Oncolytic Virus and Anti–4-1BB Combination Therapy Elicits Strong Antitumor Immunity against Established Cancer

Oncolytic viruses (OV) offer an exciting new anticancer strategy with the potential to target both localized tumors and more advanced metastatic lesions (1–3). A distinct advantage of using OV is their capacity to replicate selectively within tumor cells leading to tumor cell death. Several DNA viruses genetically engineered to enhance tumor tropism while reducing pathogenicity are currently undergoing clinical investigation, including modified strains of adenovirus, herpes simplex virus, and vaccinia virus (Vv; refs. 3, 4). Vv, in particular, causes no known disease in humans, except for rare individuals with severely impaired immune function, providing a firm foundation for its safe use. Nevertheless, as a further safety measure, genetic engineering of the Western Reserve (WR) strain of Vv was undertaken to generate either the Vvdd mutant, by double deletion of the viral thymidine kinase (TK) and viral growth factor (VGF) genes, or the Vsp mutant by deletion of 2 viral serpin genes SP1 and SP2. Importantly, both Vv mutants retain the potent oncolytic activity of wild-type Vv but display a significant reduction in pathogenicity when examined in vivo (5).

Following OV administration in mice, some studies have shown significant antitumor effects against localized tumors, with evidence suggesting that the host immune response may contribute to viral induced antitumor responses (6–8). Although detailed mechanistic studies are lacking, results suggested that enhanced therapeutic efficacy might be achieved by using OV in combination with agents that increase the activation and strength of host antitumor immune responses.

The cellular receptor 4-1BB (CD137) is an important mediator of T-cell activation and persistence, particularly for cytotoxic CD8⁺ T cells. Administration of agonist anti–4-1BB antibody can impact on tumors in mice and in some models eradicate small tumors when used alone or in combination with other antibodies (9–11). Thus, combining OV administration with anti–4-1BB antibody may result in enhanced antitumor effects. Herein, we used syngeneic mouse models of breast and colon cancer to investigate whether and how combination therapy using oncolytic Vvdd and anti–4-1BB agonist antibody could enhance therapeutic effects in both localized and disseminated disease settings.

The B6 mouse 24JK sarcoma and the BALB/c mouse 4T1.2 breast cancer lines were kindly provided by Dr. Patrick Hwu (NIH, Bethesda, MD) and Dr. Robin Anderson (Peter MacCallum Cancer Centre, VIC, Australia), respectively. The human cervical cancer line HeLa-S3 and the African green monkey kidney cell line CV-1 (CCL-70 clone) were both obtained from American Type Culture Collection and maintained at 37°C in 5% CO₂ in RMPI-1640 media supplemented with 10% heat-inactivated fetal calf serum (FCS), 2 mmol/L glutamine, 1 mmol/L sodium pyruvate, 0.1 mmol/L nonessential amino acids, 100 U/mL penicillin, and 100 μg/mL streptomycin (Life technologies). The mouse colon adenocarcinoma line MC38 was obtained from Dr. Jeff Schlom (NIH, Bethesda, MD), and the AT-3 tumor cell line was derived as previously described (12). These lines were maintained in complete Dulbecco's modified Eagle's medium supplemented as above.

All vaccinia viruses used in this study were derived from the WR strain. Two low pathogenic mutants containing deletions in the TK and VGF genes (Vvdd) or both serpin 1 and serpin 2 genes (Vsp) have been previously described (5, 13). Viral stocks were prepared as previously reported (14).

Tumor cells were plated in a 96-well tissue culture plate at 1 × 10⁴ cells per well (100 μL) and incubated overnight at 37°C. The virus was then titrated using 10-fold serial dilutions [multiplicity of infection (MOI) from 1 to 1 × 10⁻⁵] in complete medium, before overlaying 100 μL of each dilution onto the individual cell monolayers, in triplicate. Plates were incubated at 37°C for 72 hours and the viability of the respective tumor cells determined using a MTT formazan viability assay (15). Optical density (OD) of the wells was read at 450 nm on a plate reader (Molecular Devices VERSAmax). The mean viability of the virus infected tumor cells for each viral dilution was calculated as a percentage relative to the control wells treated with media alone (100% survival) ± SEM.

Tumors and the inguinal draining lymph nodes and tumors were removed from mice at day 3, 9, and 20 posttreatment (n = 5–7 per group). Assessment of infiltrate immediately following virus administration was evaluated at day 3 posttreatment, whereas comparison of the monotherapeutic treated groups compared with the combined therapy (Vvdd and anti–4-1BB) was done at days 9 and 20 posttreatment. Lymph nodes were homogenized, passed through a 70 μm nylon sieve, and resuspended in Fc blocking solution (anti-mouse FcR clone 2.4G2 in 2% FCS/PBS) for 20 minutes at 4°C. The AT-3 tumors were disrupted mechanically and digested further with collagenases (0.3%)/DNase (1500 U) solution (Sigma) for 40 minutes at 37°C on a shaker incubator, and a single cell suspension generated by homogenization as described above. Cells were stained with a cocktail of antibodies for TCRβ, CD4, CD8, NK1.1, CD11c, CD11b, and Ly6G cell surface markers. Samples were incubated for 40 minutes at 4°C and analyzed by flow cytometry (LSR-II; BD Biosciences). The data was analyzed using FlowJo software v8.5.3 (Stanford University, Palo Alto, California), and the mean percentage of viable cells for each cell type ± SEM was determined at various time points for each treatment group.

Wild-type B6 and IFNγ^−/− mice were obtained from The Walter and Eliza Hall Institute of Medical Research (Parkville, Australia). Mice of 6 to 10 weeks of age were used for in vivo experiments in accordance with the Peter MacCallum Cancer Centre Animal Experimentation Ethics Committee guidelines. Tumors were established in mice by subcutaneous injection of 5 × 10⁵ breast carcinoma AT-3 cells or 1 × 10⁶ colon carcinoma MC38 cells into the left and/or right abdominal flank of wild-type B6 or for some experiments B6.IFNγ^−/− mice. Once tumors had reached 20 to 30 mm², mice were either treated with Vvdd virus alone (1 × 10⁹ plaque-forming unit (pfu) per injection delivered intratumorally on days 0 and 2 from start of treatment), or twice intraperitoneally with anti–4-1BB (clone 3H3, IgG2a, caprylic acid/ammonium sulfate purified, endotoxin free) antibody alone (25, 50, or 100 μg per injection delivered on days 4 and 6 from start of treatment), or a combination of Vvdd and anti–4-1BB antibody (25 μg). Control groups were left untreated. Tumor growth in mice was monitored by twice weekly measurements using electronic calipers (length × height) and the area of the tumor determined for each mouse. The mean tumor area ± SEM was recorded for each treatment group. To assess spontaneous metastases, mice were culled on day 38 posttreatment and the lungs immediately perfused with 800 μL of 10% neutral buffered formalin (NBF) via intratrachea injection. The lungs were then excised and fixed in 10% NBF overnight, processed, and stained with hematoxylin and eosin (H&E) to determine the number of lung lesions. The mean number of lung metastases ± SEM for each group was plotted. In some experiments, mice (n = 6 per group) were treated intraperitoneally with the following depleting antibodies: 100 μg α-MAC4 Ig control, α-CD8β (clone 53.5.8), or α-CD4 (clone GK1.5) (days −1, 0, 1 from start of treatment, plus twice weekly thereafter), α-Ly-6G (clone NIMP-R14) (days −1, 0, plus twice weekly thereafter), or 100 μL α-asialoGM1 (WAKO) (days −1, 0, plus once a week thereafter) to determine the impact of the various cellular subsets on the antitumor response following combination therapy.

The difference in tumor growth and survival (mean area) between the treatment groups was analyzed by 2-way ANOVA with P < 0.05 considered significant. Difference in number of lung metastases between groups was analyzed by a Mann–Whitney test. The difference in mean percentage ± SEM of each immune cell type in the draining lymph nodes and tumors between treatment groups was analyzed by 2-way ANOVA at each time point.

The host range of vaccinia virus is known to be broad and capable of infecting numerous species. To determine the susceptibility of a panel of mouse cancer cell lines to killing by 2 oncolytic mutants of Vv (Vsp and Vvdd) an in vitro cytotoxicity assay was done and results quantified using a MTT viability assay. The results revealed that all 4 cell lines (breast 4T1.2 and AT-3, colon MC38, and sarcoma 24JK) were similarly susceptible to killing by the 2 mutant viruses and that viral induced cell death was observed at low concentrations of virus (MOI ≥ 10⁻³; Fig. 1). Our results indicated that both Vsp and Vvdd mutants of Vv are capable at producing direct cytotoxic antitumor effects in vitro at low doses against various mouse cancer lines of different histologic origins. Further investigation into how Vvdd modulates the immunogenicity of tumor cells in vitro found no induction of apoptosis as measured by Annexin V⁺ or calreticulin translocation at the indicated time points (Supplementary Fig. S1A and B). Interestingly, we observed a trend of increased HMGB1 and ATP release from AT-3 cells after 16-hour treatment with Vvdd (Supplementary Fig. S1C and D). These results indicated that Vvdd OV does not cause apoptotic immunogenic cell death of these cell lines. Rather, we propose that Vvdd is inducing lytic cell death (necrosis) of the tumor cells, which is supported by other studies (16).

Recent evidence has suggested that the host immune response plays an important role in the antitumor effects mediated by OV (6–8). Hence, we hypothesized that the use of an OV in combination with an immune stimulatory reagent (anti–4-1BB antibody) may result in enhanced antitumor effects. Although Vsp and Vvdd mutant vaccinia viruses showed similar cytotoxicity of tumor lines in vitro, we used the Vvdd virus strain initially for combination studies, although it would also be interesting in future studies to compare combined therapeutic effects using the Vsp strain. We first examined combination therapy against the AT-3 breast carcinoma line, given that it allowed us to determine antitumor efficacy in both localized and metastatic disease sites. In the following experiment, we tested whether administration of Vvdd vaccinia virus and anti–4-1BB antibody could impact on the growth of established subcutaneous AT-3 breast carcinoma tumors in B6 mice. We chose intratumoral delivery of Vvdd, given higher viral titers were observed in both AT-3 and MC38 tumors compared with systemic administration (Supplementary Fig. S2). Intratumoral treatment of mice with Vvdd alone was able to significantly reduce AT-3 tumor growth compared with nontreated control tumors, although these effects were modest and not significantly different to treatment with anti–4-1BB antibody alone. In contrast, the combined use of Vvdd and anti–4-1BB antibody showed enhanced growth inhibition of AT-3 tumors that was statistically significant to either treatment alone (Fig. 2A). Furthermore, mice treated with the combined therapy resulted in an increased percentage survival that was significant compared with either treatment alone (Fig. 2B). Increasing the dosage of anti–4-1BB monoclonal antibody (mAb) alone up to 100 μg did not significantly enhance the therapeutic effect as observed with the combined therapy (Fig. 2C). We also examined whether our therapy could impact on concomitant tumors in recipient mice that were not directly treated. In this experiment we found that only the combined Vvdd and anti–4-1BB antibody treatment of established AT-3 tumors could significantly impact on AT-3 tumors injected on the contralateral flank (Supplementary Fig. S3). This result suggested that the combination approach may not only be effective against primary tumors but could potentially impact on distant metastasis at other sites.

Given that combined Vvdd and anti–4-1BB therapy resulted in strong antitumor effects against localized AT-3 tumors, we next wanted to determine whether this therapy would impact on breast cancer metastases in vivo. In this experiment, mice were injected subcutaneously with AT-3 tumor cells and lungs were obtained from both nontreated groups and mice treated with combined Vvdd and anti–4-1BB therapy at day 38 posttreatment and subjected to H&E staining to evaluate the number of metastatic lung lesions. Strikingly, this data revealed a significant reduction in the number of metastatic lesions present in the lungs of mice treated with the combination therapy compared with the nontreated control group (Fig. 3A–E). To show the broad utility of this therapy, Vvdd and anti–4-1BB antibody combined treatment also significantly reduced the growth of MC38 colon carcinoma cells in mice compared with either treatment alone (Supplementary Fig. S4). Collectively, these data showed for the first time that the combination therapy of Vvdd and anti–4-1BB antibody could significantly enhance antitumor effects against established cancer and metastases in syngeneic immune-competent mice.

To gain insight into the mechanism responsible for the enhanced antitumor effects observed in mice treated with a combination of Vvdd and anti–4-1BB mAb, the draining lymph nodes and tumors of mice from each treatment group were taken at various time points posttreatment and analyzed by flow cytometry to quantify the different immune cell subtypes present. With the exception of a decrease in CD8⁺ T cells, we found no significant difference in other innate or adaptive cell types infiltrating the draining lymph nodes following virus administration at day 3 posttreatment (before anti–4-1BB antibody injection) compared with nontreated mice (Fig. 4A and B). In contrast, there was a notable increase in the percentage of both CD11c⁺ and CD11b⁺ cells at day 9 posttreatment in the lymph nodes of mice treated with Vvdd, which was significantly increased following combined Vvdd and anti–4-1BB antibody treatment (Fig. 4C). Interestingly, the number of CD8⁺ T cells decreased further at day 9 following Vvdd treatment alone or combined therapy suggesting that these cells may have migrated toward the tumor site (Fig. 4D). There was no difference in percentage of CD4⁺ T cells at either time point following Vvdd treatment or combined therapy compared with nontreated mice. As expected, there were limited numbers of NK (NK1.1⁺) or NKT cells (NK1.1⁺ TCRβ⁺) within the draining lymph node, which was not altered with any of these therapies.

We next assessed immune infiltrate into AT-3 tumors following treatment. We found a significant increase in the percentage of both myeloid and lymphoid cell types posttherapy. This included an increase in the percentage of CD11b⁺ cells and CD11b⁺Ly6G⁺ (neutrophils) at day 3 following Vvdd therapy (Fig. 5A). Both of these cell types remained elevated after Vvdd treatment alone or Vvdd combined with anti–4-1BB antibody at day 9 posttreatment (Fig. 5C). Interestingly, systemic delivery of 1 × 10⁶ enriched CD11b⁺ cells could not transfer immunogenicity to naive recipients bearing established AT-3 tumors (data not shown). An important observation was that the percentage of neutrophils persisted in tumors at day 20 following combined therapy compared with the single-agent groups and nontreated controls (Fig. 5E). In addition to these myeloid cell types, we observed an increase in the percentage of CD8⁺ T cells and natural killer (NK) cells at day 9 following either Vvdd treatment alone or combination therapy (Fig. 5D), and the percentage of CD8⁺ T cells was sustained at day 20, particularly following combined therapy (Fig. 5F). In contrast, we found that there was no significant differences in the percentage of lymphoid cell types in AT-3 tumors at day 3 in either Vvdd treated or nontreated mice (Fig. 5B), and there was no significant change in the percentage of CD4⁺ T cells or NKT cells following combined Vvdd and anti–4-1BB antibody treatment (Fig. 5B, D, and F). Our data suggested that oncolytic vaccinia virus present in the local tumor region stimulates an increased number of antigen-presenting cells (APC) migrating to the draining lymph nodes from as early as day 3 postadministration of virus but also induces the rapid decline of CD8⁺ T lymphocytes residing in the lymph nodes. In parallel to this, there was an elevated infiltration of CD8⁺ T cells, NK cells, and myeloid cells at the tumor site following Vvdd treatment, with the addition of systemically delivered anti–4-1BB mAb permitting the increased percentage of neutrophils and CD8⁺ T cells, which were sustained out to day 20 posttreatment.

Although our flow cytometry results had shown an increased immune infiltrate correlating with therapeutic effects following combined Vvdd and anti–4-1BB treatment, we next investigated whether these cell types were functionally important in the antitumor response by depleting these subsets before therapy. For these experiments, AT-3 tumors were established in mice as previously described and then groups of mice were treated with depleting antibodies for either CD8⁺, CD4⁺ T cells, NK cells, or neutrophils before treatment with Vvdd and anti–4-1BB mAb. The results showed that depletion of CD8⁺ T cells (α-CD8), NK cells (α-asialoGM1), and neutrophils (α-Ly-6G) significantly reduced treatment-induced growth inhibition of AT-3 tumors compared with control Ig antibody–treated mice (Fig. 6). Mice depleted of CD4⁺ T cells before therapy showed comparable antitumor effects to Ig control–treated mice indicating that these cells do not play an important role in the therapeutic effect. Importantly, tumor-bearing mice injected with depleting antibodies alone showed similar growth kinetics to nontreated mice (Supplementary Fig. S5). Collectively, this data showed for the first time an important role for both innate and adaptive immune cells in the therapeutic responses mediated by combined Vvdd and anti–4-1BB antibody therapy.

Given that one of the mechanisms used by CD8⁺ T cells and NK cells for eliciting effective antitumor responses is through their secretion of IFNγ, we wanted to determine whether this cytokine was playing an important role in the therapeutic effects observed. In this experiment we compared therapeutic responses against established AT-3 tumors following Vvdd and anti–4-1BB mAb treatment in wild-type B6 and B6.IFNγ^−/− mice. Growth inhibition of AT-3 tumors was significantly reduced in IFNγ^−/− mice compared with wild-type mice indicating an important role for IFNγ (Fig. 7). Importantly, growth of untreated AT-3 tumors was comparable in both nontreated B6 wild-type and IFNγ^−/− mice (Fig. 7). Taken together, our data showed that oncolytic Vvdd vaccinia virus and anti–4-1BB antibody combined therapy can potently stimulate host immune responses to deliver a strong therapeutic effect against established cancer.

Oncolytic virotherapy for the treatment of cancer is gaining momentum, with several new viral agents currently undergoing phase II/III clinical evaluation including naturally occurring reovirus (Reolysin), and 2 engineered viruses incorporating the gene encoding granulocyte macrophage colony–stimulating factor (GM-CSF) into herpes simplex virus (OncoVEX) and vaccinia virus (JX-594; ref. 2, 17). Despite the exciting clinical translation of these biologic agents, the exact mechanisms by which they induce tumor regression presently remains unclear. A large number of in vivo studies using OV have been evaluated in xenograft mouse models which have indicated an inherent capacity of OV to be able to infect, replicate, and propagate within the tumor resulting in tumor cell lysis (13, 18–25). However, more recently it has become apparent that a complex interplay between the virus, tumor, and the host immune system exists that determines the outcome of virotherapy (6–8, 25, 26). Given that current evidence suggests an important role of the host immune system in the antitumor effect mediated by OV, we hypothesized that enhanced therapeutic responses might be achieved using a combined approach involving administration of vaccinia virus with an immune agonist anti–4-1BB antibody. Other studies have reported strong antitumor effects combining systemic anti–4-1BB administration with other reagents delivered intratumorally such as CpG or interleukin-12 and GM-CSF (27, 28). Our results showed that administration of Vvdd in combination with systemic delivery of anti–4-1BB antibody could significantly inhibit the growth of established subcutaneous syngeneic cancer cell lines AT-3 and MC38, compared with either treatment alone, and reduce spontaneous AT-3 metastatic lesions in the lung. It is presently unclear whether this is entirely because of direct effects on primary tumor growth or because of emerging metastasis being destroyed by the elicited immune response. We showed that the antitumor effects following combined therapy correlated with increased numbers of dendritic cells in tumor draining lymph nodes and the increased presence of CD8⁺ T cells, NK cells, and neutrophils at the tumor site. Indeed, enhanced antitumor responses following combination therapy was dependent on these cell types as depletion of each of these subsets reduced the overall therapeutic effect. Furthermore we showed a key role for IFN-γ, an important cytolytic cytokine secreted by both CD8⁺ T and NK effector cells. Thus this study describes for the first time that using an immune agonist antibody can significantly increase antitumor responses mediated by OVs in a combined attack against established cancer.

One advantage of using OVs such as vaccinia is their capacity for expressing various transgenes, including costimulatory molecules or cytokines, to induce stronger antitumor responses. The incorporation of B7.1 or lymphocyte function–associated antigen-3 into vaccinia virus has been shown to induce tumor-specific T-cell responses in metastatic melanoma patients (29–31). Similarly, an oncolytic adenovirus armed with TNF-related apoptosis-inducing ligand was shown to induce efficient melanoma cell killing and reduced tumor growth in mice (32). Another study showed that oncolytic vaccinia virus expressing 4-1BBL could mediate effective antitumor responses against B16 melanoma cells in mice compared with nonengineered virus, although these effects were observed only when combined with host lymphodepletion (29). This is less than ideal as immunotherapies involving host lymphodepletion is often associated with toxicity and may result in depletion of immune cells important for the therapeutic effect. In contrast, we observed potent antitumor effects in vivo in 2 different cancer models following combined vaccinia virus and anti–4-1BB therapy that did not require prior conditioning of the host.

Previous studies have reported that engagement of 4-1BB on dendritic cells using anti–4-1BB mAb enhances their production of inflammatory cytokines and enhances their ability to activate T cells (33). Our study is consistent with this as treatment of mice with Vvdd alone increased both numbers of CD11b⁺ and CD11c⁺ cells in the tumor draining lymph node which were further increased following administration of anti–4-1BB antibody. Interestingly we observed a concomitant decrease in numbers of CD8⁺ T cells from day 9 following treatment, suggesting that these cells have become activated following engagement of APCs in the lymph node and migrated to the tumor site.

To support this contention, analysis of the tumor cell infiltrate by flow cytometry revealed a significant increase in the number of CD8⁺ T cells in mice following treatment with Vvdd and anti–4-1BB mAb therapy from day 9 onwards compared with either treatment alone. We also observed a significant increase in number of NK cells at the tumor site following combined therapy. An important role for both of these effector cells was shown in subsequent depletion experiments. Although the exact role of how these cells were mediating antitumor effects is presently unclear, our studies revealed that this was possibly through secretion of the cytokine IFN-γ, given the therapeutic effect was reduced in IFN-γ gene–targeted mice. Future studies will determine whether IFN-γ is exerting antitumor effects through mechanisms such as enhancing antigen presentation on tumor cells, increasing tumor susceptibility to apoptosis and/or inducing cell-cycle arrest (34–38), or through the recruitment and activation of endogenous immune effector cells, regulation of leukocyte–endothelium interactions, and/or induction of antiangiogenesis (34, 39–41).

Interestingly, our data revealed that both Vvdd alone and in combination with anti–4-1BB antibody increased numbers of neutrophils (CD11b⁺ Ly6G⁺) at the tumor site that were sustained up to day 20 in the combined treatment group. The importance of these cells was further supported by depletion experiments using α-Ly-6G mAb, which resulted in reducing the therapeutic effect (42). Although many studies have linked neutrophils with having a protumorigenic role (43), our results are consistent with another study showing that neutrophils could mediate antitumor effects by blocking vital blood supply to the tumor following their activation (8). Furthermore, neutrophils have the capacity to release reactive oxygen species, secrete proteases, and impair microvascular perfusion and may be exerting antitumor effects in our model through a number of these different mechanisms (8, 43–46). Another recent study suggested that neutrophils can be polarized into N1 and N2 phenotypes that possess either protumorigenic and antitumorigenic roles similar to macrophages (43, 47). Whether this is the case in our model remains to be clarified. Neutrophils have been reported to express 4-1BB, and ligation of this receptor has been shown to enhance their survival (48, 49). Our data is consistent with this as we observed sustained numbers of neutrophils at the tumor site following therapy. Nevertheless, both the presence of viral induced danger signals mediated by Vvdd and 4-1BB costimulation in combination was required to mediate this effect.

In conclusion, our studies have shown for the first time that administration of an immune agonist antibody can enhance the antitumor properties of an OV against localized and disseminated cancer. However, it remains unclear what impact anti–4-1BB also had on the anti-vaccinia immune response and potentially decreasing this could further potentiate this therapy. Our data highlighted a critical role for both adaptive and innate immune host cells in mediating the therapeutic effects, suggesting that future therapies combining immune agonist antibodies with OVs that can stimulate the endogenous host immune system could lead to improving therapeutic outcomes in patients.

D.L. Bartlett has ownership interest (including patents) in Jennerex; he is a consultant and is on the advisory board of Jennerex. No potential conflicts of interest were disclosed by the other authors.

Conception and design: L.B. John, L.J. Howland, M.J. Smyth, M.H. Kershaw, P.K. Darcy.

Development of methodology: L.B. John, L.J. Howland, Z.S. Guo, P.K. Darcy.

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L.B. John, L.J. Howland, J.K. Flynn, A.C. West, C. Devaud, C.P. Duong, T.J. Stewart, J.A. Westwood, Z.S. Guo, P.K. Darcy.

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, and computational analysis): L.B. John, L.J. Howland, J.K. Flynn, A.C. West, T.J. Stewart, M.J. Smyth, M.H. Kershaw, P.K. Darcy.

Writing, review, and/or revision of the manuscript: L.B. John, L.J. Howland, T.J. Stewart, D.L. Bartlett, M.J. Smyth, P.K. Darcy.

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.J. Howland, D.L. Bartlett, P.K. Darcy.

Study supervision: M.H. Kershaw, P.K. Darcy.

The authors thank the technicians of the Peter MacCallum Cancer Centre Experimental Animal Facility for animal care and the Histology Department for their assistance in processing of H&E sections of lung tissue following therapy and Nicole McLaughlin for preparation of the depleting antibodies.

This work was funded by Project Grants from the National Health and Medical Research Council (NHMRC), Cancer Council of Victoria, and the Susan Komen Breast Cancer Foundation. T.J. Stewart was supported by a National Breast Cancer Foundation Fellowship and M.J. Smyth by an NHMRC Australian Research Fellowship. M.H. Kershaw and P.K. Darcy were supported by a NHMRC Senior Research Fellowship and Career Development Award, respectively.

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.