Research ArticleCancer

Antibody-Based Delivery of Interleukin-2 to Neovasculature Has Potent Activity Against Acute Myeloid Leukemia

See allHide authors and affiliations

Science Translational Medicine  04 Sep 2013:
Vol. 5, Issue 201, pp. 201ra118
DOI: 10.1126/scitranslmed.3006221

Abstract

Acute myeloid leukemia (AML) is a rapidly progressing disease that is accompanied by a strong increase in microvessel density in the bone marrow. This observation prompted us to stain biopsies of AML and acute lymphoid leukemia (ALL) patients with the clinical-stage human monoclonal antibodies F8, L19, and F16 directed against markers of tumor angiogenesis. The analysis revealed that the F8 and F16 antibodies strongly stained 70% of AML and 75% of ALL bone marrow specimens, whereas chloroma biopsies were stained with all three antibodies. Therapy experiments performed in immunocompromised mice bearing human NB4 leukemia with the immunocytokine F8-IL2 [consisting of the F8 antibody fused to human interleukin-2 (IL-2)] mediated a strong inhibition of AML progression. This effect was potentiated by the addition of cytarabine, promoting complete responses in 40% of treated animals. Experiments performed in immunocompetent mice bearing C1498 murine leukemia revealed long-lasting complete tumor eradication in all treated mice. The therapeutic effect of F8-IL2 was mediated by both natural killer cells and CD8+ T cells, whereas CD4+ T cells appeared to be dispensable, as determined in immunodepletion experiments. The treatment of an AML patient with disseminated extramedullary AML manifestations with F16-IL2 (consisting of the F16 antibody fused to human IL-2, currently being tested in phase 2 clinical trials in patients with solid tumors) and low-dose cytarabine showed significant reduction of AML lesions and underlines the translational potential of vascular tumor–targeting antibody-cytokine fusions for the treatment of patients with leukemia.

INTRODUCTION

The formation of new blood vessels (angiogenesis) is a rare process in the healthy adult but represents an absolute requirement for the viability and growth of solid tumors (1, 2). Because this neovascularization is mediated by angiogenic molecules released by tumor cells or by host cells (such as macrophages and lymphocytes), there is intense research activity aiming at the inhibition of tumor progression by means of a molecular blockade of proangiogenic factors (3, 4).

In addition to the development of inhibitors of angiogenesis, a second therapeutic strategy [“vascular targeting” (5, 6)] has been developed to take advantage of tumor blood vessels for the selective pharmacodelivery of potent therapeutic payloads (such as drugs, cytokines, radionuclides, photosensitizers, and toxins) to disease sites (712). Vascular tumor targeting relies on monoclonal antibodies, which specifically recognize markers on newly formed blood vessels. Markers of angiogenesis have historically been discovered by observing a selective staining of newly formed blood vessels in large immunohistochemical screening campaigns (1315). More recently, systematic transcriptomic (16) and proteomic (1719) strategies have been developed for the identification and validation of targets expressed at neovascular sites.

Our group has extensively studied the in vivo performance of vascular-targeting antibodies in solid tumors in mouse models of cancer (2023) and in patients (2426). Although we and others initially focused on vascular-targeting approaches in solid tumors, we recently discovered that certain markers of angiogenesis are selectively and abundantly expressed in most lymphoma types and can be efficiently targeted in vivo using “armed” antibodies (25, 2729).

Splice isoforms of fibronectin and of tenascin-C represent some of the best-characterized markers of angiogenesis (6, 30). Specifically, the alternatively spliced extradomains EDA and EDB of fibronectin, as well as the extradomain A1 of tenascin-C, are virtually undetectable in normal adult tissues but are strongly expressed at sites of physiological angiogenesis and tumor angiogenesis (31). The human monoclonal antibodies F8, L19, and F16 specifically recognize the EDA and EDB domains of fibronectin and the A1 domain of tenascin-C, respectively, and have been shown to selectively target tumors in mouse models of cancer and in patients (24, 3235). Although these antibodies recognize blood vessels of different cancer types (31, 36, 37), the expression of other markers of angiogenesis appears to be more restricted to hematological malignancies [for example, Bst-2 expression in lymphomas (28)].

In 2000, Padró et al. reported that a high degree of neovascularization can be observed in the bone marrow of patients with AML. The authors found that microvessel density was strongly increased in the bone marrow of AML patients compared to normal bone marrow. Patients enjoying a complete remission (CR) after induction therapy exhibited a reduction of microvessel density to values comparable to the ones of control subjects, whereas higher vessel counts could be detected in patients with residual disease (38).

Here, we analyzed the expression of the EDA and EDB domains of fibronectin and of A1 domain of tenascin-C, which are undetectable in healthy bone marrow (39), in freshly frozen bone marrow and chloroma (extramedullary AML tumor) biopsies of AML and acute lymphoid leukemia (ALL) patients, showing significant staining with the F8 and F16 antibodies. The immunocytokine F8-IL2, consisting of the F8 antibody fused to human interleukin-2 (IL-2) (33, 40), selectively localized to subcutaneously grafted AML tumors and mediated substantial tumor growth retardation in mice. The combination of F8-IL2 with cytarabine led to long-lasting tumor eradication in 40% of treated mice in an immunocompromised mouse model, and 100% of treated mice in an immunocompetent setting, in a process mediated by CD8+ T cells and natural killer (NK) cells. The findings promise to be of clinical significance, because an AML patient, who had previously relapsed from multiple lines of therapy and was then treated with the F16-IL2 immunocytokine and low-dose cytarabine, exhibited a rapid disappearance of 18F-fluorodeoxyglucose positron emission tomography (18-FDG-PET) uptake in AML lesions (multiple chloromas).

RESULTS

Immunochemical analysis of bone marrow biopsies

Twenty-one bone marrow biopsies (17 AML and 4 ALL patients) and 2 chloroma (AML) specimens were freshly frozen and analyzed using the clinical-stage F8, L19, and F16 antibodies (patient characteristics are described in tables S1 and S2). The use of freshly frozen material was necessary because these antibodies do not work in paraffin (36, 41), thus preventing the use of larger collections of formalin-fixed, paraffin-embedded specimens.

Figure 1 shows representative findings of an immunofluorescence and immunohistochemical analysis performed with a set of 13 AML specimens and 4 ALL specimens, which were collected in Muenster (additional staining results and patient characteristics can be found in fig. S1 and table S1). In addition to the F8, L19, and F16 antibodies, the KSF antibody [specific to hen egg lysozyme (42)] was used as the negative control. Blood vessels were costained using an antibody specific to von Willebrand factor (vWF). In general, F8 and F16 exhibited the strongest immunochemical staining, typically associated with vascular structures. Quantification of the immunofluorescence data, based on the percentage of area stained per microscope field, revealed that a considerable fraction of the specimens analyzed (9 of 13 for F8, 9 of 13 for F16, and 5 of 13 for L19 in AML; 3 of 4 for F8, 3 of 4 for F16, and 2 of 4 for L19 in ALL) displayed a substantially stronger vascular staining with the tumor-targeting antibodies than the negative control antibody (Fig. 1, C and D).

Fig. 1 Splice isoforms of oncofetal fibronectin (EDA and EDB) and tenascin-C are expressed in bone marrow biopsies of acute leukemia patients.

Sections of bone marrow biopsies of AML and ALL patients were analyzed by immunofluorescence, as well as immunohistochemistry (IHC) using the clinical-stage F8, L19, and F16 antibodies (which recognize splice isoforms of oncofetal fibronectin and tenascin-C) (red) and the KSF antibody (which reacts with hen egg lysozyme and was used as negative control) (red). In immunofluorescence procedures, blood vessels were additionally costained using an antibody specific to vWF (green). (A and B) Representative stainings of (A) AML specimens and (B) ALL specimens. Scale bars, 100 μm. In general, F8 and F16 exhibited the strongest immunochemical staining. The immunofluorescence stainings of leukemia patient biopsies were quantified by means of the percentage area of staining for each antibody. (C) In AML patient biopsies, 9 of 13 specimens were substantially stained with F8, 9 of 13 with F16, and 5 of 13 with L19. (D) ALL samples displayed strong staining in three of four cases for F8, three of four for F16, and two of four for L19, respectively.

Similar staining patterns were observed in an additional smaller set of four AML bone marrow biopsies, which were obtained from Berne or Kiel and analyzed by immunohistochemistry. In this set, three of four specimens were stained with F8, four of four with F16, and zero of four with L19 compared to the negative control (fig. S2A).

Two specimens of chloromas (extramedullary tumors of AML) were also analyzed and showed clear vascular staining with F8, F16, and L19 (Fig. 2 and fig. S2, B and C).

Fig. 2 Splice isoforms of oncofetal fibronectin (EDA and EDB) and tenascin-C are expressed in chloromas and colocalize with vasculature.

(A) Sections of chloromas from AML patients were analyzed by immunohistochemistry using the clinical-stage F8, L19, and F16 antibodies, demonstrating that EDA, EDB, and A1 of tenascin-C are expressed in chloroma specimens. Scale bars, 100 μm. (B) To show vascular colocalization, we performed immunofluorescence using the F16 antibody (green), in combination with an antibody specific to vWF (red), as marker of blood vessels. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (blue). Scale bar, 100 μm.

Immunochemical analysis of mouse models of leukemia and biodistribution studies

The L19 and F8 antibodies exhibit identical binding affinity toward the cognate human and murine antigen (34, 35), whereas F16 does not cross-react with murine A1 domain of tenascin-C (32). For this reason, we used the L19 and F8 antibodies to stain murine leukemia samples in a search for a suitable model in which to perform biodistribution and therapy studies.

In a first step, we subcutaneously implanted human AML cell lines NB4, HL60, and THP1 in nude mice. Immunofluorescence analysis of the resulting tumors showed an intense staining with the F8 antibody in the NB4 and HL60 model. Staining intensities with L19, or in the THP1 model, were weaker (Fig. 3A). Subsequently, we established an orthotopic model of HL60 leukemia by intravenous injection of tumor cells in severe combined immunodeficient (SCID) mice, followed by analysis of blasts in blood and bone marrow specimens. Unfortunately, unlike in the situation with human specimens, the bone marrow of mice bearing HL60 leukemia did not exhibit any detectable F8 staining, whereas leukemic cells could easily be detected using anti-human CD44 antibodies (fig. S3). For this reason, we continued our investigations with the subcutaneous NB4, HL60, and THP1 AML models in nude mice, which mimic the disease settings of chloromas.

Fig. 3 F8-IL2 selectively localizes to subcutaneous human AML tumors in mice and can promote reduction in tumor progression as well as complete tumor eradication in combination with cytarabine.

Human AML cell lines NB4, HL60, and THP1 were subcutaneously implanted in nude mice. (A) Immunofluorescence analysis of the resulting tumors showed an intense staining with the F8 antibody and weak staining with L19 (red). Blood vessels were costained using an anti-CD31 antibody (green). Scale bars, 100 μm. (B to D) Biodistribution experiments were performed by intravenous injection of radioiodinated F8-IL2 (black bars) or of the anti-lysozyme KSF-IL2 fusion protein, used as negative control of identical molecular format (white bars). (B) In NB4 xenografts (n = 6), F8-IL2 displayed tumor accumulation with 5.3% injected dose per gram (ID/g) of tumor 24 hours after injection and a tumor-to-blood ratio of 18.8. HL60 (n = 3) (C) and THP1 (n = 3) (D) xenografts displayed no tumor-targeting effects. Nude mice bearing subcutaneously grafted NB4 tumors (~40 mm3) were treated with intravenous injections (n = 5) of either saline (×), F8-IL2 (▪) (30 μg, days 8, 11, 14, and 17), KSF-IL2 (□) (same dose and schedule), or cytarabine (▴) (100 mg/kg daily, days 8 to 12). Furthermore, the effect of F8-IL2 in combination with cytarabine (▵) was studied. (E and F) In addition, the combination of F8-IL2 with cytarabine led to long-lasting complete tumor eradication in 40% of the treated mice [*, significant for F8-IL2 versus saline (P = 0.0014) and KSF-IL2 (P = 0.0025) after day 17; **, after day 15 for combination versus saline (P = 0.0017) and KSF-IL2 (P = 0.0021), n = 5, two-tailed Student’s t test]. Data represent mean tumor volumes ± SD. Arrows depict days of treatment with immunocytokines (black) and cytarabine (gray). (G) Monitoring of the body weight showed that the combination treatment reduced average body weight, without exceeding the tolerated body weight loss of 15%. Data represent average percent body weight relative to first day of therapy ± SD.

Figure 3 (B to D) shows quantitative biodistribution results obtained by intravenous injection of radioiodinated preparations of the immunocytokine F8-IL2 (40) (black bars) or of the anti-lysozyme KSF-IL2 fusion protein, used as negative control of identical molecular format (white bars). A preferential accumulation in the subcutaneous neoplastic lesions was observed only for NB4 tumors, with 5.3% injected dose per gram of tumor 24 hours after injection and a tumor-to-blood ratio of 18.8. For this reason, the NB4 model was selected for subsequent therapy experiments.

Therapy experiments in mice

Nude mice bearing subcutaneously grafted NB4 tumors (~50 mm3) were treated with intravenous injections of saline, F8-IL2, or KSF-IL2 (used as negative control immunocytokine) on days 7, 10, and 13 (fig. S4, A and B). Substantial tumor growth retardation was observed up to day 23 in the case of F8-IL2 (but not KSF-IL2).

In a second experiment, immunocytokine therapy was compared to the action of cytarabine (100 mg/kg daily for five consecutive days). Cytarabine is frequently used at this dose and schedule in preclinical experiments (43), but we also studied the effect of the drug in a preliminary therapy experiment in our AML model (fig. S4, D and E). Furthermore, the effect of F8-IL2 (30 μg of F8-IL2 every third day, four injections in total) in combination with cytarabine (100 mg/kg daily for five consecutive days) was studied (Fig. 3, E and F). Administration of F8-IL2, but not KSF-IL2 or cytarabine, significantly delayed tumor progression up to day 27. In addition, the combination of F8-IL2 with cytarabine led to long-lasting complete tumor eradication in 40% of the treated mice.

In some cancer models, the antitumor activity of IL-2 is primarily mediated by T cells rather than by NK cells (42). For this reason, the effect of F8-IL2 was studied also in an immunocompetent mouse model (C57BL/6), bearing subcutaneously grafted C1498 tumors (~75 mm3). Immunofluorescence analysis of the established chloromas revealed intense staining with the F8 antibody and moderate staining with L19 (Fig. 4A). Biodistribution experiments, performed as described above, showed a selective accumulation of F8-IL2 in the neoplastic lesions, with 3.1% injected dose per gram of tumor 24 hours after injection and a tumor-to-blood ratio of 8.7 (Fig. 4B). In this setting, F8-IL2 (30 μg every third day, three injections in total) mediated potent tumor growth retardation up to day 30. The combination of F8-IL2 and cytarabine (100 mg/kg daily for five consecutive days) led to complete tumor eradication in 100% of the mice (Fig. 4, C and D).

Fig. 4 F8-IL2 treatment significantly reduces tumor progression and can promote complete tumor eradication in combination with cytarabine in subcutaneous murine AML model.

The murine AML cell line C1498 was subcutaneously implanted in C57BL/6J mice. (A) Immunofluorescence analysis of the resulting tumors showed an intense staining with the F8 antibody and weaker staining with L19 (red). Blood vessels were costained using an anti-CD31 antibody (green). Scale bars, 100 μm. (B) Biodistribution experiments were performed by intravenous injection of radioiodinated F8-IL2 (black bars) or KSF-IL2, used as negative control of identical molecular format (white bars). F8-IL2 displayed 3.1% injected dose per gram of tumor 24 hours after injection and a tumor-to-blood ratio of 8.7 (n = 5). C57BL/6J mice bearing subcutaneously grafted C1498 tumors (~75 mm3) were treated with intravenous injections (n = 5) of either saline (×), F8-IL2 (▪) (30 μg, days 7, 10, and 13), KSF-IL2 (□) (same dose and schedule), or cytarabine (▴) (100 mg/kg daily, days 7 to 11). Furthermore, the effect of F8-IL2 in combination with cytarabine (▵) was studied. (C and D) The combination of F8-IL2 with cytarabine led to long-lasting complete tumor eradication in 100% of the treated mice [**, significant after day 11 for F8-IL2 versus KSF-IL2 (P = 0.0018) and saline (P = 0.0015); #, significant for combination group versus KSF-IL2 (P = 0.0024) and saline (P = 0.0016) after day 11, n = 5, two-tailed Student’s t test]. Data represent mean tumor volumes ± SD. Arrows depict days of treatment with immunocytokines (black) and cytarabine (gray). (E) Monitoring of the body weight showed that the combination treatment reduced average body weight, without exceeding the tolerated body weight loss of 15%. Data represent mean percent body weight relative to first day of therapy ± SD. (F and G) In vivo immunodepletion experiments were performed in C57BL/6J mice bearing subcutaneously grafted C1498 tumors (n = 5). F8-IL2 promoted significant tumor growth retardation in undepleted (▪) and CD4+ T cell–depleted (○) mice, whereas animals with CD8+ T cell (•) or NK cell depletion (□) displayed a tumor progression comparable to that of undepleted and saline-treated (×) mice. Data represent average tumor volumes ± SD. Arrows depict days of treatment with immunocytokines (black) and antibodies for depletion (gray). (H) No body weight loss was observed. Data represent mean percent body weight relative to first day of therapy ± SD.

In vivo depletion experiments, performed in C57BL/6J mice bearing subcutaneously grafted C1498 tumors, showed that the therapeutic activity of F8-IL2 was completely conserved after depletion of CD4+ T cells, whereas tumor growth inhibition was lost as a result of NK cell or CD8+ T cell depletion (Fig. 4, F and G).

Clinical evaluation of F16-IL2 in combination with low-dose cytarabine

In an AML patient with rapidly progressing, generalized chloroma disease, an individual treatment was attempted using F16-IL2 in combination with low-dose cytarabine. The patient had previously experienced multiple relapses and had received allogeneic stem cell transplantations from two different unrelated donors. A pretherapeutic 18-FDG-PET/computed tomography (CT) scan revealed multiple hypermetabolic thoracic and abdominal chloroma nodules, with standard uptake values of up to 14.1 (Fig. 5). A lesion located in the hilum of the liver leading to cholestasis and a deep cervical mass leading to difficulties in swallowing and vessel compression had to be simultaneously irradiated. However, the patient presented multiple other extramedullary AML lesions in areas not targeted by radiotherapy. PET/CT images acquired on day 14 after therapy initiation with F16-IL2 (30 MIU of IL-2 equivalents intravenously on day 1, 50 MIU on day 8) and low-dose cytarabine (5 mg twice daily subcutaneously on days 1 to 10) showed a nearly complete metabolic response of both the irradiated and nonirradiated lesions after systemic treatment, which was accompanied by a partial morphological response in CT scans (Fig. 5). Notably, the clinical symptoms such as difficulties in swallowing and restricted head mobility improved shortly after systemic therapy was initiated, even before the first application of radiotherapy.

Fig. 5 PET/CT images of an AML patient with multiple chloroma manifestations treated with the immunocytokine F16-IL2 in combination with low-dose cytarabine.

A 51-year-old female patient with refractory AML after two allogeneic stem cell transplantations and rapidly progressing, disseminated extramedullary AML manifestations was treated with F16-IL2 (30 MIU intravenously on day 1, 50 MIU on day 8) and low-dose cytarabine (5 mg twice daily subcutaneously on days 1 to 10). 18-FDG-PET/CT images were acquired before (day 0) and on day 14 after initiation of therapy. The two lesions marked with an arrow were irradiated simultaneously to systemic therapy because of local complications (cholestasis, swallowing difficulties, and vessel compression). Clinical improvement started 1 day after first infusion of F16-IL2 and before start of radiotherapy. A nearly complete metabolic response of irradiated and nonirradiated extramedullary AML lesions could be documented as early as on day 14.

DISCUSSION

The initial chemotherapy in AML comprises a first phase of induction and a second phase of consolidation. In most of the patients, the induction treatment leads to CR, defined as microscopic disappearance of leukemic disease along with the return of normal hematopoiesis. However, despite the introduction of more efficacious consolidation regimens, a large proportion of AML patients in CR will subsequently experience relapses with poor prospects of long-term survival. A relapse is assumed to be the result of expansion of residual leukemic cells that have escaped the initial chemotherapy. The antileukemic function of T cells and NK cells has formed the background for the clinical use of IL-2, with the aim to eliminate residual leukemia and hence reduce the relapse rate in AML. Results of clinical trials using IL-2 monotherapy in AML patients have been disappointing (44), but a recent phase 3 study has demonstrated that postconsolidation treatment with the combination of histamine dihydrochloride and IL-2 significantly prevents relapse in AML patients (45).

There is a strong preclinical and clinical rationale suggesting that tumor-homing immunocytokines can display a superior anticancer activity compared to nontargeted IL-2 (9, 12). Vascular-targeting immunocytokines have progressed to phase 2 clinical trials for the treatment of patients with solid tumors (9, 12, 26) and have recently been evaluated preclinically for the therapy of lymphomas (27). Here, we show that the clinical-stage F8 and F16 antibodies (and, to a lower extent, L19) stain a considerable fraction of AML and ALL bone marrow biopsies as well as AML chloromas in immunohistochemical analysis. We report the therapeutic effect of the immunocytokine F8-IL2, alone and in combination with cytarabine, in subcutaneous models of AML. F8-IL2 treatment promoted significant tumor growth retardation and synergized with cytarabine, allowing the complete eradication of tumors that could not be cured with cytarabine alone.

Complete leukemia eradication was only observed in the syngeneic immunocompetent mouse model of AML, but not in immunodeficient mice bearing human leukemia cells, suggesting that T cells were necessary for cancer eradication. In vivo depletion experiments indicated that the antitumor effect of F8-IL2 was mediated by both NK cells and CD8+ T cells, whereas CD4+ T cells did not appear to significantly contribute to the therapeutic action in the C1498 model. Leukemia cells appear to be particularly sensitive to the action of antibody-based IL-2 delivery, because complete tumor eradications were never observed in the past, using IL-2–based immunocytokines in other models of cancer (27, 33, 4648). In a clinical perspective, it is important to consider that lymphocyte and NK cell counts are essentially normal in AML patients, which are otherwise often leukopenic (49). Thus, both NK cells and CD8+ T cells should be available for promoting the anticancer effect of IL-2–based immunocytokines.

In the setting of compassionate use, a heavily pretreated AML patient with rapidly progressing, generalized chloroma disease was treated with a combination of F16-IL2 and low-dose cytarabine. PET/CT images before and on day 14 after therapy initiation revealed marked reduction of the AML lesions. These findings illustrate the potential of the clinical application of immunocytokines in combination with cytarabine. In principle, such results could theoretically have been achieved with cytarabine alone, yet in our opinion, the rapid and nearly complete response to combination therapy in a patient who has failed multiple lines of cytarabine-based chemotherapy in the past and who has now received a very low dose of cytarabine is more supportive of a significant contribution of the immunocytokine. Furthermore, being after allogeneic transplantation, the donor cellular immune system might have added to the efficacy of the immunocytokine. However, although the patient’s chronic skin graft-versus-host disease slightly worsened upon therapy with F16-IL2, the most relevant side effects were elevated body temperature and transient pain at the tumor sites after each application with otherwise good tolerability, suggesting a targeted therapeutic immune effect.

Bone marrow specimens of one AML patient showed positive staining both at presentation and after CR because of induction therapy. Together with the observation that the bone marrow sample of an AML patient in aplasia was strongly stained by F8 and F16, this suggests that the antigen expression persists after chemotherapy (fig. S5). These findings are somewhat unexpected when taking into account the reports by Padró et al. (38), which show that AML patients enjoying a CR after induction therapy exhibited a reduction of microvessel density to values comparable to the ones of control subjects. However, antigen persistence could reflect the high chemical stability of extracellular matrix components (such as fibronectin and tenascin splice isoforms) generated during neovascularization processes. This feature is likely to favorably influence the development of armed antibody therapeutics, which can persist at the site of disease for several days.

A clear limitation of our study is the lack of therapy experiments performed in systemic disease, which involves the bone marrow. We attempted to reproduce the promising targeting results that were obtained with chloroma models, in orthotopic mouse models of AML with florid proliferation of blasts in the bone marrow. Unfortunately, the models that we investigated so far did not express the EDA domain of fibronectin, in contrast to what is observed in human AML specimens. In general, IL-2–based immunocytokines have previously been used to eradicate established neuroblastoma metastases in the bone marrow (50). This observation indicates the use of immunocytokines for the therapeutic delivery of IL-2 to the diseased bone marrow in other pathologies, such as leukemias.

F8-IL2 has not yet been studied in clinical trials, but F16-IL2 and L19-IL2 have been extensively studied in more than 200 patients with cancer (9, 12). F16-IL2 has been studied in two phase 1b trials in combination with paclitaxel or with doxorubicin (45). L19-IL2 is currently being tested for the treatment of metastatic melanoma in a phase 2b study in combination with dacarbazine (26). It has also been studied as monotherapy treatment for patients with renal cell carcinoma (51) and is completing a phase 1b trial in combination with gemcitabine in patients with pancreas cancer. Both immunocytokines have shown promising safety profiles and high combinability in patients (26, 51).

The incidence of AML increases with age, and there is an unmet medical need for the treatment of elderly patients, who do not respond to approved therapeutic modalities and who do not tolerate aggressive chemotherapeutic regimens. The side effects of IL-2–based immunocytokines are generally mild, when these products are used at doses up to 67.5 MIU per week (26, 51). The management of patients becomes more problematic, when high-dose IL-2 regimens are used, with repeated administrations of >40 MIU doses to young patients in the intensive care setting (52).

The strong and selective antigen expression in acute leukemia, the emerging use of IL-2–based immunocytokines in clinical trials, as well as the orthogonal profiles of side effects with F8/F16-IL2 and cytarabine provide a rationale for the combined use of these agents for the treatment of AML patients (particularly those who are not eligible for bone marrow transplantation or suffer from chloroma disease).

MATERIALS AND METHODS

Study design

Immunohistological analysis of AML patient biopsies was performed with the clinical-stage human monoclonal antibodies F8, F16, and L19. The use of freshly frozen material was made necessary by the fact that these antibodies do not work in paraffin (36, 41), which prevented the use of larger collections of formalin-fixed, paraffin-embedded specimens. Samples were collected at three medical centers (Muenster, Berne, and Kiel) for routine histological and cytological analyses. Semiquantitative immunofluorescence was performed on the largest sample set (n = 17), and the staining was additionally verified by immunohistochemistry of consecutive sections of the same samples. Median patient age was 61 years (20 to 82 years). In case of AML, the study cohort represented the most frequent FAB subtypes; however, the subtypes M3, M6, and M7 were not available for analysis. The ALL samples included two common ALL and two T-ALL. The bone marrow of AML and ALL patients was highly infiltrated by leukemic blasts (median, 80%). Tumor-bearing mice were generated by subcutaneous injection of AML cells to analyze the therapeutic effect of F8-IL2 (because the F16 antibody, which showed the strongest staining in patient biopsies, is not cross-reactive to mouse) alone and in combination with cytarabine. Initial mouse experiments (model setup, preliminary biodistribution studies, and dose-finding experiments) were performed with three mice per group to limit the use of animals. Mouse models, which were selected for therapy experiments, were tested in targeting studies with n = 6 (NB4) and n = 5 (C1498). Therapy experiments were performed with five mice per group. The therapy endpoint was defined when tumors reached a set volume [>1000 mm3 (NB4) or 1200 mm3 (C1498)]. Animals with tumors of 20 to 100 mm3 on first day of therapy were included in experiments and staged to maximize uniformity among the groups. The therapeutic effect of F8-IL2 was studied in two mouse models, with two experiments performed in each model. The clinical application of F16-IL2 in combination with cytarabine was tested on the basis of compassionate use. The preliminary clinical data shown here represent the finding in one AML patient.

Tissues

Bone marrow core biopsy and bone marrow aspiration (iliac crest) were obtained for routine histological and cytological analyses. A fragment of each bone marrow biopsy was embedded in cryoembedding medium and immediately frozen at −80°C. Two sets of specimens were analyzed. The larger set from Muenster consists of bone marrow biopsies of 16 patients. The sample set from Berne and Kiel consists of four bone marrow biopsies from AML patients and two chloroma biopsies. Additional patient characteristics can be found in table S1.

Cell lines, animals, and xenograft models

The human AML cell line HL60 and the murine AML cell line C1498 were purchased from the American Type Culture Collection. The human AML cell lines NB4 and THP1 were obtained from the German Resource Center for Biological Material. Cell lines were cultured according to the supplier’s recommendations. Six- to 8-week-old female CB17/lcr SCID and BALB/c nude mice were purchased from Charles River Laboratories. Six- to 8-week-old female C57BL/6J mice were purchased from Elevage Janvier. For the localized xenograft (chloroma) models, 107 HL60, NB4, or THP1 cells were subcutaneously injected into the flank of 8- to 10-week-old BALB/c nude mice. C1498 cells (106) were injected into the flank of 8- to 10-week-old C57BL/6J mice. All animal experiments were performed on the basis of project license (198/2008) administered by the Veterinäramt des Kantons Zuerich and approved by all participating institutions.

Antibodies and therapeutic agents

F8 is a human monoclonal antibody specific to the EDA domain of fibronectin (34). The expression and characterization of the F8-IL2 immunocytokine, as well as the control immunocytokine KSF-IL2, have previously been described (40). F16 is a tumor-targeting antibody specific to the domain A1 of human tenascin-C (32). L19 is a human monoclonal antibody specific to the EDB domain of fibronectin (35). KSF is specific to hen egg lysozyme and does not show any specificity toward human antigens (42). All antibodies were biotinylated in the SIP format and carried a comparable number of biotin molecules. Cytarabine was purchased in solution from Sandoz.

Immunofluorescence

Tissue samples were snap-frozen, embedded in cryoembedding medium (Richard-Allan Scientific Neg-50; Thermo Scientific), and then stored at −80°C. Consecutive tissue sections of 10-μm thickness were prepared with a Microm HM 505N. Immunofluorescence was performed as previously described (31). For staining of the patient samples from the Muenster set, the following primary antibodies were used: biotinylated F8, L19, F16, or KSF (2 μg/ml) and polyclonal rabbit anti-human vWF (Dako). For the tumor and bone marrow sections of the mouse models, the following primary antibodies were used: biotinylated F8, L19, F16, or KSF (2 μg/ml) and monoclonal rat anti-CD31 antibody (1.6 μg/ml) (BD Biosciences). Detection of the biotinylated antibodies was performed with streptavidin Alexa Fluor 594 (Invitrogen). As secondary detection antibody for the anti-vWF antibody, Alexa Fluor 488 goat anti-rabbit (Invitrogen) antibody was used. The anti-CD31 antibody was detected with Alexa Fluor 488 donkey anti-rat (Invitrogen). The samples were analyzed on an Axioskop2 mot plus microscope (Zeiss) with Zeiss AxioVision v4.0 acquisition software.

Semiquantitative analysis of immunofluorescence

The immunofluorescence stainings of leukemia patient biopsies were semiquantitatively analyzed with ImageJ (http://rsb.info.nih.gov/ij/) by computing the percentage area of staining for each antibody. The values were plotted as a boxplot (SigmaPlot).

Immunohistochemistry

Immunohistochemistry of leukemia patient samples was performed as previously described (29). Biotinylated F8, L19, and F16 (2 μg/ml) were used as primary antibodies. For the analysis of the Muenster samples, KSF (2 μg/ml) was used as a negative control. For the Berne/Kiel samples, negative control was performed by omitting the primary antibody. For detection, streptavidin–alkaline phosphatase complex (1:1000 dilution) (Biospa) was used followed by subsequent reaction with the phosphate substrate Fast Red TR Salt (Sigma-Aldrich). Sections were counterstained with hematoxylin solution Gill No. 1 (Sigma-Aldrich), mounted with Glycergel mounting medium (Dako), and analyzed on an Axiovert S100 TV microscope (Zeiss) with Zeiss AxioVision v4.0 acquisition software.

Quantitative biodistribution studies

F8-IL2 and KSF-IL2 were radioiodinated with 125I, purified, and validated for immunoreactivity as previously described (46). Radiolabeled F8-IL2 or KSF-IL2 (20 μg of antibody, 20 μCi per mouse) was injected intravenously into BALB/c nude mice bearing subcutaneous NB4 (n = 6), HL60 (n = 3), or THP1 (n = 3) tumors, or C57BL/6 mice bearing subcutaneous C1498 (n = 5) tumors. Mice were sacrificed after 24 hours, and the organs were excised and weighed. Radioactivity was measured with a Packard Cobra gamma counter. Radioactivity content of representative organs was expressed as percentage of the injected dose per gram of tissue (%ID/g ± SE).

Therapy studies in localized xenograft (chloroma) models

NB4 cells (107) were injected subcutaneously into the flank of 6- to 8-week-old female BALB/c nude mice. C1498 cells (106) were injected subcutaneously into the flank of 6- to 8-week-old female C57BL/6 mice. When tumors were established (20 to 100 mm3), mice were staged to maximize uniformity among the groups (n = 5) and injected into the lateral tail vein with either saline, 30 μg of F8-IL2, 30 μg of KSF-IL2, cytarabine (100 mg/kg), or a combination of F8-IL2 and cytarabine (at same dosage). Treatment schedule for the immunocytokines was every third day for four and three injections, respectively. Cytarabine was administered daily for 5 days. The mice were monitored daily, and tumor growth was measured every second day with a digital caliper using the following formula: volume = length × width × width × 0.5. Animals were sacrificed when the tumor reached a volume of >1200 mm3. The data were displayed as average values ± SD.

In vivo depletion of NK cells, CD4+, and CD8+ T cells

Subcutaneous C1498 tumors were established in C57BL/6 as described above. Tumor-bearing mice (n = 5) were injected intraperitoneally with either saline, anti–asialo GM1 (30 μl) (Wako Chemicals), anti-CD4 (Bio X Cell), or anti-CD8 (Bio X Cell) antibodies on days 6, 9, 12, and 15. F8-IL2 (30 μg) was administered intravenously on days 7, 10, and 13. The mice were monitored as described above. The data were displayed as average values ± SD.

Treatment of a chloroma patient

A female patient was diagnosed with complex karyotype AML FAB M1 in February 2007 at 45 years of age. Upon being refractory to induction chemotherapy with cytarabine and daunorubicin (7 + 3) and high-dose cytarabine and mitoxantrone (HAM), she received unrelated donor allogeneic stem cell transplantation after conditioning with fludarabine, cytarabine, amsacrine, cyclophosphamide, anti-thymocyte globulin, and 2 × 2 Gy total body irradiation in October 2007, whereupon she enjoyed a first CR. In July 2008, she experienced AML relapse, and therapy with low-dose cytarabine (persistent AML), high-dose cytarabine, and an allogeneic stem cell boost resulted in a second CR, which lasted more than 2.5 years. However, in May 2011, a second AML relapse was diagnosed. Although chemotherapy with HAM and a donor stem cell boost were ineffective, allogeneic stem cell transplantation from a different unrelated donor led to a short-lasting CR until January 2012, when the third relapse occurred. Chemotherapy with FLAG-IDA and an additional donor stem cell boost resulted in another phase of CR. In November 2012, however, extramedullary AML manifestations were diagnosed and confirmed histologically in both mammary glands (whereas the bone marrow was still leukemia-free with stable donor cell chimerism) and palliative local radiotherapy (37 Gy) was performed. Three months later, multiple chloroma nodules occurred in the abdominal skin and were again treated with local radiotherapy (36 Gy) until the beginning of May 2013. Three weeks later, intra- and extrahepatic cholestasis led to the diagnosis of rapidly progressing, disseminated chloroma disease with manifestations in the left deep cervical area, in the mediastinum, in the hilum of the liver, in the mesenterium, in the hilum of the right kidney, in the manubrium sterni, in the transverse process of the ninth thoracic vertebra, and subcutaneously in the left gluteal area. In this desperate situation, our patient received the immunocytokine F16-IL2 (30 MIU intravenously on day 1, 50 MIU on day 8) and low-dose cytarabine (5 mg twice daily subcutaneously on days 1 to 10). Ethical approval for this individual compassionate use therapy was obtained from the joint ethical board of the University of Muenster and the locoregional physician’s chamber of Westfalen-Lippe. Written informed consent was obtained in accordance with the Declaration of Helsinki. A phase 1b study with F16-IL2 in combination with chemotherapy is running at multiple centers (NCT01134250, NCT01131364) and had shown that doses up to 55 MIU can be safely administered to patients (53).

Statistical analysis

The staining of bone marrow biopsies of AML patients with F8, F16, and L19 was analyzed compared to the negative control antibody KSF. Values exceeding the 95% confidence interval (average + 2 × SD) of KSF (3.4% for AML samples; 2.6% for ALL samples) were considered positive.

In therapy experiments in mice, the differences in tumor volume between therapeutic groups were compared with the two-tailed Student’s t test with α = 0.05. Taking the Bonferroni correction into account, significance was determined when P ≤ 0.003.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/5/201/201ra118/DC1

Materials and Methods

Fig. S1. Splice isoforms of oncofetal fibronectin (EDA and EDB) and tenascin-C are expressed in bone marrow biopsies of acute leukemia patients.

Fig. S2. Splice isoforms of oncofetal fibronectin (EDA and EDB) and tenascin-C are expressed in bone marrow and chloroma biopsies of acute leukemia patients.

Fig. S3. EDA is not expressed in orthotopic models of leukemia, despite disease progression in blood and bone marrow.

Fig. S4. Treatment of subcutaneous NB4 tumors with F8-IL2 promotes tumor growth retardation.

Fig. S5. Antigen expression persists in bone marrow biopsies of AML patients in complete response and aplasia.

Table S1. Patient characteristics of the sample set from Muenster.

Table S2. Patient characteristics of the sample set from Berne and Kiel.

REFERENCES AND NOTES

  1. Funding: This work was supported by the Swiss National Science Foundation, the ETH Zuerich, the European Union (ADAMANT Project and IMMOMEC Project), the Swiss Cancer League, the Swiss Bridge Foundation, and the Stammbach Foundation. Author contributions: K.L.G. conceived and performed the experiments and wrote the manuscript; C.S. conceived the experiments, performed initial immunohistochemical experiments, and contributed to the manuscript; L.G. supplied F16-IL2 for clinical application; K.F. produced the fusion proteins F8-IL2 and KSF-IL2 and helped in the generation of the orthotopic model; T.P. and W.K. collected, characterized, and provided human leukemia specimens and corrected the manuscript; W.E.B. collected, characterized, and provided human leukemia specimens and contributed to the manuscript; D.N. conceived the experiments, supervised the experimental work, and wrote the manuscript. Competing interests: D.N. is a co-founder and shareholder of Philogen. L.G. is the head of clinical operations at Philogen (http://www.philogen.com), the company that owns the F8, F16, and L19 antibodies. The other authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article