100 mg ml 1 (1,190 nM), with 120 s association and 600 s dissociation times. Surface plasmon resonance measurements were performed using the Biacore 2000 system. Recombinant hCD25-His, hCD122-His or mCD25 (R&D systems) were covalently immobilized onto CM5 sensor chips (GE) to B200 response units using the amine coupling kit (GE). Binding of IL-2-Fc variants to hCD25, hCD122 or mCD25 was recorded using a 30 ml min 1 ow rate, with 60 s of association and 300 s of dissociation followed by regeneration of immobilized ligands in 0.1 M glycine,0.1 M NaCl pH 3. Curve tting of hCD25 and mCD25 kinetic data was performed using the BIAevaluation software. One-armed IL-2WTFc was produced in order to validate the curve tting strategy used for bivalent IL-2-Fc variants. Briey, IL-2WTFc and Fc-His constructs (both in pCEP4 expression vector) were co-transfected into Expi293 cells (Thermo Fisher), followed by purication of the
monovalent IL-2WTFc/Fc-His pairing with a HisTrap column (GE) using an imidazole gradient elution (AKTA protein purication system, GE).
Assessment of binding and activity on CTLL-2 cells. The murine CTLL-2 cell line (IL-2-dependent, CD25high) was cultured in complete RPMI 1640(10% FBS, 2 mM L-glutamine, 50 U ml 1 penicillin, 50 mg ml 1 streptomycin,1 mM sodium pyruvate, 10 mM HEPES, 55 mM 2-mercaptoethanol) supplemented with 1 ng ml 1 hIL-2 (Peprotech). CTLL-2 proliferation was measured by incorporation of radiolabelled thymidine. Briey, 5 103 cells were
cultured in the presence of IL-2-Fc for 24 h, followed by addition of 3H thymidine and culturing for a further 24 h. Cells were collected onto a glass bre lter and radioactivity was quantied using a liquid scintillation counter (Perkin Elmer). Prior to binding and signalling assays, CTLL-2 cells were starved of IL-2 for 6 h in order to remove surface-bound cytokine. Binding of IL-2-Fc variants to CTLL-2 cells (2 105 cells per well) was allowed for 20 min on ice. Unbound IL-2-Fc was
removed by washing in ow cytometry buffer (2% FBS, 2 mM EDTA in PBS) and surface-bound IL-2-Fc was detected by ow cytometry using a FITC-conjugated anti-mIgG2a antibody (clone R19-15, BD). Signalling in CTLL-2 cells was measured by ow cytometric detection of pSTAT5 after IL-2-Fc stimulation. Cells (2 105 per well) were stimulated with IL-2-Fc variants for 10 min at 37 C,
followed by xation (10 min at 37 C in BD Cytox), permeabilization (30 min on ice in BD Phosow Perm Buffer III) and staining with an AF488-conjugated anti-pSTAT5 antibody (clone 47/Stat5-pY694, BD) for 45 min at RT.
Ex vivo stimulation of mouse splenocytes and human PBMC. Red blood cell-depleted splenocytes from an IL-7 transgenic mouse were incubated with Fc-Block (anti-mouse CD16/CD32, BD) for 20 min on ice, followed by washing in ow cytometry buffer and staining with uorophore-conjugated anti-CD8a and anti-CD44 antibodies. Stained cells were seeded in a 96-well plate (1 106 cells per
well) and re-suspended in serially diluted IL-2-Fc variants. Stimulation was allowed for 10 min at 37 C, followed by washing with ow cytometry buffer, xation, permeabilization and intracellular staining with AF488-conjugated anti-pSTAT5 (clone 47/Stat5-pY694, BD). The levels of pSTAT5 in the CD8 CD44high population were determined by ow cytometry. For ow cytometry of human
PBMC, the following antibodies were used: Pacic Blue-conjugated anti-CD8 (clone RPA-T8, BD), APC-conjugated anti-CD4 (clone S3.5, Thermo Fisher), PE-conjugated anti-FoxP3 (clone 259D/C7, BD). Stimulation of PBMC was allowed for 10 min at 37 C in the presence of 1 mg ml 1 (Treg stain) or10 mg ml 1 (CD8 T-cell stain) IL-2-Fc. Detection of intracellular pSTAT5 was performed as described above. Buffy coats from normal donors were obtained from the Australian Red Cross Blood Service. Informed consent was obtained from all subjects and approval for this study was obtained from the human research ethics committee of the St Vincents Hospital (Sydney, Australia).
Serum ELISA. Mice were injected with a single dose (i.p.) of 20 mg hIL-2, 3 mg hIL-2 15 mg mAb or 16.8 mg IL-2-Fc variants. Blood samples were taken
at specied time points via tail vein bleeding and serum was separated by centrifugation after resting for 2 h at RT. hIL-2 was detected using an anti-hIL-2 mAb (clone 5344.111, BD) for capture and an anti-hIL-2 biotinylated polyclonal antibody (BAF202, R&D systems) for detection. Detection of IL-2/mAb and IL-2-Fc variants was performed using an anti-mIgG2a mAb (clone R11-89, BD) for capture and BAF202 for detection. Blocking and dilutions were performedin 1% bovine serum albumin in PBS and individual standard curves were built to determine the serum levels of each variant.
Adoptive transfer of MP CD8 cells. The MP CD8 reporter assay was performed as previously described7. MP CD8 cells from IL-7 Tg mice were uorescence-activated cell sorting-sorted and labelled with carboxyuorescein succinimidyl ester (CFSE) for adoptive transfer into C57BL/6 recipients. After cytokine treatment, donor MP CD8 cells in the spleens of recipient mice were identied by ow cytometry as Thy-1.1 and their proliferation measured by CFSE dilution.
Fluorescence microscopy. Spleen tissue from mice treated with Alexa Fluor 647-labelled IL-2-Fc was collected 12 h post injection and frozen in optimal cutting temperature (OCT) embedding medium (Sakura). Tissue sections (5 mm) were cut using a Leica CM3050 S Cryostat, followed by xation in acetone (7 min) and rehydration in PBS. After blocking with Protein Block solution (Dako), sections were stained with anti-B220-biotin, PE-conjugated anti-FoxP3 and BV421-conjugated anti-CD4 (in Antibody Diluent solution, Dako), followed by washing in PBS 1%
Tween 20 and staining with AF488-conjugated streptavidin (Thermo Fisher). Sections were mounted with Fluoromount (Sigma) and analysed using a Leica DM5500 microscope.
Binding to macrophages ex vivo. Red blood cell-depleted splenocytes from C57BL/6 mice were stained with anti-CD11b-PE, anti-F4/80-FITC and anti-CD3e-PE-Cy7. Stained cells were aliquoted into 96-well plates (1 106 cells per well),
washed in ow cytometry buffer and incubated with 10 mg ml 1 IL-2-Fc. After
10 NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373 ARTICLE
IL-2-Fc binding, cells were washed in ow cytometry buffer and stained with2 mg ml 1 anti-human IL-2-biotin (clone 5344.111, BD). Finally, cells were stained with streptavidin-APC (eBioscience) prior to ow cytometric analysis.
Complement deposition assay. CTLL-2 cells were collected 23 days after passaging and surface-bound hIL-2 was stripped by washing in RPMI 1640,2% FBS, pH 3 for 20 s. Since complement deposition is calcium-dependent, all steps were performed in EDTA-free buffers. Accordingly, washing and incubation steps were performed in high-glucose DMEM supplemented with 0.5% bovine serum albumin and 0.08% sodium azide. IL-2-stripped CTLL-2 cells were aliquoted into 96-well plates (3.5 105 cells per well) and re-suspended in 42 mg ml 1 IL-2-Fc
(500 nM). After IL-2-Fc binding, cells were washed in media and re-suspended in puried mouse complement (Cedarlane Laboratories) diluted 1:2. Complement deposition was allowed for 2 h on ice, followed by washing and staining with anti-mIgG2a-AF488 and anti-mC1q-biotin (clone RmC7H8, Cedarlane Laboratories). Cells were then washed and stained with streptavidin-APC (eBioscience). Finally, cells were re-suspended in HEPES-buffered saline supplemented with 2% FBS and 1 mM CaCl2 for ow cytometric analysis.
Statistical analysis. Data are displayed as means.e.m., (*Po0.05, **Po0.01, ***Po0.001, ****Po0.0001). Statistical analyses included one-way and two-way analysis of variance with Bonferroni post hoc test for multiple comparisons, and two-tailed unpaired Students t-tests in data sets comparing two groups. Survival was displayed using KaplanMeier plots and compared by the GehanBreslowWilcoxon test. Data were analysed using the Prism software (GraphPad).
Data availability. The data supporting the ndings of this study are available within the article and its Supplementary Information les and from the corresponding authors on reasonable request.
References
1. Boyman, O. & Sprent, J. The role of interleukin-2 during homeostasis and activation of the immune system. Nat. Rev. Immunol. 12, 180190 (2012).
2. Morgan, D. A., Ruscetti, F. W. & Gallo, R. Selective in vitro growth of T-lymphocytes from normal human bone marrows. Science 193, 10071008 (1976).
3. Devos, R. et al. Molecular-cloning of human interleukin-2 carrier DNA and its expression in Escherichia coli. Nucleic Acids Res. 11, 43074323 (1983).4. Taniguchi, T. et al. Structure and expression of a cloned cDNA for human interleukin-2. Nature 302, 305310 (1983).
5. Lotze, M. T. et al. High-dose recombinant interleukin 2 in the treatment of patients with disseminated cancer: responses, treatment-related morbidity, and histologic ndings. JAMA 256, 31173124 (1986).
6. Rosenberg, S. A. IL-2: the rst effective immunotherapy for human cancer.J. Immunol. 192, 54515458 (2014).7. Taniguchi, T. & Minami, Y. The IL-2IL-2 receptor system: a current overview. Cell 73, 58 (1993).
8. Robb, R. J., Greene, W. C. & Rusk, C. M. Low and high afnity cellular receptors for interleukin 2. Implications for the level of Tac antigen. J. Exp. Med. 160, 11261146 (1984).
9. Sakaguchi, S., Yamaguchi, T., Nomura, T. & Ono, M. Regulatory T cells and immune tolerance. Cell 133, 775787 (2008).
10. Lotze, M. T. et al. In vivo administration of puried human interleukin 2. II. Half life, immunologic effects, and expansion of peripheral lymphoid cellsin vivo with recombinant IL 2. J. Immunol. 135, 28652875 (1985).
11. Yang, J. C. et al. Randomized study of high-dose and low-dose interleukin-2 in patients with metastatic renal cancer. J. Clin. Oncol. 21, 31273132 (2003).
12. Boyman, O., Kovar, M., Rubinstein, M. P., Surh, C. D. & Sprent, J. Selective stimulation of T cell subsets with antibody-cytokine immune complexes. Science 311, 19241927 (2006).
13. Krieg, C., Letourneau, S., Pantaleo, G. & Boyman, O. Improved IL-2 immunotherapy by selective stimulation of IL-2 receptors on lymphocytes and endothelial cells. Proc. Natl Acad. Sci. USA 107, 1190611911 (2010).
14. Letourneau, S. et al. IL-2/anti-IL-2 antibody complexes show strong biological activity by avoiding interaction with IL-2 receptor alpha subunit CD25. Proc. Natl Acad. Sci. USA 107, 21712176 (2010).
15. Carmenate, T. et al. Human IL-2 mutein with higher antitumor efcacy than wild type IL-2. J. Immunol. 190, 62306238 (2013).
16. Liu, D. V., Maier, L. M., Haer, D. A. & Wittrup, K. D. Engineered interleukin-2 antagonists for the inhibition of regulatory T cells. J. Immunother. 32, 887894 (2009).
17. Rao, B. M., Girvin, A. T., Ciardelli, T., Lauffenburger, D. A. & Wittrup, K. D. Interleukin-2 mutants with enhanced alpha-receptor subunit binding afnity. Protein Eng. 16, 10811087 (2003).
18. Levin, A. M. et al. Exploiting a natural conformational switch to engineer an interleukin-2 superkine. Nature 484, 529533 (2012).
19. Mitra, S. et al. Interleukin-2 activity can be ne tuned with engineered receptor signaling clamps. Immunity 42, 826 (2015).
20. Boyman, O., Ramsey, C., Kim, D. M., Sprent, J. & Surh, C. D. IL-7/anti-IL-7 mAb complexes restore T cell development and induce homeostatic T cell expansion without lymphopenia. J. Immunol. 180, 72657275 (2008).21. Rubinstein, M. P. et al. Converting IL-15 to a superagonist by binding to soluble IL-15R alpha. Proc. Natl Acad. Sci. USA 103, 91669171 (2006).
22. Weinblatt, M. E. et al. A trial of etanercept, a recombinant tumor necrosis factor receptor: Fc fusion protein, in patients with rheumatoid arthritis receiving methotrexate. N. Engl. J. Med. 340, 253259 (1999).
23. Vazquez-Lombardi, R. et al. Challenges and opportunities for non-antibody scaffold drugs. Drug Discov. Today 20, 12711283 (2015).
24. Jindal, R. et al. Spontaneous resolution of acute rejection and tolerance induction with IL-2 fusion protein in vascularized composite allotransplantation. Am. J. Transplant. 15, 12311240 (2015).
25. Zheng, X. X. et al. Favorably tipping the balance between cytopathic and regulatory T cells to create transplantation tolerance. Immunity 19, 503514 (2003).
26. Millington, T. et al. Effects of an agonist interleukin-2/Fc fusion protein,a mutant antagonist interleukin-15/Fc fusion protein, and sirolimus on cardiac allograft survival in non-human primates. J. Heart Lung Transplant. 31, 427435 (2012).
27. Bell, C. J. et al. Sustained in vivo signaling by long-lived IL-2 induces prolonged increases of regulatory T cells. J. Autoimmun. 56, 6680 (2015).
28. Zheng, X. X. et al. IL-2 receptor-targeted cytolytic IL-2/Fc fusion protein treatment blocks diabetogenic autoimmunity in nonobese diabetic mice.J. Immunol. 163, 40414048 (1999).29. Tzeng, A., Kwan, B. H., Opel, C. F., Navaratna, T. & Wittrup, K. D. Antigen specicity can be irrelevant to immunocytokine efcacy and biodistribution. Proc. Natl Acad. Sci. USA 112, 33203325 (2015).
30. Zhu, E. F. et al. Synergistic innate and adaptive immune response to combination immunotherapy with anti-tumor antigen antibodies and extended serum half-life IL-2. Cancer Cell 27, 489501 (2015).
31. Kontermann, R. E. Strategies for extended serum half-life of protein therapeutics. Curr. Opin. Biotechnol. 22, 868876 (2011).
32. Wang, X., Rickert, M. & Garcia, K. C. Structure of the quaternary complex of interleukin-2 with its a, b, and gc receptors. Science 310,11591163 (2005).
33. Heaton, K. M., Ju, G. & Grimm, E. A. Human interleukin 2 analogues that preferentially bind the intermediate-afnity interleukin 2 receptor lead to reduced secondary cytokine secretion: implications for the use of these interleukin 2 analogues in cancer immunotherapy. Cancer Res. 53, 25972602 (1993).
34. Weir, M. P., Chaplin, M. A., Wallace, D. M., Dykes, C. W. & Hobden, A. N. Structure activity relationships of recombinant human interleukin-2. Biochemistry 27, 68836892 (1988).
35. Rosenstein, M., Ettinghausen, S. E. & Rosenberg, S. A. Extravasation of intravascular uid mediated by the systemic administration of recombinant interleukin 2. J. Immunol. 137, 17351742 (1986).
36. Sondermann, P., Huber, R., Oosthuizen, V. & Jacob, U. The 3.2- crystal structure of the human IgG1 Fc fragmentFcgRIII complex. Nature 406, 267273 (2000).
37. Wines, B. D., Powell, M. S., Parren, P. W., Barnes, N. & Hogarth, P. M. The IgG Fc contains distinct Fc receptor (FcR) binding sites: the leukocyte receptors FcgRI and FcgRIIa bind to a region in the Fc distinct from that recognized by neonatal FcR and protein A. J. Immunol. 164, 53135318 (2000).
38. Li, X., Kostareli, E., Suffner, J., Garbi, N. & Hammerling, G. J. Efcient Treg depletion induces T-cell inltration and rejection of large tumors. Eur. J. Immunol. 40, 33253335 (2010).
39. Curran, M. A., Montalvo, W., Yagita, H. & Allison, J. P. PD-1 and CTLA-4 combination blockade expands inltrating T cells and reduces regulatory T and myeloid cells within B16 melanoma tumors. Proc. Natl Acad. Sci. USA 107, 42754280 (2010).
40. Twyman-Saint Victor, C. et al. Radiation and dual checkpoint blockade activate non-redundant immune mechanisms in cancer. Nature 520, 373377 (2015).
41. Rech, A. J. et al. CD25 blockade depletes and selectively reprograms regulatory T cells in concert with immunotherapy in cancer patients. Sci. Transl. Med. 4, 134ra62 (2012).
42. Rasku, M. A. et al. Transient T cell depletion causes regression of melanoma metastases. J. Transl. Med. 6, 12 (2008).
43. Yamada, Y. et al. Differential effects of denileukin diftitox IL-2 immunotoxin on NK and regulatory T cells in nonhuman primates. J. Immunol. 188, 60636070 (2012).
44. Ott, P. A., Hodi, F. S. & Robert, C. CTLA-4 and PD-1/PD-L1 blockade: new immunotherapeutic modalities with durable clinical benet in melanoma patients. Clin. Cancer Res. 19, 53005309 (2013).
45. Bulliard, Y. et al. Activating Fc g receptors contribute to the antitumor activities of immunoregulatory receptor-targeting antibodies. J. Exp. Med. 210, 16851693 (2013).
NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications 11
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15373
46. Bulliard, Y. et al. OX40 engagement depletes intratumoral Tregs via activating FcgRs, leading to antitumor efcacy. Immunol. Cell Biol. 92, 475480 (2014).
47. Simpson, T. R. et al. Fc-dependent depletion of tumor-inltrating regulatory T cells co-denes the efcacy of antiCTLA-4 therapy against melanoma. J. Exp. Med. 210, 16951710 (2013).
48. Edelman, G. M. et al. The covalent structure of an entire gG immunoglobulin molecule. Proc. Natl Acad. Sci. USA 63, 7885 (1969).
49. Duncan, A. R. & Winter, G. The binding site for C1q on IgG. Nature 332, 738740 (1988).
50. Kim, J. M., Rasmussen, J. P. & Rudensky, A. Y. Regulatory T cells prevent catastrophic autoimmunity throughout the lifespan of mice. Nat. Immunol. 8, 191197 (2007).
Acknowledgements
This work was supported by the National Health and Medical Research Council. We thank Professor Alexander Rudensky (Howard Hughes Medical Institute and Memorial Sloan-Kettering Cancer Center, New York) for supplying the Foxp3DTR/GFP mice.
Author contributions
D.C. and J.S. conceived and supervised the study. R.V.-L. designed and performed most of the experiments. T.G.P., K.E.W., J.S. and D.C. contributed to experimental design. C.L., D.Z., J.J., P.S., E.K.D. and C.K. performed or contributed to specic experiments. R.V.-L., J.S. and D.C. wrote the manuscript.
Additional information
Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications
Web End =http://www.nature.com/ http://www.nature.com/naturecommunications
Web End =naturecommunications
Competing interests: The authors declare no competing nancial interests.
Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/
Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/
Web End =reprintsandpermissions/
How to cite this article: Vazquez-Lombardi, R. et al. Potent antitumour activity of interleukin-2-Fc fusion proteins requires Fc-mediated depletion of regulatory T-cells. Nat. Commun. 8, 15373 doi: 10.1038/ncomms15373 (2017).
Publishers note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
Web End =http://creativecommons.org/licenses/by/4.0/
r The Author(s) 2017
12 NATURE COMMUNICATIONS | 8:15373 | DOI: 10.1038/ncomms15373 | http://www.nature.com/naturecommunications
Web End =www.nature.com/naturecommunications
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Copyright Nature Publishing Group May 2017
Abstract
Interleukin-2 (IL-2) is an established therapeutic agent used for cancer immunotherapy. Since treatment efficacy is mediated by CD8+ and NK cell activity at the tumour site, considerable efforts have focused on generating variants that expand these subsets systemically, as exemplified by IL-2/antibody complexes and 'superkines'. Here we describe a novel determinant of antitumour activity using fusion proteins consisting of IL-2 and the antibody fragment crystallizable (Fc) region. Generation of long-lived IL-2-Fc variants in which CD25 binding is abolished through mutation effectively prevents unwanted activation of CD25+ regulatory T-cells (Tregs) and results in strong expansion of CD25- cytotoxic subsets. Surprisingly, however, such variants are less effective than wild-type IL-2-Fc in mediating tumour rejection. Instead, we report that efficacy is crucially dependent on depletion of Tregs through Fc-mediated immune effector functions. Our results underpin an unexpected mechanism of action and provide important guidance for the development of next generation IL-2 therapeutics.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer