We have previously shown that talactoferrin-alfa (TLF), a recombinant human lactoferrin, is an immunomodulatory protein that is active against implanted tumors, both as a single agent and in combination with chemotherapy. In this study, we show that talactoferrin is active against autochthonous tumors in a transgenic mouse line, which is more analogous to human cancers, and identify key mechanistic steps involved in the anticancer activity of oral TLF. BALB/c mice transgenic for the rat neu (ErbB2) oncogene (BALB-neuT) treated with oral TLF showed a significant delay in carcinogenesis, with 60% tumor protection relative to vehicle-treated mice at week 21. Oral TLF also showed tumor growth inhibition in wild-type BALB/c mice implanted with neu+ mammary adenocarcinoma, with one third displaying a long-lasting or complete response. Oral TLF induces an increase in intestinal mucosal IFN-γ production and an increase in Peyer's patch cellularity, including expansion of CD8+ T lymphocytes and NKT cells, and the enhancement of CD8+ T-cell cytotoxicity. In IFN-γ knockout mice, there is an absence of the TLF-induced Peyer's patch cellularity, no expansion of CD8+ T lymphocytes and NKT cells, and loss of TLF anticancer activity. TLF antitumor activity is also lost in mice depleted of CD8+ T cells and in CD1 knockout mice, which lack NKT activity. Thus, the inhibition of distant tumors by oral TLF seems to be mediated by an IFN-γ–dependent enhancement of CD8+ T- and NKT cell activity initiated within the intestinal mucosa. [Cancer Res 2007;67(13):6425–32]
- mammary adenocarcinoma
- neu (ErbB2)
In recent years, immunotherapeutic approaches to treating cancer, with either adoptive transfer of immunity or stimulation of the endogenous immune system, have shown increasing promise. Talactoferrin-alfa (TLF), an immunomodulatory agent currently in late-stage cancer clinical trials, acts through a novel mechanism of action. Talactoferrin is a recombinant human lactoferrin (rhLF) produced by fermentation in Aspergillus ( 1). Lactoferrin is an important endogenous immunomodulatory protein, belonging to the transferrin family of iron-binding glycoproteins ( 2). Endogenous lactoferrin is broadly distributed within the body and is found in exocrine secretions (e.g., milk, tears, and saliva; ref. 3) and in neutrophil-specific granules ( 4). Orally administered human lactoferrin (hLF) has been shown to have anti-infective and anticancer activity in animal models ( 5). hLF has been shown to enhance host protection against a broad range of infections ( 6– 8). We have previously shown that oral TLF inhibits the growth of implanted tumors at distant sites ( 9) and potentiates the antitumor activity of cisplatin ( 10). However, little is known about the anticancer and immunomodulatory activity of oral TLF against autochthonous tumors. In this study, we examined the influence of oral TLF on spontaneously arising mammary carcinomas in rat neu (ErbB2) transgenic mice. This transgenic mouse model bears mammary carcinomas that become evident after the progressive stage of tumorigenesis, thereby providing a relationship between the tumors and surrounding tissue ( 11). We show that oral TLF has antitumor activity in transgenic mice.
Orally administered hLF has been shown to have systemic immunostimulatory effects including (a) stimulation of Th1 cytokine response in splenocytes and lymph node cells ( 12– 14), (b) boosting of effector activity in peritoneal macrophages and splenic natural killer (NK) cells ( 15), and (c) enhancement of leukocyte numbers in blood and lymphoid tissues ( 16).
Oral TLF primarily targets the intestinal epithelial cells and intestine-associated immune system ( 17) and is not absorbed systemically ( 18– 21). Oral lactoferrin also enhances interleukin-18 (IL-18) production ( 10, 19), a cytokine with pleiotropic effects on immune cell activation and function ( 22). However, the effect of oral lactoferrin on the intestinal and systemic immune system has not been fully elucidated.
Here, we show in wild-type BALB/c mice bearing implanted neu+ mammary adenocarcinomas that chronic oral TLF triggers significant tumor inhibition, and one third of treated mice display a long-lasting or complete response. Oral TLF seems to initiate an immune activation in the gut and the gut-associated lymphoid tissue (GALT). This results in an IFN-γ–dependent increase in number of NKT cells and CD8+ T lymphocytes in small intestinal Peyer's patches and systemic CTL activity. By using knockout or immunodepleted mice, we have observed that TLF anticancer activity is IFN-γ dependent and requires CD8+ T lymphocyte and NKT cell activity.
Materials and Methods
Mice. Virgin female BALB/c mice transgenic for rat neu oncogene expressed under the control of mouse mammary tumor virus promoter (BALB-neuT mice; ref. 23), wild-type BALB/c mice, BALB/c mice IFN-γ gene knockout (BALB-IFNγKO; ref. 24), and CD1 gene knockout (BALB-CD1KO; ref. 25) were bred in our animal facility. Mice were randomly assigned to control and treatment groups and were treated concurrently. Mammary glands of BALB-neuT mice ( 23) and the challenge site of wild-type and KO BALB/c mice ( 26) were inspected at weekly intervals to note tumor appearance. Tumor masses were measured with caliper in two perpendicular diameters in a blind fashion to determine the longest diameter. Progressively growing masses (>2-mm mean diameter) were regarded as tumors. In BALB-neuT mice, tumor multiplicity was calculated as the cumulative number of incident tumors per total number of mice and reported as mean ± SE ( 23). Mice were treated according to the European Union guidelines. CD8 lymphocyte depletion was done by i.p. injections of 0.2 mL of HBSS containing 500 μg anti-CD8 monoclonal antibody (mAb; TIB-105 hybridoma, Lyt-2; American Type Culture Collection) or vehicle, 1 day before and 2, 5, 8, 11, 14, and 17 days after tumor challenge (day 0). Flow cytometry of residual blood collected 3 days after the last injection showed that CD8+ T cells were decreased to a level below 1 in 5,000 cells.
Cell line and tumor challenge. TUBO cells are a cloned line generated from a BALB-neuT mouse mammary gland carcinoma and express large amounts of neu protein ( 11). TUBO cells were cultured in DMEM with 20% fetal bovine serum (FBS). Mice were challenged s.c. in the left flank with 0.2 mL of a single suspension of a minimal lethal dose of TUBO cells (1 × 105) in HBSS ( 27) and monitored for 105 days. At the end of this period, tumor-free mice were classed as survivors, whereas those with tumors were sacrificed. Mice with a tumor mass exceeding 10 mm in mean diameter were sacrificed for humane reasons.
TLF administration. TLF was provided in a phosphate-buffered solution [6 mmol/L sodium phosphate monobasic monohydrate, 9 mmol/L sodium phosphate dibasic heptahydrate, 50 mmol/L sodium chloride (pH 7.0 ± 0.5)] at a concentration of 100 ± 10 mg/mL. It was stored frozen at −20°C before dosing. Mice were randomly divided into experimental groups and treated orally by gavage with TLF (1,000 mg/kg, 3 mg/m2) or vehicle. In mice with implanted tumors, TLF administration occurred once daily, 5 days per week for 3 weeks, except when otherwise specified. In the transgenic mice experiments, TLF was administered once daily, 5 days per week, 3 weeks on and 1 week rest. TLF administration was started on week 6 and continued until week 30.
Counts and flow cytometry of Peyer's patch cells. Three days after oral TLF ingestion, mice were sacrificed, and the first three Peyer's patches of small intestinal epithelium were collected, and a single-cell suspension was obtained and counted in a blind fashion using a Burker's chamber. Cells (3 × 105) excluding trypan blue (Life Technologies/Invitrogen) were stained immediately with mAb anti-CD8a phycoerythrin (PE)–conjugated (Cedarlane), anti-CD49b PE (PharMingen), and anti-CD3 FITC (PharMingen). After 30 min of incubation at 4°C, the cells were washed in Dulbecco's modified PBS supplemented with 0.1% sodium azide and 2% FBS and analyzed with a CyAn ADP machine (DakoCytomation) using Summit 4.2 software (DakoCytomation).
In vivo cytotoxicity assay. In vivo cytotoxicity assay was done as previously described ( 28). Briefly, a single-cell suspension of 10 × 106 naive spleen cells (Spc) per milliliter was labeled with two different concentrations (0.5 μmol/L, CFSE low or 5.0 μmol/L, CFSE high) of the fluorescent dye CFSE (Molecular Probes). Spc labeled with 5.0 μmol/L were also pulsed with the dominant 9-mer 63-71 peptide of the protein product of rat neu oncogene with H-2Kd restriction element ( 29) for 1 h at room temperature. The two Spc populations were mixed together in equal amounts and injected i.v. into control and treated mice that had received or did not receive TUBO cells s.c. Mice were sacrificed 48 h later, and single-cell suspensions from spleens were processed individually to evaluate CFSE high and CFSE low cells by CyAn ADP (DakoCytomation). Specific cytolytic activity was calculated as follows: 100 × (percentage CFSE low cells − percentage CFSE high cells) / percentage CFSE low cells ( 28).
IFN-γ production and intracellular staining. Small intestinal epithelial tissue was collected and homogenized as described by Kuhara et al. ( 30). Briefly, a known weight of small intestinal tissue was placed in a lysis buffer consisting of PBS, 1% NP40, 0.5% Na deoxycholate, and 0.1% SDS containing 10 mg/mL of phenylmethylsulfonyl fluoride, a general protease inhibitor (Sigma-Aldrich). The tissue was mechanically homogenized, and the suspension was centrifuged at 15,000 rpm for 10 min. The clear supernatant was collected in 100-μL aliquots and stored at −80°C until use. An ELISA kit (R&D Systems) was used to determine the IFN-γ content of the aliquots.
The Peyer's patches were incubated for 30 min at 4°C with directly conjugated FITC anti-CD8 (Cedarlane). The cells were washed with Dulbecco's modified PBS with 0.1% sodium azide and 2% FBS. Cell pellets were resuspended in 1 mL of Fix/Perm (eBioscience), and the samples were incubated overnight at 4°C. After two washes with Permeabilization buffer (eBioscience), the samples were incubated at 4°C with 2 μL of Fc receptor blocker (PharMingen) for 15 min followed by PE anti-IFN-γ (PharMingen) for 30 min. The cells were washed twice before conducting flow cytometric analysis. Single- and double-stained samples were used for instrumentation setup. Flow cytometric analysis was done with a CyAn ADP machine (DakoCytomation) using Summit 4.2 software (DakoCytomation).
Statistics. Differences in tumor incidence were evaluated with the Mantel-Haenszel log-rank test. Tumor multiplicity, number of positive cells at flow cytometry, and ELISA were analyzed using the Student's t test (two tailed).
Oral TLF inhibits autochthonous carcinogenesis in neu transgenic mice. The neu transgenic mice used in this study have a 100% penetrance of mammary carcinogenesis. Mammary cells overexpressing neu first manifest atypical hyperplasia by about 4 weeks of age. In week 7, the hyperplastic condition progresses to multifocal preneoplastic lesions ( 11). At least one carcinoma (>2-mm mean diameter) is palpable in each untreated control mouse by week 21 ( Fig. 1A ), and by week 31, a tumor mass is palpable in all 10 mammary glands ( Fig. 1B). Oral TLF (3 mg/m2) significantly delayed cancer progression, extending both the mouse disease-free survival and reducing tumor multiplicity ( Fig. 1A and B).
Oral TLF inhibits the growth of implanted neu+ tumors. BALB-neuT mouse autochthonous mammary carcinogenesis is an inexorable but slow process and requires about 20 weeks to become clinically evident ( 11). In contrast, implantation of neu+ TUBO cells in wild-type BALB/c mice is a more rapid process. Implantation of a 100% lethal dose of neu+ TUBO cells results in the formation of a tumor mass (>2-mm mean diameter) in about 3 weeks ( 26). Even in the more rapidly growing TUBO model, TLF showed anticancer activity ( Fig. 2 ). In the vehicle-treated control mice, the tumors grew steadily, and all 14 mice were sacrificed by day 54 due to large tumor outgrowth ( Fig. 2). By contrast, a markedly delayed growth of TUBO tumor was evident in about half of the mice receiving oral TLF. In 27% of them, 105 days after challenge, the tumor masses were <3-mm mean diameter. This unique finding shows that oral TLF inhibits tumor proliferation, and that its effect persists for over 3 months after completion of dosing.
Peyer's patch reactivity associated with TLF administration. Oral TLF has been shown to have several immunologic effects ( 31), and we have tried to characterize the key mechanisms that play a critical role in inhibiting TUBO challenge. Because the GALT is an important target for oral TLF, we evaluated the cellularity of Peyer's patches in tumor-challenged mice after 3 days of TLF treatment. The number of cells in the Peyer's patches increased by about a third ( Fig. 3A ), suggesting a high proliferative activity following TLF administration. Although the relative percentage of B220+ B cells, CD4+ T cells, and CD49b+ cells only slightly changed (data not shown), the percentage of CD8+ T cells (5×; Fig. 3B) and CD49b+/CD3+ NKT cells (2.5×; Fig. 3C) increased substantially when compared with controls. The absolute number of CD8+ T cells was 0.5 to 1 × 105 in control mice and increased to 2.7 to 4.5 × 105 in TLF-treated mice, whereas that of CD49b+/CD3+ NKT cells increased from 1.3 to 2.3 × 104 to 4.3 to 7.2 × 104. No significant increase in CD8+ T-cell numbers was observed in mesenteric lymph nodes and spleen (data not shown).
Both CD8+ T and NKT cells are required for TLF-activated tumor inhibition. As CD8+ T lymphocytes and NKT cells are important effectors involved in immune-mediated killing of cancer cells ( 32, 33), we assessed whether these TLF-activated cell populations play a significant role in the inhibition of TUBO cell tumors. BALB/c mice either remained untreated ( Fig. 4C ) or were depleted of CD8+ cells through i.p. administrations of anti-Lyt-2 hybridoma (TIB-105 hybridoma) before TUBO cell challenge and during TUBO cell growth ( Fig. 4A). Oral TLF no longer showed any antitumor activity in the CD8+ T cell–depleted mice. A similar loss of TLF anticancer activity was observed in BALB/c mice KO for the CD1d gene, which is required for NKT cell development (ref. 31; Fig. 4B).
To confirm the role of CD8+ T lymphocytes in the anticancer response induced by oral TLF, we did an in vivo CTL assay. In vivo cytotoxicity was evaluated in both naive BALB/c mice treated for 2 weeks with TLF only and mice challenged with TUBO cells 2 days after the beginning of TLF administration. TLF alone did not enhance the cytotoxic response against the dominant rat neu peptide ( Fig. 4D). Following TUBO cell challenge, an anti-neu cytotoxic response was evident and was enhanced when the TUBO challenge was combined with TLF treatment.
Importance of IFN-γ in TLF-activated antitumor resistance. Because IFN-γ is a key cytokine involved in early immune events, we examined the effect of oral TLF on IFN-γ production. When mice were treated with TLF for 24 h, there was a 3-fold increase in intestinal IFN-γ ( Fig. 5A ). There was also an increase in CD8+ T cells with intracytoplasmatic IFN-γ in the Peyer's patches ( Fig. 5B). These data suggest that IFN-γ plays an important role in the immunostimulatory events leading to oral TLF anticancer activity. We used IFN-γ KO mice to confirm the role played by IFN-γ in mediating TLF anticancer activity. In the BALB-IFNγKO mice, the TLF-induced increase in Peyer's patch cellularity is no longer evident, and the proportion of CD8+ T cells and CD49b+/CD3+ NKT cells is no longer increased ( Fig. 3D–F). Furthermore, the anticancer activity mediated by oral TLF is also absent in BALB-IFNγKO mice ( Fig. 4C).
Our investigations have shown that oral administration of TLF significantly inhibits both the development of tumors in neu transgenic mice and the growth of neu+ implanted tumors. In an early event, oral TLF triggered the release of IFN-γ by the intestinal mucosa lymphocytes. This was followed by an IFN-γ–dependent increase in Peyer's patches cellularity and an expansion of CD8+ T cells and CD49b+/CD3+ NKT cells. The critical role played by IFNγ, NKT cells, and neu-specific CD8+ T lymphocytes in mediating TLF anticancer activity was shown using knockout mice, mice depleted of CD8+ T cells, and an in vivo cytotoxicity assay.
We have previously reported antitumor activity of oral TLF against a range of implanted tumors in mice ( 10). Furthermore, TLF has also shown apparent anticancer activity in human clinical trials including phase I/II clinical trials with a range of solid tumors ( 19) and phase II trials in renal cell cancer ( 34) and non–small-cell lung cancer ( 35). In the present study, we have also assessed the ability of oral TLF to inhibit the chronic but inexorable progression of autochthonous mammary lesions. BALB-neuT mice are transgenic for the rat neu transforming oncogene under the transcriptional control of the long terminal repeat sequence of the mammary tumor virus. The virgin BALB-neuT female mouse provides one of the most aggressive models of neu mammary carcinogenesis with autochthonous mammary tumors that spontaneously develop in situ and are hence much more analogous to human cancers than implanted tumors. These mice develop mammary carcinomas in all their mammary glands with 100% incidence. The atypical mammary hyperplasia condition is generated by cells overexpressing neu and is first evident at 4 weeks of age. It progresses to multifocal preneoplastic lesions around week 7, and multiple invasive carcinomas are pathologically detectable in each mammary gland by week 22 ( 36, 37). This progression is accompanied by the accumulation of CD4+CD25+Foxp3+ T regulatory cells in the draining lymph nodes and mammary lesions ( 28) and suppressive immature myeloid cell expansion in the blood and spleen ( 38). Oral TLF provided protection against tumor development in this autochthonous model, with a 20% increase in median tumor-free survival. Because neu-induced mammary carcinogenesis in this transgenic model is very similar to tumor progression in humans, TLF may be of value in inhibiting the progression of early cancerous lesions in patients in addition to its promise in treating established tumors ( 34, 35).
Cancer patients are usually immunosuppressed, with the tumor itself often generating immunosuppressive factors. Chemotherapy itself further damages immunologic function along with the other adverse events resulting from toxicity to rapidly growing cells. Because oral TLF does not manifest systemic adverse effects (it has been administered to over 475 people without a single drug-related serious adverse event), it promises a more favorable efficacy/toxicity trade-off than conventional chemotherapy. In addition, oral lactoferrin has been shown to help protect against common adverse events associated with chemotherapy including cytopenias ( 39) and gut toxicities ( 40).
Immunotherapeutic approaches have been viewed as an important additional modality in treating cancer, and restoration of immune function could help responders remain disease-free for extended periods of time. Unfortunately, approved approaches including IL-2 and IFN-γ infusion have been associated with limited efficacy and high toxicities. Because the complex immune system requires multiple humoral and cellular factors acting in concert to induce effective immunostimulation, it is not surprising that infusion of individual cytokines have limited efficacy. Furthermore, to achieve physiologically relevant concentrations of cytokines within the immune system, extremely high systemic levels are required, resulting in the severe toxicities associated with cytokine administration. In contrast, as expected with an orally administered protein, TLF is not systemically bioavailable ( 19, 21). Instead, TLF acts in the gut and GALT, the largest immune organ in the body and a physiologically relevant site of lactoferrin activity. In this work, we have identified some of the early steps involved in the ability of TLF to induce systemic anticancer activity through immunostimulation at the GALT.
We have previously shown that oral TLF induces IL-18 production within intestinal mucosa ( 10). In this work, we examined changes in other key cytokines and immune cells. Considering that IL-18 stimulates production of IFN-γ ( 41), a cytokine with great immunomodulatory activity, the finding that IFN-γ is a key early cytokine required for oral TLF anticancer activity was not unexpected. There was a marked increase in IFN-γ production in the intestinal mucosa and CD8+ T cells in Peyer's patches, and the TLF-induced increases in small intestinal Peyer's patch cellularity and CD8+ T and NKT cells numbers were abolished in IFN-γ knockout mice. The importance of these events was further supported by the loss of TLF antitumor activity in mice deficient in IFN-γ, NKT cells, and CD8+ T cells, consistent with the role played by these mediators. CD8+ T lymphocytes are known to be key cytotoxic effector cells mediating tumor killing and following polyclonal activation in the gut; CD8+ T lymphocytes may be expected to migrate to distant tumor draining lymph nodes and tumors to mediate tumor killing. Present data showing that TLF treatment leads to the generation of an enhanced cytotoxic activity against neu suggest that when cytotoxic T cells are activated by xenogeneic neu peptides, TLF enhances their specific cytotoxic activity. This enhancement may result from a direct activity of TLF on antigen poised CD8+ T cells. However, recent observations suggest that TLF also induces dendritic cell maturation. 3 TLF-differentiated dendritic cells loaded with tumor cell antigens may migrate to lymphoid organs (e.g., Peyer's patches and lymph nodes) and present neu peptides to CD8+ T cells. Alternatively, CD8+ T cells polyclonally triggered by TLF in Peyer's patches may receive effective differentiation signals from antigen-pulsed dendritic cells maturated in the presence of TLF or TLF-induced downstream cytokines ( Fig. 6 ).
Although CD8+ T cells are required for tumor inhibition, the loss of the protective activity of TLF in BALB-CD1KO mice suggests that NKT cells are also necessary for the antitumor effect of TLF. NKT cells may act as both effector and regulatory cells ( 42). NKT cells migrating to and accumulating within tumors play an effector role, whereas circulating NKT cells are also regulatory ( 43). The key role that NKT cells assume in the antitumor reaction triggered by TLF is not surprising as it mirrors the powerful antitumor activity displayed by NKT cells triggered by α-galactosylceramide ( 33). On the other hand, the disappearance of TLF-elicited protection against neu+ tumors in mice lacking either CD8+ T or NKT cells shows that these two cell populations with markedly different specificity are both required.
The characterization of these mechanisms fits in well with previous data suggesting IL-18 and IFN-γ release, along with T and NK cell activation, as possible key events in the antitumor activity of oral TLF ( 10, 30). It is not surprising to observe the role of TLF in innate and adaptive immunity because this is consistent with the physiologic role of lactoferrin. Transgenic mice expressing human lactoferrin exhibited enhanced Th1 immune polarization ( 13). Furthermore, lactoferrin found in colostrum and milk has been shown to play an important immunomodulatory role, including helping to establish the GALT and strengthen the Th1 axis systemically.
TLF is a promising new immunomodulatory molecule with a novel mechanism of action. The molecule seems to be safe and well tolerated and has exhibited anticancer activity against renal cell cancer, non–small-cell lung cancer, and other tumor types in preclinical work and in phase I/II and phase II human clinical trials. In this work, we identify some of the early events involved in TLF-induced immunomodulation and link them to TLF anticancer activity. Additional work is under way to identify additional critical roles that may be played by TLF, including dendritic cell maturation and enhancement of the cross-presentation of tumor antigens to CD8+ T cells.
Grant support: Italian Association for Cancer Research, Italian Ministries for the Universities and Health, University of Torino, Regione Piemonte Bando Reg. Ric. Sci. Applicata 04.
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.
We thank Dr. Preeti Ismail for her substantial help with a critical review and revision of the manuscript and Deborah Duke and Irene Merighi for their editorial and technical help.
↵3 M. Spadaro, manuscript completed.
- Received November 3, 2006.
- Revision received February 26, 2007.
- Accepted April 5, 2007.
- ©2007 American Association for Cancer Research.