Human IL-2 Mutein with Higher Antitumor Efficacy … Journal of Immunology Human IL-2 Mutein with Higher Antitumor Efficacy Than Wild Type IL-2 Tania Carmenate,* Anabel Pacios,*

of May 16, 2018.This information is current as

Efficacy Than Wild Type IL-2Human IL-2 Mutein with Higher Antitumor

Kalet LeónErnesto Moreno, Karina Garcia-Martínez, Dasha Fuente and Tania Carmenate, Anabel Pacios, Michel Enamorado,

ol.1201895http://www.jimmunol.org/content/early/2013/05/15/jimmun

published online 15 May 2013J Immunol 

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2013 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Human IL-2 Mutein with Higher Antitumor Efficacy ThanWild Type IL-2

Tania Carmenate,* Anabel Pacios,* Michel Enamorado,* Ernesto Moreno,*

Karina Garcia-Martınez,* Dasha Fuente,† and Kalet Leon*

IL-2 has been used for the treatment of melanoma and renal cell carcinoma, but this therapy has limited efficacy and severe

toxicity. Currently, it is assumed that part of the limited efficacy is due to the IL-2–driven preferential expansion of regulatory

T cells, which dampen the antitumor immunity. In this study, we characterize a human IL-2 mutant with higher antitumor

efficacy and lower toxicity than wild type human IL-2 (wtIL-2). The mutant differs from wtIL-2 by four mutations at the interface

with the a subunit of IL-2R. The IL-2 mutant induces in vitro proliferation of CD8+CD44hi and NK1.1 cells as efficiently as does

wtIL-2, but it shows a reduced capacity to induce proliferation of CD4+Foxp3+ regulatory T cells. The IL-2 mutant shows a higher

antimetastatic effect than does wtIL-2 in several transplantable tumor models: the experimental metastasis model of MB16F0

melanoma and the experimental and spontaneous metastasis models for the mouse pulmonary carcinoma 3LL-D1222. Relevantly,

the IL-2 mutant also exhibits lower lung and liver toxicity than does wtIL-2 when used at high doses in mice. In silico simulations,

using a calibrated mathematical model, predict that the properties of IL-2 mutein are a consequence of the reduction, of at least

two orders of magnitude, in its affinity for the a subunit of IL-2R (CD25). The human IL-2 mutant described in the present work

could be a good candidate for improving cancer therapy based on IL-2. The Journal of Immunology, 2013, 190: 000–000.

Interleukin-2 is a member of the family of common g-chaincytokines, which have central roles in T lymphocyte gener-ation, activation, and homeostasis (1). IL-2 is a 15-kDa

glycoprotein produced primarily by activated CD4+ and CD8+

T lymphocytes. The high-affinity receptor for IL-2 is composedof three subunits: IL-2Ra (CD25), IL-2Rb (CD122), and g-chain(CD132) (2). This trimeric form is expressed constitutively onCD4+Foxp3+ regulatory T cells (Tregs) and on activated CD4+ andCD8+ T cells. In contrast, the intermediate-affinity form composedof b-and common g-chains is also functional and highly expressed inmemory CD8+ T cells and NK cells.For many years, IL-2 was a cytokine central for protective im-

munity because of the demonstration of its potent capacity to induceT cell growth in vitro (3, 4). Since the 1980s, the role of IL-2 inT cell expansion and activation encouraged its use in melanomaand renal cell carcinoma treatments. Several clinical trials showedthe limitations of this therapy; although some patients experiencedcomplete or long-lasting responses, only 15–20% of treated patientsreceived some clinical benefit (5). Moreover, the high doses re-quired to obtain such results induce high toxicity, with vascular leaksyndrome being the most frequent and severe complication (6).IL-2, together with other common g-chain cytokines, plays a

central role in the generation and maintenance of the T lymphocyteresponse. It was reported that the strength and duration of the IL-2signal controls both the primary and secondary expansion of theCD8+ T cell population (7). In addition, IL-2 regulates Th1, Th2,

and Th17 differentiation based on its modulation of the expressionof IL-4R, IL-12R, or IL-6R (8). The in vivo function of IL-2 wasfirst questioned after the characterization of IL-2–deficient mice.Instead of the expected immune-deficient phenotype, such mice areable to develop a normal antiviral response (9); moreover, afterseveral weeks they develop an ulcerative colitis–like disease, and50% of them die (10).The description and characterization of Tregs by Sakaguchi

et al. (11) and, later, the demonstration of the crucial role forIL-2 in the generation and homeostasis of these cells definitivelychanged the opinion about the primary function of IL-2 in vivo(12). Now, it is assumed that part of the limitation in efficacy ofIL-2 therapy in cancer patients is due to the IL-2–driven expansionof Tregs, which, in turn, dampens antitumor immunity (13). In ad-dition, an increased number of Tregs has been found in many cancerpatients; in some cases, such as melanoma and ovarian cancer, highnumbers of Tregs correlate with a poor prognosis (14). Manyattempts have been made to improve IL-2–based therapy, and sev-eral of them were based on developing IL-2 variants with modifiedproperties (15, 16). In this article, we present the preclinical eval-uation of a human IL-2 (hIL-2) mutant that causes a preferentialexpansion of CD8+ and NK cells over Tregs. The mutant behaves asan IL-2 agonist, suitable for cancer treatment with higher efficacyand lower toxicity than wild type hIL-2 (wtIL-2).

Materials and MethodsCell lines and culture conditions

The murine T cell line CTLL2 was donated by Dr. A. Santos (Center ofGenetic Engineering and Biotechnology, La Habana, Cuba). This line isIL-2 dependent and constitutively expresses the abg form of IL-2R. Thecells were cultured in RPMI 1640 (Life Technologies) supplemented with10% heat-inactivated FBS, 50 IU/ml human rIL-2 (Center of GeneticEngineering and Biotechnology), 2 mM L-glutamine, 50 U/ml penicillin,and 50 mg/ml streptomycin. Murine MB16F0 melanoma cells and murine3LL-D122 lung carcinoma cells were maintained in DMEM F12 (LifeTechnologies) supplemented with 10% heat-inactivated FBS, 2 mM L-glutamine, 50 U/ml penicillin, and 50 mg/ml streptomycin. All cells weremaintained at 37�C under a humidified 5% CO2 atmosphere. Tumor cells

*Systems Biology Department, Center of Molecular Immunology, Havana, CP11600, Cuba; and †National Center for Laboratory Animal Breeding, Havana, CP10800, Cuba

Received for publication August 16, 2012. Accepted for publication April 4, 2013.

Address correspondence and reprint requests to Tania Carmenate, Center of Molec-ular Immunology, Havana, CP 11600, Cuba. E-mail address: [email protected]

Abbreviations used in this article: E, helper; hIL-2, human IL-2; LN, lymph node;M, memory; R, regulatory; TE, 10 mM Tris, 1 mM EDTA; Treg, regulatory T cell;wtIL-2, wild type human IL-2.

Copyright� 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1201895

Published May 15, 2013, doi:10.4049/jimmunol.1201895 by guest on M

ay 16, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

were harvested using trypsin/EDTA; for in vivo experiments, cells wereresuspended in PBS.

Design and production of hIL-2 mutein

For production of the hIL-2 mutein, the synthetic genes containing thecorresponding mutations from the original hIL-2 gene, as well as the fulloriginal hIL-2 gene, were obtaining from Geneart (Berlin, Germany). Thegenes were cloned into the commercial vector pET28a (Novagen, Darm-stadt, Germany). This vector contains an N-terminal His-tag, and the genesof interest are under the control of lac operator. Escherichia coli cells[strain BL21(DE3); Invitrogen] were transformed with the hIL-2 mutant orwtIL-2 expression plasmids using the manufacturer’s protocol. Trans-formed cells were allowed to grow in 200 ml Lysogenic broth mediumuntil the A620nm reached 0.6; subsequently, 4 mM IPTG was added, andincubation continued for 6 h, after which cells were harvested by centri-fugation and stored as a pellet at 280�C.

Purification of wtIL-2 and hIL-2 mutein from insolublematerial

The frozen pellets were suspended in 10 mMTris, 1 mM EDTA (pH 8) (TE)and sonicated using an ultrasonic cell disrupter (IKA). In each case, theinsoluble material was harvested by centrifugation (18,000 3 g) andwashed successively with 4 M Urea-TE and 1% Triton X-100–TE using anUltra Turrax T8 homogenizer. For further purification, we used an HPLCsystem (Pharmacia, Uppsala, Sweden); the proteins were extracted with6 M Guanidinium hydrochloride–TE at 0.1 g/ml (wet weight), and rena-turation was carried out by dialysis. For further purification, proteins wereapplied to a reverse-phase C4 column (Vydac). In this final chromatography,the recombinant polypeptides were purified using an H2O-acetonitrile-trifluoroacetic acid system, with a linear gradient (30–85% of acetonitrile)and 0.6 ml/min flow. Finally, the proteins were dialyzed against 10 mMacetate (pH 4), filtered through 0.2 mm filters, and stored at 4�C.

Abs and flow cytometry

For flow cytometry, cell suspensions from spleen and lymph nodes (LNs)were prepared according to standard protocols. All fluorochrome-conjugatedmAbs used were from eBioscience, unless otherwise stated; FITC-conjugatedanti-CD3 (145-2C11), PECy5.5-conjugated anti-CD4(L3T4), PE-conjugatedanti-NK1.1 (PK136), PE-conjugated anti-CD8 (eBio H35-17.2), PECy5.5-conjugated anti-B220(RA3-6B2), PE-conjugated anti-Foxp3(NRRF-30), PE-conjugated anti-CD25 (3C7), and PE-conjugated anti–His-tag were fromR&D Systems. Intracellular Foxp3 staining sets were purchased fromeBioscience. Samples were measured using a FACScan (Becton Dickinson)flow cytometer and analyzed using FlowJo software (TreeStar). Anti-CD8(YTS169) and anti-NK1.1 (PK136), both produced at the Center of Mo-lecular Immunology, were mAbs used for in vivo Ab-depletion experiments.

Mice

Seven- to eight-week-old female C57BL/6 and BALB/c mice were obtainedfrom the National Center for Laboratory Animal Breeding (Havana, Cuba).Food and water were provided ad libitum. The experiments were performedaccording to guidelines of the International Laboratory Animals Resourcesusing standardized procedures in the Center of Molecular Immunology. Allanimal studies were conducted under a protocol approved by the Institu-tional Animal Care and Use Committee.

Cell suspension preparation

CD8+ T cells and CD4+FOXP3+ Tregs were purified from inguinal andmesenteric LN cell suspensions using CD8+ T Cell and CD4+CD25+

Regulatory T Cell Isolation kits (Miltenyi Biotec), following the manu-facturer’s recommendation. In both cases, purity was tested by flow cytom-etry. In the case of NK cell expansion from spleen cells, spleens from twomice were collected, disaggregated, and RBC lysed.

Proliferation assays

CTLL2 cells were harvested by centrifugation, washed two times, andincubated in RPMI 1640 without IL-2 for 5 h prior to the proliferationassay. A total of 104 cells/well was incubated with serial dilutions of eitherwtIL-2 or hIL-2 mutein in RPMI 1640 and allowed to grow for 48 h. Afterthat, 20 ml alamarBlue dye (Invitrogen) was added per well, and plateswere incubated for 12 h. Finally, plates were read at 540 and 630 nm, andthe percentage of reduced alamarBlue was calculated following the man-ufacturer’s recommendation.

For lymphocyte populations, 105 cells/well purified CD8+ cells or CD4+

CD25+ Tregs were cultivated on 96-well plates. In the case of CD4+CD25+

Tregs, 96-well plates were precoated with 5 mg/ml anti CD3 mAb 2C11,and cells were cultivated in the presence of 100 mg/ml anti mIL-2 mAbS45B6. For the NK1.1+ cell expansion assay, 23 106 spleen cells/wellwere cultivated on 24-well plates. In all cases, cells were incubated withdifferent concentrations of wtIL-2 or hIL-2 mutein; cells were harvestedafter 5 d and counted by flow cytometry. For determination of thenumber of cells reference fluorospheres Flow-Check (Beckman Coulter)were used.

Lung metastasis models

For experimental metastasis models, C57BL/6 mice were inoculated with1 3 105 B16F0 or 3LL-D122 cells via the tail vein on day 0. Treatmentswere given on days 1–4 by i.p. injections of saline solution, 20 mg wtIL-2,or 20 mg IL-2 variant, twice a day. When necessary, 1 mg depleting mAbsspecific for CD8 or NK1.1 molecules was injected i.p. into mice) on days1, 4, and 7. For the spontaneous metastasis model, C57BL/6 mice wereinoculated with 2 3 105 3LL-D122 cells in the footpad; when primarytumors reached 0.9 mm in diameter, they were removed surgically.Treatments were given on days 1–4 after challenge and after surgery byi.p. injection of saline solution, 20 mg wtIL-2, or 20 mg hIL-2 variant.Mice were sacrificed on day 21, lungs were removed and embedded inBouin’s solution, and metastatic nodules were counted under a binocularmicroscope.

Toxicity

For toxicity assessment, five BALB/c mice/group were inoculated with 80mg wtIL-2 or hIL-2 mutein, whereas the control group received PBS. Thetreatments were administered twice a day for 5 d. Mice were sacrificed,and lungs and livers were weighed. For histologic study, organs were fixedin 10% formalin solution, and paraffin-embedded sections were stainedwith H&E.

Statistical analysis

We used Graph Pad Prism 4.0 software for statistical analysis. In vitroproliferation curves were adjusted to sigmoid dose-response curves. Inthe case of antitumor assays, a parametric ANOVA, followed by a Bon-ferroni multiple-comparison test, was applied. For toxicity assessment,organ weights were compared using the nonparametric Kruskal–Wallis test,followed by the Dunn multiple-comparison test.

Mathematical model of the interplay between IL-2 and IL-2muteins with T lymphocytes

The mathematical model used in this study was developed and calibratedto describe the interaction between IL-2 and Th cells (E), regulatory (R)CD4+ T cells, and memory CD8+ T cells (17). The model includes severalcompartments, which represent different LNs, in which T cells are con-fined interacting with each other, with the APCs and available solublemolecules. It also includes a compartment representing the blood (i.e., thecirculatory system), which contains only soluble molecules, IL-2, or IL-2mutants. LNs are connected to the blood compartment, allowing the freeexchange of soluble molecules. The concentration of IL-2 and muteins inthe blood is assumed to decay with a constant rate, which represents renalelimination. An external source term for these molecules is added in thiscompartment to simulate particular treatment applications. The modelincludes the dynamics of helper (E), regulatory (R), and memory (M)T cells on the different functional states of their life cycle: resting, acti-vated, and cycling cells.

No a IL-2 mutein is modeled as a soluble molecule that bears all ofthe properties of IL-2 in the model but whose conjugation affinity for thea-chain of IL-2R is reduced by the factor f (parameter f = {0 to 1}).Treatments are simulated to represent a continuous infusion of this mol-ecule in the blood for a defined period of time. This is implemented bysetting on, transiently, the external source terms in the blood compartmentof this molecule. Two parameters always control treatment application: the‘‘dose,’’ which set up the total amount per day of IL-2 mutein infused inthe blood, and the ‘‘treatment duration,’’ which set the length of timefor which continuous infusion is maintained. We explore how the doseand treatment duration determine the outcome of the system simula-tion, as well as whether this treatment can condition a significantpreferential expansion (dominance) of Th cells, Tregs, or M cells in the LNcompartment.

For calibration of the mathematical model, the majority of the parameterswere fixed to values taken directly or derived from available independentexperimental data; just a few parameters remain unknown, and their in-fluence on the result was explored within a range of biologically reasonablevalues.

2 IL-2 MUTEIN

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

ResultsMutein production

An hIL-2 mutein was designed that differs from wtIL-2 only atpositions R38, F42, Y45, and E62; all of these residues weresubstituted by alanine. For production convenience, Cys125 wassubstituted by Ser, because this mutation was shown not to changethe biological activity of hIL-2 (18). The synthetic genes encodingeither the mutein or wtIL-2 were cloned into the pET28a vectorand expressed in the BL21 (DE3) E. coli strain. Proteins wereexpressed as inclusion bodies and were purified to homogeneityfollowing the purification protocol described by Moya et al. (19)for human rIL-2 purification.

IL-2 mutant behaves as an IL-2 agonist in vitro

Fig. 1 shows the molecular structure of hIL-2 mutein with thepositions of mutated residues. Point mutations on the mutant areconcentrated in the region in contact with the a subunit of hIL-2R.We used flow cytometry to assess the capacity of the new mole-cule to bind to the CTLL-2 cells. This cell line expresses the high-affinity abg IL-2R just as the activated T cells do, with $5–10-fold excess of the free a-chain (20). For detection of His-taggedrecombinant proteins bound to the cell surface, we used a PE-conjugated anti–His-tag mAb. At the high concentrations used,wtIL-2 and hIL-2 mutein bind in the same way to the cell surface.Furthermore, a competition assay was performed to establishwhether the hIL-2 variant is able to inhibit the binding of the anti-CD25 mAb, clone 3C7, which competes with wtIL-2 (Fig. 2A).As a result, when cells were preincubated with hIL-2 mutein,binding of anti-CD25 mAb was not inhibited, showing the samemean fluorescence intensity 275 6 5.74 for cells incubated withanti-CD25 mAb alone versus 270 6 35.7 for cells preincubatedwith the hIL-2 mutant. As expected, the preincubation of cellswith wtIL-2 decreased the mean fluorescence intensity of CD25labeling to 108 6 9.71. The anti-CD25 mAb is able to bind toCTLL-2 cells even after wtIL-2 preincubation; the excess of asubunit not assembled with the rest of receptor chains must ex-plain this residual binding. This result indicates that the hIL-2mutant has, at least, less capacity than wtIL-2 for binding the asubunit receptor, and it was termed no-a mutein.We used the CTLL-2 cell line–proliferation assay to determine

the signaling capacity of no-a mutein in vitro. This cell line ishighly sensitive to IL-2 and is commonly used for IL-2 activitydetermination (21). Cells were allowed to grow in the presence ofeither wtIL-2 or no-a mutant. Fig. 2B shows the proliferationcurves, after 48 h of incubation; the no-a mutant was able to in-duce cell proliferation by behaving as an IL-2 agonist. The twoproteins induced the same maximal level of proliferation; how-ever, a 38-fold reduction in no-a mutant activity was observedwhen comparing EC50. In contrast, the mutations do not seem to

influence the global conformation of the molecule, because no-amutant is able to bind to the cell and induce proliferation.

No-a mutein conserves the capacity to direct stimulate effectorcells, but not Tregs, in vitro

We compared no-a mutein and wtIL-2 with regard to their abilityto stimulate effector cells or CD4+Foxp3+ Tregs. As effector cellswe chose CD8+ T cells and NK cells, which mainly express the bgintermediate-affinity form of IL-2R. First, CD8 cells purified frommesenteric LNs of C57/BL6 mice were cultured with differentconcentrations of wtIL-2 or mutated hIL-2. After 5 d, a consider-able increase in the percentage of CD8+CD44hi cells was ob-served, reaching 62.13% for wtIL-2 and 64.15% for no-a mutein.The mutant and wtIL-2 were able to stimulate the proliferation ofCD8+ T cells; the total numbers of CD8+CD44hi cells were de-termined, and the proliferation curves obtained were similar forthe two molecules (Fig. 3A).The capacity to induce the differentiation of NK cells from

mouse splenocytes was also tested (Fig. 3B). At the highestconcentration, the no-a mutein was as effective as wtIL-2, in-ducing the same percentages of NK1.1+B220+ cells. Nevertheless,this effect decreased faster for no-a mutant, with an EC50 of 5.96mg/ml for the mutant and 0.48 mg/ml for wtIL-2. The differencesobserved might be due to the 10% of NK cells that express thehigh-affinity form of IL-2R, because they respond better to wtIL-2than to the no-a mutant.We also evaluated the in vitro expansion of Tregs driven by

either wtIL-2 or no-a mutein (Fig. 3C). After Treg enrichment,the sample was analyzed by flow cytometry: CD4+Foxp3+ Tregsaccounted for 80% of the sample, and the most important con-taminant was CD4+Foxp32 T cells (up to 16%). The cells wereactivated with plate-bound anti-CD3 mAb, and we used 100 mg/ml of anti-murine IL-2 mAb S4B6 to inhibit the IL-2 that effectorT cells produce after activation. After culture, the total number ofTregs was determined, and proliferation curves were obtained(Fig. 3C, right panel). As expected, the activity of the mutein overTregs was significantly lower than was wtIL-2 activity, with anEC50 of 1.58 6 0.9 mg/ml for the mutant and 1.8 6 1.2 ng/ml forwtIL-2; almost 1000-fold more no a mutant was required toachieve the same expansion level. Also, when the expanded pop-ulation was assayed by flow cytometry, expansion of a populationof CD42Foxp32 T cells was seen in the samples treated with dif-ferent concentrations of the no-a mutein. This population, whichaccounted for ,2% in the initially purified lymphocytes, accountedfor .30% at the end of culture. This in vitro–expanded populationwas later identified as CD8+CD44hi T cells. This result indicatesthat, in the presence of different lymphocyte populations, the mu-tant is available for either abg IL-2R– or for bg IL-2R–expressingcells. In conclusion, the hIL-2 mutein nearly completely conservesthe capacity of wtIL-2 to induce proliferation of bg IL-2R–bearingcells, whereas its capacity to induce the in vitro proliferation ofabg IL-2R–expressing cells is severely affected. Particularly, thedecrease in signaling capacity through the high-affinity IL-2R wasnotable for Tregs. We do not expect that the same behavior takeplaces in vivo, where different T cell subsets are able to produceIL-2, and, in turn, the cytokine may stimulate Tregs. For this reason,we performed the following in vivo experiments to study the anti-tumor effect of the mutein in comparison with wtIL-2.

No-a mutein shows greater antitumor effect than wtIL-2

The antitumor effect induced by no-a mutant was compared withthat induced by wtIL-2 in the experimental metastases model ofthe mouse melanoma line MB16F0 (Fig. 4). This is a very sig-nificant experimental model because human melanoma is one of

FIGURE 1. Point mutations introduced on the IL-2/IL-2Ra interface.

(A) Molecular surface representation of the IL-2 molecule in green. The

residues contacting with IL-2/Ra are in red, and the mutated residues are

in blue. (B) Ribbon representation of the IL-2 molecule in green and IL-

2/Ra in cyan. The side chains of the mutated residues are shown in blue.

The Journal of Immunology 3

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

the diseases treated with wtIL-2. Previously, it was demonstratedthat MB16F0 cells do not express IL-2R and do not respond towtIL-2 exposure in vitro. C57/BL6 mice were inoculated i.v. onday 0 with 2 3 105 tumor cells and were treated on days 1–4 withsaline as control, wtIL-2, or no-a mutein. Considering that anti-tumor effector cells, mainly CD8 and NK cells, express IL-2Rbg,and that the no-a mutein fully signals through this receptor, a 20-mg dose was used in both cases. Mice were sacrificed on day 21,lungs were removed, and metastases were counted under a bin-ocular stereoscope. Native hIL-2–treated and no-a mutant–treatedmice exhibited a reduction in the number of lung metastases;however, the antimetastatic effect of no-a mutein was superior,being different from the control group (p , 0.01), as well as thegroup treated with wtIL-2 (p , 0.05).It is known that effector lymphocytes, CD8+ T cells and par-

ticularly NK cells, are relevant effectors for the antimetastaticresponse. Therefore, we tested whether depletion of such pop-ulations abrogates the antitumor effect observed in vivo after no-amutant treatment. Depletion capacities for all Abs were assessedby flow cytometry analysis of mouse spleen cell suspensions.Depletion with both mAbs, anti CD8 and anti-NK1.1, abrogatedthe antimetastatic effect induced by treatment with no-a mutein(Fig. 4D); consequently, this population seems to be the majoreffector cells mediating the delayed growth of lung metastasesinduced by mutein treatment.The in vivo antitumor effect of no-a mutein treatment was also

assayed in the experimental metastases model of the 3LL-D122cell line (Fig. 5). We verified that 3LL-D122 cells do not expressIL-2R and do not respond to wtIL-2 exposure in vitro, just likeMB16F0 cells. As described above, C57BL6 mice were inoculatedi.v. with 2 3 105 tumor cells and treated on days 1–4 with salineas control, wtIL-2, or no-a mutein. Mice treated with no-a muteinshowed a marked reduction in nodule numbers, being statisticallydifferent from the control group, whereas mice treated with wtIL-2 did not show any reduction compared with the control group.Furthermore, using the same tumor cell line, we compared theeffect of no-a mutein with that of wtIL-2 in the model of spon-taneous metastases, which better resembles the clinical setting forhuman disease. As described in the previous section, mice wereinoculated in the footpad with 23 105 3LL-D122 cells and treated

for 4 d after challenge. Tumor growth was monitored; whentumors reached 0.8–0.9 mm in the control group, they were sur-gically removed, and mice were treated for 4 d with the sametreatments. The no-a mutein also showed a clear antimetastaticeffect in this therapeutic setting, being better than wtIL-2 andstatistically different from the control group treated with PBS(p , 0.05).

No-a mutein induces less toxicity than wtIL-2

The toxicity induced by the high doses of IL-2 used is one of thelimitations of IL-2–based therapy. The most frequent complicationinduced by IL-2 therapy is vascular leak syndrome, resulting inedema and lymphocyte infiltration in several organs. We decidedto compare the toxicity induced by wtIL-2 and the no-a mutein.BALB/c mice were injected i.p. with 80 mg of each protein, twicea day, for 5 d; this dose is four times higher than that used toachieve antitumor effect. After treatment, a significant incrementin the weights of lungs and livers was observed in the grouptreated with wtIL-2 but not in the groups treated with no-a muteinor PBS. Moreover, histopathological studies of the lungs andlivers of IL-2–treated mice revealed perivascular lymphocyticinfiltrations. Such infiltrations were not observed in organs fromno-a mutein–treated mice, which were similar to organs from thecontrol group (Fig. 6). These results suggest that, because the no-amutein characterized in this work induces less toxicity than doeswtIL-2, it could be used at higher doses than wtIL-2 to obtainbetter efficacy.

Modeling the therapeutic impact of IL-2 muteins witha reduced capacity to bind to CD25

We resorted to mathematical modeling to theoretically address thetherapeutic impact of eliminating or severely reducing IL-2’s ca-pacity to interact (bind) with the a-chain of its receptor (CD25).We used a mathematical model that was developed and calibratedby our group (17). This model focuses on the complex interplay ofIL-2 with the dynamics of regulatory CD4+CD25+FoxP3++ T cells(R), CD4+ Th cells (E), memory CD8+CD44+ T cells (M), and NKcells (M). In particular, it takes into account our current knowl-edge of the differential and dynamical expression of the a-, b-,and g-chains of IL-2R. To address the immune-stimulating po-

FIGURE 2. IL-2 mutein is less efficient than wtIL-2 in activating the IL-2abg receptor in vitro. (A) Flow cytometry graphs showing the ability of IL-2

mutein to bind to the CTLL2 cell surface. Direct binding, filled graph negative control, open graph wtIL-2 (upper left panel) or IL-2 mutein (upper right

panel) bound to the cells and detected with an anti 6hist- MAb-PE. Flow cytometry graphs showing the competition assay (lower panels). Negative control

(filled graph), cells labeled with anti–CD25 MAb-PE (open graph), and CTLL2 cells previously incubated with IL-2 (shaded graph; left panel) or IL-2

mutein (shaded graph; right panel) and then labeled with the anti–CD25 MAb-PE. (B) CTLL2 colorimetric proliferation assay. Data represent the per-

centage of reduced alamarBlue reduction. The mutein is able to sustain cell proliferation with a sp. act. that is 38-fold lower than wtIL-2. All experiments

were performed three times.

4 IL-2 MUTEIN

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

tential of different therapies, the model is set to a steady-stateequilibrium, where Tregs (R) effectively (dominate) regulate theexpansion of autoreactive effector T lymphocytes (E cells) andmemory cells (M cells). Such a steady state is interpreted in themodel as natural tolerance. Then the system dynamics is perturbedby simulating the injection of a desired agent, further evaluatingwhether it is driven away from the natural-tolerance steady state.We explored, through mathematical modeling, the effect of

injections with no-a mutants on T lymphocytes. The no-a mutantsare simulated as molecules that share all of the quantitativeproperties of wtIL-2 in the model, but their affinity for bindingCD25 is reduced by the factor f (parameter f = {0 to 1}). Fig. 7Ashows the effects of a dose of 40 mg of wtIL-2 (thin lines) or no-aIL-2 (f = 0) for 5 d on different lymphocyte populations. Notehow, in these simulations, the system dynamics starts in a steadystate, with a significantly high number of Tregs (R) and a lownumber of autoreactive effector T cells (E) and NK and memorycells (M). This is the natural-tolerance steady state. The system isperturbed for 5 d; afterward, the dynamic evolution differs fortreatment with IL-2 mutant or wtIL-2. The injection of wtIL-2seems to preferentially expand the Treg population, whereas in-jection of IL-2 mutein barely expands the Tregs, favoring insteada significant expansion of the NK and memory CD8+CD44+

T cells (M cells in the model). In both cases, the effect of thetreatment is transient, and the system returns to the natural-tolerance steady state. To quantify this transient differential ef-fect on lymphocytes, we calculated, at day 6 (just after thetreatment is finished), the ratio of the increase in effector cells(E+M cells in the model) to the increase in Tregs, as a function ofthe injected dose for IL-2 muteins, with different values of f (Fig.7B). It seems that IL-2 muteins with a more severely reducedcapacity to bind CD25 (f , 0.01) have a wider range of low-intermediate treatment doses with which they exhibit a higher ca-pacity than wtIL-2 to preferentially expand the effector lymphocytes.Moreover, treatment with higher and sustained doses of IL-2 or

muteins could induce a permanent change in the system dynamics,leading to a full breakdown of the natural-tolerance steady state.Fig. 7C illustrates this result for treatment with no-a mutein (f =0) at a dose of 0.5 mg/d for 15 d. Note how the system evolvesthere into a new steady state, which is characterized by a signifi-cant expansion of the autoreactive effector T cells (M and E cells)and a net reduction in the number of Tregs (R). Such a steady statewas interpreted in the work of Garcıa-Martınez et al. (17) as asteady state of autoimmunity. In Fig. 7D, the minimal dose of no-a mutant required in the model simulation to achieve the lattereffect is quantified. This is shown as a function of the mutant’s

FIGURE 3. IL-2 mutein conserves IL-2’s capacity to stimulate effector cells but not Tregs. (A) Effect on purified CD8 cells. Cells were cultured with

titrated concentrations of native IL-2 or IL-2 agonist (10 mg/ml to 10 ng/ml) for 5 d. Flow cytometry graphs of cells gated on CD8+ [filled graphs: cells

without treatment; open graphs: cells cultured with native IL-2 (left panel) or IL-2 agonist (middle panel)]. Proliferation curves adjusted to a sigmoid dose-

response curve (right panel). (B) Effect on NK cell expansion. Total splenocytes from erythrocyte-depleted C57BL6 mice were cultured with titrated

concentrations of native IL-2 or IL-2 agonist (10 mg/ml to 16 ng/ml) for 5 d. NK1.1+B220+ cells were measured by flow cytometry. Dot plots show the

percentage of NK cells for each molecule (left panels). Proliferation curves show that IL-2 mutein was only 12-fold less active than was native IL-2 when

EC50 for each curve was compared (right panel). (C) Effect on purified Tregs. Purified Tregs were cultured in the presence of wtIL-2 or IL-2 agonist for 5 d.

Dot plots show the composition of the expanded populations (left panels). Proliferation curves of CD4+Foxp3+ T cells show that EC50 is 1000-fold higher

for wtIL-2. Data are cumulative from three different experiments.

The Journal of Immunology 5

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

capacity to bind to CD25 (f) for two mutant decay rates in theblood. The solid line represents the normal IL-2 decay rate (half-life of 9 min) (22), and the dashed line represents the decay ratereported for an IL-2 molecule fused to Fc (half-life of 7 h) (23).As can be seen in Fig. 7, wtIL-2 (f = 1) requires higher dosesthan does mutant IL-2 to effectively destroy the natural-tolerancesteady state. The minimal dose required for the mutant IL-2 isreduced significantly by increasing its half-life in blood. Inter-

estingly, the model predicts that a significant reduction, of at leasttwo orders of magnitude (f , 0.01), in the capacity of IL-2 to bindCD25 is required to see a clear differentiation of mutein propertieswith respect to wtIL-2.

DiscussionIn the 1990s, IL-2 was approved by the U.S. Food and DrugAdministration for the treatment of melanoma and renal cellcarcinoma. The clinical efficacy achieved with IL-2 therapy has notmet expectations, because only 15–20% of patients respond to thetherapy. But it is relevant that some of those patients experiencetotal or long-lasting responses (5). The IL-2–driven expansion ofTregs in vivo has been related to this poor efficacy. In addition,high-dose treatment with IL-2 induces severe toxicity, with vas-cular leak syndrome being the major symptom. Therefore, thesearch for improvements in IL-2 as a therapeutic agent remainsa very attractive topic.

FIGURE 4. Antitumor effect of

IL-2 and IL-2 mutein in the MB16F0

experimental metastases model. (A)

Schedule of treatment for mice

bearing MB16F0 cells. (B) Photo-

graphs of representative lungs of

each group. (C) MB16F0 nodule

counts. Cumulative data from three

different experiments (n = 8 mice/

group). Bars represent means; error

bars represent SEM. IL-2 mutein

shows greater antimetastatic effect

than does native IL-2. **p , 0.001,

*p , 0.05, Bonferroni multiple-

comparison test. (D) In vivo CD8+

cell depletion abrogates the anti-

metastatic effect on IL-2 mutein. (E)

In vivo NK1.1+ cell depletion abro-

gates the antimetastatic effect on

IL-2 mutein. *p , 0.01, Bonferroni

multiple-comparison test.

FIGURE 5. Antitumor effect of

IL-2 and IL-2 agonist in the 3LL-

D122 mouse lung carcinoma model.

(A) Treatment schedule for mice

bearing 3LL-D122 cells: experi-

mental metastases model (upper

panel) and spontaneous metastases

model (lower panel). (B) Photo-

graphs of representative lungs for

each group in the experimental me-

tastases model. (C) 3LL-D122 nodule

counts from the experimental metas-

tases model. (D) 3LL-D122 nodule

counts from the spontaneous metas-

tases model. Figures show cumulative

data from three different experiments.

Error bars represent SEM. *p, 0.05,

versus control, Bonferroni test.

6 IL-2 MUTEIN

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

In this study, a new IL-2 mutant, termed no-a mutant, wasdesigned for its noticeably reduced capacity to stimulate Tregs andits capacity to induce the proliferation of TCD8+ and NK cells.The no-a mutein behaves as an IL-2 agonist in vitro, and itexpands CD8+ cells and NK cells as well as does wtIL-2. Incontrast, the hIL-2 mutein shows an ∼40-fold reduced capacity tostimulate the proliferation of the CTLL2 cell line, an abg IL-2R–expressing cell line, and an even greater reduced capacity (1000-fold) to stimulate CD4+CD25+Foxp3+ cells in vitro. The no-amutant was also assayed in two tumor models in mice. Aftertreatment, the number of lung nodules was reduced in the

3LLD122 and melanoma MB16F0 cell lines, and the muteinshowed a stronger anti-metastatic effect than did wtIL-2. Thisantimetastatic effect was abrogated when NK or CD8+ cells weredepleted. In addition, in preliminary studies, the no-a muteinseems to induce less toxicity in mice than does wtIL-2.From the three-dimensional structure of the quaternary IL-2/

IL-2Rabg complex solved by Wang et al. (24), and using Pymolsoftware for structure visualization and calculation, we locatedthe mutated residues on the hIL-2 molecule surface. All mutatedresidues are separated by#5 A from the a receptor subunit. Whena computational alanine scanning of the IL-2/IL-2Ra interfacewas performed, we found that all of the mutated residues makesignificant contributions to the IL-2/IL-2Ra interaction, suggest-ing that the hIL-2 mutein does not bind to, or has a reduced ca-pacity to bind to, the a receptor subunit. This suggestion wasverified, in part, by the fact that IL-2 mutein is not able to displacethe binding of anti-CD25 mAb (3C7) to the CTLL2 surface,whereas wtIL-2 competes with 3C7 mAbs and induces a markedreduction in labeling intensity. Individual mutations in the resi-dues R38 and F42 were reported by Heaton et al. (25) as beingimportant for disrupting IL-2 interactions with the a-chain. In thiscase, we included mutations in these positions and at Y45 and E62at the same time; consequently, we expected a dramatic decreasein IL-2 affinity by the a-chain. In the study by Heaton et al. (26),R38A and F42K variants showed only a 50% reduction in sp. act.;in our work, the introduction of four mutations in the same mol-ecule induced a stronger (40-fold) decrease in sp. act. over theabg IL-2R–bearing CTLL 2 cell line.Several hIL-2 agonists, with mutations on different regions of

the molecule surface, have been described; most of them weredesigned and tested in vitro for structural characterization of theIL-2 molecule. Nevertheless, some previous studies with IL-2muteins aimed to obtain variants with a greater therapeutic in-dex than hIL-2. The work of Shanafelt et al. (15) is one of the mostsignificant; a mutant, named BAY 50-4798, exhibited antitumoractivity and was better tolerated than hIL-2. Different from themutein described in the present study, BAY 50-4798 was mutated

FIGURE 6. IL-2 mutein induces less toxicity that does wtIL-2. Five

BALB/c mice/group were inoculated with 80 mg of wtIL-2 or IL-2 mutein

for 5 d. Lungs and livers were weighed, and histological analyses were con-

ducted. (A) Organ weights after treatment. (B) Representative regions of H&E-

stained lungs and livers (original magnification 310 [upper panels], 340

[lower panels]). Data are representative of three independent experiments.

FIGURE 7. Simulations of transient injections with

no-a IL-2 muteins and wtIL-2. (A) Kinetics outcomes

of these treatments, using a dose of 40 mg of wtIL-2

(thin lines) or no-a IL-2 (thick lines) for 5 d (time point

indicated by the vertical arrows). Solid and dashed lines

correspond to the kinetic evolution of the total number

of effector cells (E+M) and Tregs (R), respectively. (B)

Quantification of the ratio of the fold increase in ef-

fector cells (E+M) and the fold increase in Tregs (R), at

day 6 after the start of treatment, as a function of the

dose used. Thick lines correspond to treatment using

no-a IL-2 mutants (f = 0) or wtIL-2 (f = 1). Thin lines

correspond to no-a IL-2 mutants with different values

of f: f = 0.01 (points line), f = 0.05 (dashed line), and

f = 0.5 (continuous line). (C) Kinetic outcome of

transient injection with no-a IL-2 mutants, using a dose

of 0.5 mg for 2 wk. Solid and dashed lines correspond

to the kinetic evolution of the total number of effector

cells (E+M) and Tregs (R), respectively. (D) Dose de-

pendency predicted as a function of the value of f. Both

lines correspond to different half-lives of the mutant:

t = 20 min (continuous curve) and t = 3 h (points curve).

The Journal of Immunology 7

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

at the interface with either IL-2Rb or IL-2Rg, but the residuesimplicated in the interaction with IL-2Ra remained untouched.Other investigators also mutated the residues implicated in the

interaction with IL-2Ra. In 2003, Rao et al. (16) tested the hy-pothesis that a mutein with incremented affinity for IL-2Ra musthave greater biological potency; nevertheless, these mutants didnot exhibit increased activity in vitro. In contrast, Heaton et al.(25) tested whether R38 or F42 variants, which must have affectedthe interaction with IL-2Ra, were able to reduce the secretion ofIL-1b, TNF-a, TNF-b, and IFN-g. PBMCs stimulated with bothmutants induce the same lytic activity as does wtIL-2, but theconcentration of inflammatory cytokines tested in the supernatantof such culture decreased compared with that produced by wtIL-2stimulation. The investigators postulated that such single-pointmutated variants could be used for immunotherapeutic purposes,and they must evoke less systemic toxicity, based on the idea thatthe major side effects of IL-2 treatment are caused primarily bythe secretion of IL-1b and TNF-a. More recently, Krieg et al. (27)postulated that IL-2–mediated pulmonary edema depends on thedirect effect on lung endothelial cells that express IL-2Rabg ina functional form and that blocking the a-chain is enough to reducethe pulmonary toxicity induced by IL-2. This idea also explainsa possible low toxicity induced by mutants described by Heaton andeven more for the mutein described in the present work, in whichthe interaction with the a subunit seems to be totally disrupted.Indeed, in this study, we obtained primary in vivo evidence showinglower toxicity induced in mice by the no-a mutein. This latterproperty is important because the high toxicity associated withIL-2–based therapy is one of its major drawbacks.Furthermore, we used a previously calibrated mathematical model

to assess the impact of treatment with a mutein of IL-2 with a reducedcapacity to bind to CD25 on lymphocyte dynamics. The results of oursimulations suggest that an IL-2 mutein with severely reduced or nullcapacity to bind to a-chain of IL-2R (CD25) is a more potent inducerof immunity than is wtIL-2. Such increased capacity relies on thequantitative nature of the relationship between IL-2 and lymphocytedynamics, particularly on the regulated differential expression of thea- and b-chains of IL-2Rs among lymphocytes. A significant re-duction in the interaction of IL-2 with CD25 (of at least two ordersof magnitude, f , 0.01) reduces its preferential use by Tregs (a factcharacteristic of wtIL-2 in vivo dynamics), making the IL-2 mutantequally accessible to Tregs and Th cells and even more preferen-tially used by NK and memory CD8 T cells. Such an imbalance issufficient to drastically alter the overall T cell dynamic response.Furthermore, our simulations predict that, in the no-a muteins, thelower the capacity to bind to CD25 and the smaller their decay ratein vivo, the greater their immune-stimulatory potential.The experimental evidence reported in this article shows that an

hIL-2 mutant, with a severely reduced binding capacity to the IL-2Ra subunit, has a higher therapeutic efficacy than does wtIL-2in two experimental metastasis models in mice. Such increasedtherapeutic efficacy might be explained, following our in vitroexperiments and the in silico simulations, by the preferential useof the mutein by NK cells and CD8 memory T cells and not by theCD4+CD25+ Tregs. The dynamic effects predicted in silico forthis type of IL-2 mutein seem to be qualitatively similar to thosepredicted in the same calibrated mathematical model for treat-ments with immune complexes of IL-2 and anti–IL-2 mAbs, as faras the mAb block the binding of IL-2 to CD25 (17). From a morequantitative point of view, the predicted effects are clearly stron-ger for mutants with a greater reduced capacity to bind CD25 andlonger expected life spans in vivo.Consistent with the latter ideas, Boyman et al. (28) reported

a substantially higher immune-stimulatory effect in vivo for some

IL-2/IL-2 mAbs complexes compared with wtIL2. They describedthe unexpected property of the so-called ‘‘stimulator IL-2 mAbs’’to dramatically expand, in vitro and in vivo, the memory CD8+

CD44hi T cells, which express IL-2Rbg in higher amounts than doCD4+ T cells. Subsequently, Krieg et al. (27) demonstrated thattreatment with such immune complexes, which they called ‘‘IL-2/mAbCD122,’’ reduced the number of lung nodules in the MB16F0experimental metastasis model. Furthermore, it was demonstratedthat the increase in life span of IL-2 in vivo, which the mAbsconfers, is not sufficient to achieve the superagonist effect ofstimulator immune complexes. In contrast, treatment of mice withIL-2–IgG fusion protein and concomitant blocking of the CD25molecule with the mAb PC-61 was the only way to reproduce theextensive proliferation of CD8+CD44hi T cells induced by the IL-2/S4B6 mAb immune complex (29). Although the direct bindingof stimulator anti–IL-2 mAbs, S4B6 and 5H4, was not demon-strated in the original work, Boyman et al. (28) postulated thatthose mAbs bind to the IL-2 molecule at the IL-2Ra interface.However, the hypothesis was recently demonstrated by Rojaset al. (30) via fine epitope mapping of the anti–mIL-2 mAbs usingphage display technology. The above-mentioned mAbs indeedbind to the IL-2 molecule by the same region of interaction with areceptor chain. The IL-2 mutein described in the present study isa direct demonstration that blocking the interaction of IL-2 withthe CD25 chain is sufficient to enhance IL-2 action over CD8+

CD44hi cells, as well as other IL-2Rbg–expressing cells, such asNK cells. In the case of no-a mutein, a negligible increment inIL-2 life span is expected, because the mutein has the same mo-lecular weight as does wtIL-2. However, their fusion to Fc mol-ecules or other carrier proteins might substantially increase theirlife span, as well as their in vivo activity.To our knowledge, this is the first time that an IL-2 mutein, which

binds only IL-2Rbg, has been tested in vivo for antitumor activ-ity, showing a stronger effect than wtIL-2. Different from otherstudies, we mutated four residues on the same molecule; never-theless, the purified mutant was stable and conserved the capacityto bind to the cell surface. Multiple mutations of relevant residuesat the interface with the a subunit seem to produce a drasticdisruption in the interaction with this IL-2R subunit and seem tobe related to better antitumor activity. Based on our in vitro and insilico results, we postulate that, in vivo, the no-a mutein mustbehave in a similar way to the IL-2/S4B6 immune complex,preferentially stimulating the proliferation of memory phenotypeCD8+ cells and NK cells that express IL-2Rbg in large amounts.This direct effect of no-a mutein on lymphocyte populationsremains to be tested. The dual role of IL-2—promoting theexpansion of effector cells while maintaining the generation andhomeostatic proliferation of Tregs—is a limitation for its use incancer and HIV therapy. The realization that it is possible toseparate the two functions, by selectively disrupting the interac-tion of IL-2 with CD25, is a new concept that holds promisefor novel therapeutic developments. The basic idea of segregatingdifferent functions of a cytokine by selectively disrupting its in-teraction with the different subunits of its receptor is quite ap-pealing. It might be relevant for future drug designs using IL-2, aswell as other cytokines with multimeric receptors.

DisclosuresThe authors have no financial conflicts of interest.

References1. Kovanen, P. E., and W. J. Leonard. 2004. Cytokines and immunodeficiency

diseases: critical roles of the gamma(c)-dependent cytokines interleukins 2, 4, 7,9, 15, and 21, and their signaling pathways. Immunol. Rev. 202: 67–83.

8 IL-2 MUTEIN

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

2. Lin, J. X., and W. J. Leonard. 1997. Signaling from the IL-2 receptor to thenucleus. Cytokine Growth Factor Rev. 8: 313–332.

3. Morgan, D. A., F. W. Ruscetti, and R. Gallo. 1976. Selective in vitro growthof T lymphocytes from normal human bone marrows. Science 193: 1007–1008.

4. Smith, K. A. 1988. Interleukin-2: inception, impact, and implications. Science240: 1169–1176.

5. Rosenberg, S. A. 2001. Progress in the development of immunotherapy for thetreatment of patients with cancer. J. Intern. Med. 250: 462–475.

6. Siegel, J. P., and R. K. Puri. 1991. Interleukin-2 toxicity. J. Clin. Oncol. 9: 694–704.

7. Boyman, O., and J. Sprent. 2012. The role of interleukin-2 during homeostasisand activation of the immune system. Nat. Rev. Immunol. 12: 180–190.

8. Liao, W., J. X. Lin, L. Wang, P. Li, and W. J. Leonard. 2011. Modulation ofcytokine receptors by IL-2 broadly regulates differentiation into helper T celllineages. Nat. Immunol. 12: 551–559.

9. Kundig, T. M., H. Schorle, M. F. Bachmann, H. Hengartner, R. M. Zinkernagel,and I. Horak. 1993. Immune responses in interleukin-2-deficient mice. Science262: 1059–1061.

10. Sadlack, B., H. Merz, H. Schorle, A. Schimpl, A. C. Feller, and I. Horak. 1993.Ulcerative colitis-like disease in mice with a disrupted interleukin-2 gene. Cell75: 253–261.

11. Sakaguchi, S., N. Sakaguchi, M. Asano, M. Itoh, and M. Toda. 1995. Immu-nologic self-tolerance maintained by activated T cells expressing IL-2 receptoralpha-chains (CD25). Breakdown of a single mechanism of self-tolerance causesvarious autoimmune diseases. J. Immunol. 155: 1151–1164.

12. Malek, T. R. 2003. The main function of IL-2 is to promote the development of Tregulatory cells. J. Leukoc. Biol. 74: 961–965.

13. Cesana, G. C., G. DeRaffele, S. Cohen, D. Moroziewicz, J. Mitcham,J. Stoutenburg, K. Cheung, C. Hesdorffer, S. Kim-Schulze, and H. L. Kaufman.2006. Characterization of CD4+CD25+ regulatory T cells in patients treatedwith high-dose interleukin-2 for metastatic melanoma or renal cell carcinoma. J.Clin. Oncol. 24: 1169–1177.

14. Curiel, T. J., G. Coukos, L. Zou, X. Alvarez, P. Cheng, P. Mottram, M. Evdemon-Hogan, J. R. Conejo-Garcia, L. Zhang, M. Burow, et al. 2004. Specific recruit-ment of regulatory T cells in ovarian carcinoma fosters immune privilege andpredicts reduced survival. Nat. Med. 10: 942–949.

15. Shanafelt, A. B., Y. Lin, M. C. Shanafelt, C. P. Forte, N. Dubois-Stringfellow,C. Carter, J. A. Gibbons, S. L. Cheng, K. A. Delaria, R. Fleischer, et al. 2000. AT-cell-selective interleukin 2 mutein exhibits potent antitumor activity and iswell tolerated in vivo. Nat. Biotechnol. 18: 1197–1202.

16. Rao, B. M., A. T. Girvin, T. Ciardelli, D. A. Lauffenburger, and K. D. Wittrup.2003. Interleukin-2 mutants with enhanced alpha-receptor subunit binding af-finity. Protein Eng. 16: 1081–1087.

17. Garcıa-Marınez, K., and K. Leon. 2012. Modeling the role of IL2 in the interplaybetween CD41 helper and regulatory T cells: studying the impact of IL2modulation therapies. Int. Immunol. 24: 427–446.

18. Wang, A., S. D. Lu, and D. F. Mark. 1984. Site-specific mutagenesis of thehuman interleukin-2 gene: structure-function analysis of the cysteine residues.Science 224: 1431–1433.

19. Moya, G., L. J. Gonzalez, V. Huerta, Y. Garcıa, V. Morera, D. Perez, F. Brena,and M. Arana. 2002. Isolation and characterization of modified species ofa mutated (Cys125 -Ala) recombinant human interleukin-2. J. Chromatogr. A971: 129–142.

20. Smith, K. A. 2006. The quantal theory of immunity. Cell Res. 16: 11–19.21. Gillis, S., M. M. Ferm, W. Ou, and K. A. Smith. 1978. T cell growth factor:

parameters of production and a quantitative microassay for activity. J. Immunol.120: 2027–2032.

22. Donohue, J. H., M. T. Lotze, R. J. Robb, M. Rosenstein, R. M. Braziel,E. S. Jaffe, and S. A. Rosenberg. 1984. In vivo administration of purified Jurkat-derived interleukin 2 in mice. Cancer Res. 44: 1380–1386.

23. Harvill, E. T., J. M. Fleming, and S. L. Morrison. 1996. In vivo properties of anIgG3-IL-2 fusion protein. A general strategy for immune potentiation. J.Immunol. 157: 3165–3170.

24. Wang, X., M. Rickert, and K. C. Garcia. 2005. Structure of the quaternarycomplex of interleukin-2 with its alpha, beta, and gammac receptors. Science310: 1159–1163.

25. Heaton, K. M., G. Ju, and E. A. Grimm. 1993. Human interleukin 2 analoguesthat preferentially bind the intermediate-affinity interleukin 2 receptor lead toreduced secondary cytokine secretion: implications for the use of these inter-leukin 2 analogues in cancer immunotherapy. Cancer Res. 53: 2597–2602.

26. Heaton, K. M., G. Ju, and E. A. Grimm. 1994. Induction of lymphokine-activatedkilling with reduced secretion of interleukin-1 beta, tumor necrosis factor-alpha,and interferon-gamma by interleukin-2 analogs. Ann. Surg. Oncol. 1: 198–203.

27. Krieg, C., S. Letourneau, G. Pantaleo, and O. Boyman. 2010. Improved IL-2immunotherapy by selective stimulation of IL-2 receptors on lymphocytes andendothelial cells. Proc. Natl. Acad. Sci. USA 107: 11906–11911.

28. Boyman, O., M. Kovar, M. P. Rubinstein, C. D. Surh, and J. Sprent. 2006. Se-lective stimulation of T cell subsets with antibody-cytokine immune complexes.Science 311: 1924–1927.

29. Letourneau, S., E. M. van Leeuwen, C. Krieg, C. Martin, G. Pantaleo, J. Sprent,C. D. Surh, and O. Boyman. 2010. IL-2/anti-IL-2 antibody complexes showstrong biological activity by avoiding interaction with IL-2 receptor alpha sub-unit CD25. Proc. Natl. Acad. Sci. USA 107: 2171–2176.

30. Rojas, G., A. Pupo, K. Leon, J. Avellanet, T. Carmenate, and S. Sidhu. 2013.Deciphering the molecular bases of the biological effects of antibodies againstInterleukin-2: a versatile platform for fine epitope mapping. Immunobiology 218:105–113.

The Journal of Immunology 9

by guest on May 16, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from