Interleukin-4 Receptor Inhibition Targeting Metastasis Independent of Macrophages

Rhabdomyosarcoma (RMS) is the most commonly occurring soft tissue sarcoma in children. RMS is generally classified into either the alveolar (aRMS) or embryonal (eRMS) subtype. aRMS exhibits a relatively simple mutational landscape characterized by a t(2;13) or t(1;13) chromosomal translocation between PAX3 or PAX7, respectively, and FOXO1, which result in PAX:FOXO1 oncogenic fusion proteins (1). For children presenting with fusion positive metastatic aRMS the 5-year event-free survival is 8%, an outcome which has remained unchanged since the first RMS clinical trials 40 years ago (2). Multimodal therapy regimens relying on intensified chemotherapy have not improved survival but have instead resulted in increased toxicity and potential lifelong medical complications (3). In contrast, the pediatric cancer of bone, osteosarcoma, is the most commonly diagnosed bone tumor and occurs primarily in the lower extremities among children and young adults (4, 5). When disease is localized patients have an average 5-year event-free survival of 80%. However, 15% to 18% of osteosarcoma cases are initially diagnosed with metastasis and those patients respond poorly to therapeutics resulting in a 5-year event-free survival rate of approximately 30% (4–6).

Our group has previously reported a cytokine receptor as a possible therapeutic target to abrogate the metastatic burden in aRMS. The IL4 receptor (IL4Rα) and its components are expressed in aRMS cells, and pharmacologic inhibition of this pathway results in a decrease of pulmonary and lymphatic metastasis in a genetically engineered mouse model of aRMS (7) as well as a reduction in establishment of new tumors in an orthotopic allograft model (8). IL4 signaling is mediated by two distinct receptor complexes. Type I signaling occurs when the IL4Rα chain heterodimerizes to the γ_c chain, which in turn signals directly through JAK/STAT or indirectly through AKT. Type II signaling occurs through a receptor complex composed of the IL4Rα and IL13 receptor (IL13Rα1) chain, and signals directly through JAK/STAT (9, 10). The IL4Rα chain is reportedly expressed at low levels on most normal cell types (11), and yet solid tumors such as melanoma and breast carcinoma also have upregulated IL4Rα expression (12–14). IL4Rα signaling has been well characterized for not only the multiple fundamental roles this receptor plays in the innate immune response by modulating B-cell and T_H2 T-cell differentiation (11, 15), but also for the affects that aberrations to the pathway have on modulating cancers (16–18). For example, the protumorigenic role of IL4 signaling has been demonstrated in a range of epithelial cancers such as colon, breast, and lung carcinomas (12, 19, 20). Meta-analysis of clinical studies found an elevation of IL4 producing T_H2 cells in circulating peripheral blood in 50% of cancers tested, although whether the skewing towards a T_H2 phenotype is the result of generalized inflammation or a tumor response is unknown (16). Increases in melanoma pulmonary metastasis can be induced by treating mice with rIL4 ligand (21), and shRNA silencing of IL4Rα expression in breast cancer cells leads to decreases in liver and pulmonary metastatic burden (14). To expand on these breast carcinoma findings, DeNardo and colleagues elegantly demonstrated that IL4 secreted from T_H2-polarized CD4⁺ T lymphocytes could activate M2 macrophages and potentiate mammary adenocarcinoma metastasis in a mouse model (17).

Wholly considering the biological significance of IL4Rα pathway inhibition for aRMS from our prior studies and given the abundance of other research suggesting the pro-metastatic contribution of IL4 in some adult cancers, we postulated that targeting the IL4/IL4Rα signaling axis might be effective in mediating cell growth, metastasis, or both in cancers beyond pediatric aRMS. Herein, we investigate the role of the macrophage pool on mediating metastasis in aRMS, and test the effectiveness of IL4 pathway stimulation and inhibition on cell proliferation and migration in multiple adult cancer cell lines. We additionally examine the role of IL4 signaling in osteosarcoma, a pediatric bone cancer which is often incurable when metastatic.

MDA-MB-231 (HTB-26), HCT-116(CCL-247), LnCaP (Clone FGC, CRL-1740), RH30 (CRL-2061), SJSA (CRL-2098), HOS (CRL-1543), SAOS2 (HTB-85), and U2-OS (HTB-96) cells were purchased from ATCC. MCF-7 (86012803) was purchased from Sigma Aldrich. CF-1 is a cell culture isolated from an 18-month-old male presenting with alveolar RMS (aRMS; ref. 22). PCB-204 is a cell culture isolated from the high-grade, extra-skeletal osteosarcoma of a 33-year-old male (23). MCF-7 and MDA-MB-231 were cultured in DMEM media (Thermo Fisher Scientific, 11995073) supplemented with 10% FBS (Thermo Fisher Scientific, 10437036) and 1% penicillin/streptomycin (Thermo Fisher, 15140112). All other human cells were cultured in complete RPMI media (Thermo Fisher Scientific, 11875119) supplemented with 10% FBS and 1% penicillin/streptomycin. 61323 is an osteosarcoma cell culture isolated from a tumor arising from an Osx1-Cre, Trp53^−/−, Rb1^−/− genetically engineered mouse as previously characterized (24). 61323 was cultured in DMEM media supplemented with 10% FBS and 1% penicillin/streptomycin. All human cell cultures were authenticated using short tandem repeat analysis performed by the University of Arizona Genetics Core.

Cell viability assays were performed as described previously (7). Briefly, cells were seeded at a density of 3 × 10³ cells per well in an opaque-walled 384 multiwell plate, or 1 × 10⁵ cells per well in an opaque-walled 96-well plate. To inhibit IL4 signaling pathway, cells were treated with human IL4Rα antibody (R&D systems, MAB230) in normal growth media; 10 μg/mL IgG2A isotype (R&D systems, MAB003) was used as a control. To stimulate IL4 pathway, cells were seeded in DMEM or RPMI media with 2% donor equine serum (Thermo Fisher Scientific, 16050122), then IL4 ligand (Peprotech, 214-12) or IL13 ligand (Peprotech, 210-13) were added at varying concentrations. Seventy-two hours after addition of inhibitor or ligand CellTiter-Glo 2.0 (Promega, G9341) was used according to manufacturer instructions and a BioTek Synergy 2 plate reader was used to evaluate cell viability. All data points were collected in quadruplicate and each experiment was repeated three times.

All in vivo experiments were conducted in accordance with the institution-approved IACUC protocol. In vivo studies were carried out as described previously (8, 25). Eight-week-old female SHO mice (Charles River, 474) were injected with cardiotoxin into the gastrocnemius to induce muscle injury, followed 24 hours later by inoculation with 1 × 10⁶ Rh30-RFP-LUC, U61323, or U48484 cells in sterile PBS (Thermo Fisher Scientific, 10010002). A total of 1 × 10⁶ HCT-116 cells suspended in sterile PBS were injected subcutaneous into the flank of experimental mice with no pre-injury. For studies directly investigating perturbation of the IL4 axis, mice injected with human derived cells were treated with 25 mg/kg RegN675 (Regeneron). Mice injected with murine U48484 aMRS cells were treated with 25 mg/kg RegN1103 or RegN653. Mice injected with mouse U61323 osteosarcoma cells were treated with or 2.8 mg/kg mIL-4Rα mAb (BD Bioscience, 552288) or IgG control (BD Bioscience, 554687) as previously described (7, 8). Mice were treated 2 days preceding inoculation, then 4, 10, and 17 days post inoculation. For macrophage inhibition studies, mice were treated with 25 mg/kg of anti-CD115 (CSF-1R) antibody (BioLegend, 135510) by intraperitoneal injection twice weekly. Mice in all studies were monitored three times weekly and digital caliper measurements of tumor volume was recorded two times weekly. All animals were sacrificed by CO₂ asphyxiation when tumor volume reached 2.5 cm³.

To more clearly visually observe lung metastasis in in vivo studies, Rh30 cells were transduced with RFP-LUC lentiviral particles (Gentarget, LVP677) according to manufacturer's instructions. Briefly, Rh30 cells were seeded at a density of 4 × 10⁴ cells per well of a 12-well plate. When cells were 60% confluent culture media was removed and replaced with 0.5 mL warmed DMEM and 45 μL virus was added dropwise to media. After 72 hours cells were trypsinized and placed in 1 μg/mL puromycin in complete media to select for transfected cells. The resulting RFP+, LUC+ cells used for the previously described in vivo studies.

To generate a metastatic colon carcinoma cell line, HCT-116 cells were injected via tail vein into three SHO mice. Twenty-one days following injection, mice were sacrificed by CO₂ asphyxiation and lungs were removed. Grossly observed pulmonary metastasis were excised and expanded in RPMI complete growth medium. Cells from lung metastasis #3 (isolated from mouse U70001) were chosen due to their metastatic potential and used to re-inject as an orthotopic xenograft by subcutaneous injection into SHO mice.

Significance of cell viability assays were calculated using one-way ANOVA. Statistical significance for all tumor growth curves was calculated using a repeated measures two-way ANOVA. Normality of tumor growth curves was tested using Shapiro–Wilk test. One animal injected with U7001LM cells was censored from all analysis in the RegN646 treated group due to deteriorated health of the animal (Fig. 4E). Differences in survival presented as Kaplan–Meier curves were analyzed for significance using log-rank test. Differences in pulmonary metastatic burden presented in RT-PCR data were analyzed by Mann–Whitney test. Group contrasts with regard to location of pulmonary metastasis presented in bar graph form was carried out with exact Wilcoxon tests.

To more clearly visually observe lung metastasis in in vivo studies, a lentiviral gene delivery system was used to generate a human aRMS cell line that stably expresses RFP-Luciferase. Pre-made lentiviral particles (GenTarget, LVP674) were used to transduce Rh30 human aRMS cells following manufacturer instructions. A total of 4 × 10⁴ per well of a 12-well plate were transduced with 45 μL lentivirus in 0.5 mL RPMI growth medium for 72 hours. After 72 hours, cells were trypsinized and placed in 1 μg/mL puromycin in complete media to select for transfected cells. Cells were cultured for an additional 14 days and FACS (BD FACS Aria II) was conducted to isolate 8% of cells with the most intense RFP signal. Selected cells were cultured for an additional 14 days in growth media then used for experiments. The resulting Rh30 RFP+, LUC+ cells were used for the previously described in vivo studies.

Cultured cells were lysed using RIPA buffer (Thermo Fisher Scientific, 89900) supplemented with HALT protease and phosphatase inhibitor (Thermo Fisher Scientific, 78441), and total protein was determined using BCA assay (Thermo Fisher Scientific, 23225). Sixty micrograms of protein was resolved using SDS-PAGE gel (Bio-Rad, 4561034) and transferred to PVDF membranes. Membranes were blocked using 5% BSA (Jackson ImmunoResearch Labs, 001-000-162) diluted in Tris-buffered saline with Tween-20 (Thermo Fisher Scientific, BP377-500) then incubated overnight with primary antibodies IL4Rα (AbCam, 203398) or β-Actin (AbCam, 8227), or those from Cell Signaling Technology, detecting GAPDH (5174), p-AKT (4060), AKT (9272), p-Stat6 (9361), or Stat6 (5397). HRP conjugated goat anti-rabbit (Vector laboratories, PI-1000) was used as a secondary antibody, and protein was visualized using Clarity ECL Substrate (Bio-Rad, 1705061) on an IVIS Lumina imager (Perkin-Elmer).

We previously demonstrated that IL4 signaling is active in aRMS tumor cells, and that pharmacological blockade of IL4Rα with a commercially available antibody extends survival in a genetically-engineered mouse model of aRMS by reducing lung and lymph node metastasis (7, 8, 26, 27). To extend these findings, we aimed to evaluate the activity of a clinical grade reagent in preventing metastasis in an orthotopic allograft model of aRMS. For these allograft studies we utilized RegN1103, the mouse homologous antibody to RegN668 (dupilumab, trade name Dupixent), an FDA approved IL4Rα antagonist that inhibits IL4 signaling through the type I and type II receptor complexes. SHO mice deficient in B and T cells were allografted with murine aRMS U48484 cells (28) and treated with RegN1103 (n = 10) or the isotype control RegN653 (n = 7) beginning when tumor volume reached 0.25 cm³. Survival was extended for mice treated with RegN653 isotype control (P = 0.02) compared with RegN668 (Fig. 1A). We observed no significant difference in tumor volume of mice treated with RegN1103 versus isotype control (P = 0.071; Fig. 1B). Histologic analysis was performed to determine the frequency of large, small, or any size pulmonary metastasis. As seen in our previously published work utilizing a tool antibody, mice treated with IL4Rα blocking antibody trended towards less metastasis, although this difference did not reach statistical significance (small mets P = 0.15, large mets P = 0.35, any mets P = 0.51; Fig. 1C).

To determine the contribution of the host immune system to the tumor microenvironment, we next aimed to elucidate the role of macrophages as facilitators of metastatic disease in vivo. T_H2 cell secreted IL4 has been shown to activate the pro-angiogenic M2 phenotype of macrophages, and we sought to clarify if the reduction in metastasis observed with IL4R inhibition was a result of perturbed tumor cell intrinsic IL4 signaling, or due to interrupted paracrine signaling of the immune system. B- and T-cell deficient SHO mice were allografted with U48484 murine aRMS cells harboring the Pax3:Foxo1 fusion and when tumors reached 0.25 cc mice were treated with the macrophage depleting anti-CD115 blocking antibody alone (n = 5) or anti-CD115 blocking antibody plus the IL4Rα blocking antibody RegN1103 (n = 6). Survival was not changed between mice treated with RegN1103 alone or RegN1103 combined with anti-CD115 (P = 0.96) and caliper measurements revealed no difference in tumor volume between groups (P = 0.41; Fig. 1D and E). No difference in lung metastasis was observed when Pax3:Foxo1 transcript levels were examined by RT-PCR (P = 0.17; Fig. 1F). Although not statistically significant, a trend towards more lung metastasis in the macrophage depleted group appeared to be present.

The results of these studies, when considered in the context of our previously published in vivo work with an aRMS GEMM and orthotopic allograft studies (8), suggest a role of IL4R signaling independent of B and T cells for metastatic progression. Likewise, the trend towards an increased number of lung metastasis in mice treated with the macrophage depleting anti-CD115 blocking antibody imply an M1 (antitumor) polarity of the macrophage population in prevention of metastatic disease.

To ascertain if IL4 signaling was possible in other pediatric or common adult cancers, we measured the relative abundance of the IL4Rα chain in various cancer cell lines. Whole cell lysates of HCT-116 colorectal carcinoma, MDA-MB-231 triple-negative metastatic breast adenocarcinoma, MCF-7 estrogen receptor/progesterone receptor positive metastatic breast adenocarcinoma, LnCaP metastatic prostate carcinoma, Rh30 pediatric aRMS, and primary patient-derived pediatric aRMS cells CF-1 (22) were tested by immunoblotting for endogenous expression of the IL4Rα chain (Fig. 2A). Rh30 aRMS cells expressed the highest IL4Rα protein levels of aRMS and adult cancer cell lines tested (Fig. 2A).

To investigate whether IL4 signaling was biologically active in adult cancer cell lines, we next examined the expression levels of downstream targets upon stimulation or inhibition of the IL4 pathway. Cells were serum starved overnight, then IL4 pathway activation was induced by addition of 20 ng/mL recombinant IL4 ligand or inhibited by the addition of the research tool-grade antihuman monoclonal IL4Rα blocking antibody RegN675. RegN646 that binds to IL4Rα of monkey origin was used as a negative control. In all the cell lines the activation of STAT6 in response to addition of IL4 ligand was observed, as reflected by an increase in protein levels of phospho-STAT6 (Fig. 2B–G), implying that IL4R signaling is functional in these cells. In breast carcinoma cell lines MCF-7 and MDA-MB-231, and in aRMS cells Rh30 and CF-1, activation of STAT6 with IL4 ligand was mitigated by pretreatment with RegN675. An increase in the downstream target phospho-AKT was also observed in CF-1, Rh30, MDA-MB-231, MCF-7, and HCT-116 cells following stimulation with IL4 ligand. In LnCaP cells, which have been shown to constitutively activate AKT (29), no increase or decrease was observed with any treatment (Fig. 2G).

Because IL4 signaling components were found in the adult cancers we tested, and downstream target proteins responded to changes in IL4Rα signaling, we next sought to understand if IL4Rα could act as a mitogen in these cell lines. Previous reports have underscored the wide range of effects that IL4 pathway stimulation or inhibition has on adult cancer cells. Our lab and/or others have shown IL4 ligand induces proliferation in HCT-116 colon cancer cells, embryonal RMS (eRMS) RD cells, aRMS Rh30 cells, and/or aRMS mouse U21459 cells (7, 19, 30); in addition, primary prostate cancer cells co-cultured with IL4 producing feeder cells exhibit increased clonogenic potential compared with those cultured with control cells (31). To collectively examine the role IL4Rα signaling has on cell growth in adult cancers, we performed cell viability assays using IL4 or IL13 ligand to activate the IL4Rα signaling pathway, or a commercially available IL4Rα blocking antibody to inhibit the pathway. Both IL4 (P = <0.0001) and IL13 (P = 0.0032) ligand were found to stimulate cell growth in HCT-116 human colon cancer cells, although not in a dose-dependent manner, suggesting that lower concentrations of the ligand (5 ng/mL of IL4 and 50 ng/mL of IL13) saturated the receptor. Furthermore, addition of a commercially available IL4Rα blocking antibody to HCT-116 cells reduced cell proliferation (P = 0.0212), a finding that is consistent with previous reports examining the role of IL4 in colon cancer (19). LnCap metastatic prostate cancer cells showed reduced cell viability upon treatment with IL4Rα antibody (P = 0.0011), although not in a dose-dependent manner and this response was variable between concentrations (Fig. 3A). No statistically significant change was observed in MCF-7 breast carcinoma or MDA-MB-231 breast carcinoma cells in response to treatments altering the IL4 signaling pathway.

To further extend our previously published studies which demonstrated IL4 pathway inhibition results in reduced pulmonary metastasis of aRMS (8) and a genetically engineered mouse model (GEMM; ref. 7), we next sought to evaluate the role of autocrine IL4R signaling on engraftment, growth, and metastasis in human aRMS and human colon cancer xenografts. SHO mice were given a single 25 mg/kg pretreatment dose of either anti-human IL4Rα blocking antibody or IgG isotype control 2 days prior to intramuscular injection of Rh30. Following xenograft implantation mice were again treated with RegN675 (n = 4) or RegN646 control (n = 4) on days 4, 10, and 17. Consistent with previous studies, caliper measurements revealed no difference in tumor engraftment or growth between mice treated with IL4Rα blocking antibody or control (Fig. 4A and B). Pulmonary metastatic burden was measured by the presence of human GAPDH in the lungs and revealed no metastasis in mice treated with RegN646 control. Two of four mice treated with RegN675 exhibited the presence of micrometastasis as demonstrated by high RT-PCR cycling times recorded for the human GAPDH gene of interest (P = 0.42; Ct = 35.61; mean log10 expression = −5.423; Ct = 35.59; mean log10 expression = −5.13; Fig. 4C). The difference between allograft and xenograft results might be explained by species specificity of antibodies used, suggesting that the differences in microscopic metastatic burden was not due to blockade of a tumor cell autonomous function but instead a tumor–microenvironment interaction (see Discussion).

As presented above, human colon cancer (HCT-116) was the best-case model system for a tumor cell autonomous blockade effect in a nonpediatric cancer (Fig. 2F). Thus, we next queried whether blocking IL4 signaling would inhibit engraftment, slow tumor growth, or reduce metastatic burden in vivo. SHO mice were administered a single pretreatment dose of 25 mg/kg RegN675 or RegN646 control 2 days prior to an orthotopic xenograft of a pulmonary metastatic HCT-116 cell culture (U70001LM). Following xenograft implantation, mice were again treated with RegN675 (n = 3) or RegN646 control (n = 4) on days 4, 10, and 17. No difference in tumor engraftment or growth was observed in mice treated with anti-human IL4Rα blocking antibody compared with those treated with IgG control antibody (Fig. 4D and E). To examine pulmonary metastatic burden in the U70001LM xenografted mice, the presence of human GAPDH in the lungs was assessed by R-PCR. Six of seven mice injected with U70001LM did not present with evidence of metastasis, and the GAPDH detected in one sample was amplified at high cycling time (P = 0.99; Ct = 31.06; mean log10 expression = −3.86).

Because of the high mortality associated with pediatric metastatic osteosarcoma (5), we next sought to examine the role IL4Rα inhibition might play in this bone disease. We first tested murine osteosarcoma cells U61323 (24), the immortalized cell lines U2-OS, SAOS2, HOS, and SJSA, and primary osteosarcoma patient derived culture PCB-204 (23) for IL4Rα expression (Fig. 5A). Following protein analysis determining that IL4Rα was expressed on all osteosarcoma cells tested, we next determined if signaling through this receptor effected tumor growth or metastasis in vivo by testing a commercially available anti-mouse IL4Rα neutralizing antibody (IL4Rα ab) in an orthotopic allograft model of osteosarcoma. Mice were given a single pretreatment of 25 mg/kg anti-mouse IL4Rα antibody (n = 4) or IgG isotype control (n = 4) 2 days before orthotopic injection of murine osteosarcoma U61323 cells (24) harboring the Rosa26^LUSEAP allele which expresses firefly luciferase (32). Treatment was continued on day 4, 10, and 17 following inoculation with cells. Caliper measurements revealed no difference in tumor engraftment or growth in mice treated with blocking antibody compared with those treated with IgG control antibody (Fig. 5B). To examine pulmonary metastasis, luciferase reporter expression was analyzed by RT-PCR and no difference in metastatic burden was observed (P = 0.97; Fig. 5D).

IL4 signaling has been studied for over 30 years and is now known to be an essential part of the innate immune system, where this receptor plays diverse biological roles as a master regulator of B-cell differentiation, the driver of naïve T cells into T_H2 cells, and the activator of M2 macrophage polarization (11, 15, 17). More recently, the influence of IL4 signaling on various disease states has been examined. Numerous studies have examined the effects of the IL4/IL4Rα axis on metastasis, with varying conclusions gained. With notable exceptions, few of these studies delineate if affecting IL4 signaling through the tumor cell itself or the microenvironment is responsible for increases or decreases in tumor growth and metastasis. DeNardo and colleagues (17) showed that in a mouse model of mammary carcinoma, macrophages manifested an M2 phenotype in response to the T_H2 cytokines IL4 and IL13. This response resulted in pro-angiogenic and immunosuppressive activity of tumor-associated macrophages (TAM), which leads to increased pulmonary metastasis. Herein, we tested the efficacy of stimulating or inhibiting the IL4 pathway in adult breast, colon, and prostate cancer, as well as pediatric aRMS and osteosarcoma. We demonstrated that IL4 pathway inhibition does not affect tumor engraftment or growth in an orthotopic allograft model of aRMS, and we observed a trend towards fewer pulmonary metastasis in mice treated with the murine specific mAb for IL4Rα, RegN1103; however, these differences did not reach statistical significance. We speculate that our random histology sampling may have under-represented metastasis frequency and underpowered the comparison.

In our studies, we demonstrate type I and type II IL4 pathway signaling can be activated in adult and pediatric cancers by addition of IL4 ligand, as evidenced by an increase in phosphorylation of downstream STAT6 and AKT, respectively; in aRMS, breast carcinoma and prostate cancer cell lines this result could be abrogated by additionally treating cells with the IL4Rα blocking antibody RegN675. Pharmacologically induced inhibition of the IL4 signaling pathway moderately reduced cell proliferation in HCT-116 human colorectal carcinoma and LnCaP human prostate cells in vitro, but did not translate into changes in tumor growth in vivo even though all cancer cell lines tested expressed IL4Rα.

In this context, one is intrigued to consider the role of the IL4/IL-3 axis in the lung and tumor microenvironment, as the FDA has recently approved the recombinant human IgG4 mAb dupixent (Regeneron, counterpart to research tool RegN675) with binding specificity to human IL4Rα, to treat moderate to severe asthma and eczema. The pathophysiology of asthma has been well studied, and although the disease causes are complex, an aberrant Th2 response is known to contribute to symptoms (33, 34). Dupixent is thought to dually block IL4 and IL13 signaling on immune cells, which would otherwise cause airway inflammation and remodeling in response to CD4⁺ Th2 cells (35). The trend toward decreased metastasis observed in our orthotopic allograft model could conceivably be due to interruption of Th2 induced IL4 and IL13 signaling; however, the SHO mice used in these experiments do not possess functional B or T cells, leading us to believe that IL signaling is initiated through a different cell type. Moreover, because SHO mice injected with aRMS cells and treated with an anti-CD115 blocking antibody targeting the macrophage pool did not exhibit a difference in metastatic burden compared with mice treated with combined CD115 blocking antibody and IL4Rα ab, we postulate that any antimetastatic effects of IL4 signaling inhibition could be the result of an M1 > M2 macrophage population. Other hematopoietic cell types such as basophils and mast cells have been shown to secrete IL4 (36, 37), making possible that IL4Rα inhibition prevents metastasis by an RMS-tumor microenvironment (or lung microenvironment) interaction rather than a tumor cell-intrinsic mechanism. This possibility is further supported by the differing experimental results observed between the orthotopic allograft and orthotopic xenograft model of aRMS. The human versus mouse species specificity of the antibody used for these experiments might indicate that the decreases in pulmonary metastasis mediated by the blockade of IL4 signaling is a result of changes to the host cells rather than the human xenografted RMS cells.

We also note a limitation of xenograft studies, that lack of IL4 or IL13 cross-species IL4Rα stimulation implies xenograft studies only inform autocrine tumor signaling. However, genetically engineered mice and allograft studies are fully informative of the tumor microenvironment and tumor biology. Taken together, our allograft results reported here and previously published allograft and GEM studies (7, 8) suggest that IL4Rα blockade can be further examined for treatment and maintenance therapy of pediatric RMS to prevent metastatic progression. These results may not apply to pediatric or adult osteosarcoma.

C. Keller reports C. Keller has joint ventures, collaborations, or research agreements with Genentech/Roche, Eli Lilly, Novartis, Syndax, Artisan Biopharma, and Nurix Pharmaceuticals. No disclosures were reported by the other authors.

M.M. Cleary: Data curation, formal analysis, investigation, visualization, writing–original draft, writing–review and editing. N. Bharathy: Formal analysis, investigation, visualization. J. Abraham: Formal analysis, investigation, visualization. J. Kim: Formal analysis. E.R. Rudzinski: Formal analysis. J.E. Michalek: Formal analysis. C. Keller: Conceptualization, supervision, writing–review and editing.

This work was supported in part by NCI R01 CA189299, Brighter Days Childhood Cancer Organization, and Alexa's Hope Honor Page (cc-tdi.org). We are grateful to Regeneron for supplying reagents. We thank Dr. Peter Adamson for thoughtful discussions on the design of our studies and Brenda Weigel for suggesting these studies.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.