Immunological and Nonimmunological Effects of Indoleamine 2,3-Dioxygenase on Breast Tumor Growth and Spontaneous Metastasis Formation

The role of the tryptophan-catabolizing enzyme, indoleamine 2,3-dioxygenase (IDO1), in tumor escape and metastasis formation was analyzed using two pairs of Ido1+ and Ido1− murine breast cancer cell lines. Ido1 expression in 4T1 cells was knocked down by shRNA, and Ido1 expression in NT-5 cells was upregulated by stable transfection. Growth of Ido1− tumors and spontaneous metastasis formation were inhibited in immunocompetent mice. A higher level of cytotoxic T lymphocytes was generated by spleen cells from mice bearing Ido1− tumors than Ido1+ tumors. Tumor and metastatic growth was enhanced in immunodeficient mice, confirming an intensified immune response in the absence of Ido1 expression. However, Ido1+ tumors grow faster than Ido1− tumors in immunodeficient SCID/beige mice (lacking T, B, and NK cells) suggesting that some Ido1-controlled nonimmunological mechanisms may be involved in tumor cell growth regulation. In vitro experiments demonstrated that downregulation of Ido1 in tumor cells was associated with decreased cell proliferation, increased apoptosis, and changed expression of cell cycle regulatory genes, whereas upregulation of Ido1 in the cells had the opposite effects. Taken together, our findings indicate that Ido1 expression could exert immunological and nonimmunological effects in murine breast tumor cells.

IDO1 can be expressed in many tissue and cell types, such as placenta, lung, gut, and epididymis [13,[15][16][17]. The expression of IDO1 at the maternal-fetal interface in mice placenta is believed to play a role in the protection of the fetus from immunologic rejection by maternal immune mechanisms [18]. Indeed, this has been supported experimentally wherein the pharmacological inhibition of IDO1 by 1-methyltryptophan (1MT) resulted in the rejection of allogeneic, but not syngeneic, murine fetuses [19][20][21]. Ido1 activation occurs commonly in tumor cells and/or tumordraining lymph nodes (TDLNs), and pharmacological inhibition of IDO1 with 1-MT has been shown to result in T-celldependent antitumor responses in animal models [8,[22][23][24][25][26][27]. 1-MT was observed to retard tumor outgrowth but was unable to trigger complete tumor regression [6]. In vitro studies indicate that IDO1 is capable of exerting suppressive effects directly on T cells and can activate suppressive populations of regulatory T cells [8,9]. Furthermore, soluble cytotoxic-T-lymphocyte-antigen-4-(CTLA4-) expressing T regulatory cells induce IDO1 expression in DC, converting them into regulatory antigen-presenting cells (APCs) [24,26]. Intracellular signaling via CD80/86, CD200R, and FcεRI could induce IDO1 expression by DC [9,13,25]. A tumorrelated immune escape mechanism based on tryptophan degradation by IDO1 has been proposed [22,28]. Ido1 is expressed by tumor cells; however, the level of Ido1 expression is significantly lower than in placenta. Tumor cell inhibition of immune response was only demonstrated for Ido1-transfected clones, exhibiting 1000-fold increased expression of Ido1 mRNA relative to placental levels [22,29]. Thus, a role for IDO1 in tumor immune response is indicated but requires further investigation.
In this study, we examined the impact of IDO1 on tumor growth and metastasis in immune-competent and immunedeficient mice. Two murine breast cell lines, 4T1 and NT-5, expressing Ido1 and missing Ido1 expression, respectively, were utilized. NT-5 cells were transfected with an Ido1 expression vector to establish an NT-5/Ido1+ line. Expression of Ido1 in 4T1 cells was knocked down by transfection with an Ido1− specific shRNA expressing plasmid to establish a 4T1/Ido1− line. Using these two pairs of cell lines, we examined the relationship between Ido1 expression and cancer cell growth in vitro and in vivo. Our analysis of tumor growth and metastasis, in immunocompetent and immunodeficient mice, revealed that IDO1 not only modulated the immunological system, but also played an important biological role in tumor cell proliferation, cell cycle regulation, and antiapoptotic signaling.

Tumor Cell
Lines. The NT-5 HER-2/neu-expressing tumor cell line was provided by Elizabeth Jaffe, John Hopkins University. The 4T1 mouse mammary tumor cell line was purchased from American Type Culture Collection. Cells were cultured in RPMI-1640 medium (Cellgro Mediatech, Inc, Manassas, VA) supplemented with 10% FBS (Sigma-Aldrich Co, St. Luis, MO).

Plasmid Construction and Cell
Transfection. The mammalian expression vector for Ido1 was constructed by inserting full-size mouse Ido1 cDNA into the vector pRc/CMV (Invitrogen, Life Technologies Corp., Carlsbad, CA). NT-5 cells were cloned, and IDO expression in the individual clones was evaluated. The clone with the lowest IDO1 expression was used for transfection with either Ido1 constructs or control pRc/CMV vector using Lipofectamine 2000 according to manufacturer instructions (Invitrogen). Stable transfectants (NT-5/Ido1+ and NT-5/vector) were selected by growth in a medium supplemented with 400 μg/mL G418 (Sigma-Aldrich Co). IDO-positive 4T1 breast tumor cells were also cloned, and the clone with highest IDO expression was transfected with shRNA against IDO containing plasmid and the control plasmid (SABiosences). Stably transfected 4T1/Ido1− and 4T1/vector cells were selected with 600 μg/mL of G418 (Sigma-Aldrich).

Real-Time PCR and Gene Expression
Profiling. RNA was isolated from the cells using the RNeasy kit from Qiagen (Valencia, CA), and the extracted RNA was converted to cDNA using the RT 2 First Strand Kit from Super Array Bioscience Corporation (Frederick, MD) according to manufacturer's protocols. Real-time PCR was carried out using the Super Array RT 2 Real-Time SYBR Green PCR Master Mix (Super Array Bioscience) and was performed on the ABI Prism 7700 sequence detector real-time PCR system (AB Applied Biosystems, Foster City, CA). Primers for mouse Ido1 were forward 5 -GTACATCACCATGGCGTATG-3 ; reverse: 5 -CGAGGAAGAAGCCCTTGTC-3 . Standard curves were generated from five 10-fold serial dilutions of tumor cell cDNA, and no product could be observed in the negative control lacking template. Differences in gene expression were calculated by using the Ct method and normalized to GAPDH according to the manual from Super Array Bioscience (Super Array, Bioscience Corp., Frederick, MD).
The RT 2 Profiler PCR Array System and mouse cell cycle regulation RT 2 Profiler PCR Array (Super Array, Bioscience Corp) were used. Real-time PCR detection was carried out per the manufacturer's instructions. The experimental cocktail was prepared by mixing cDNA isolated from cell lines with the RT 2 Real-Time SyBR Green/ROX qPCR Master Mix (SABiosciences Corp., Frederick, MD). The mixtures were equally aliquoted into the 96-well plate containing predispensed gene-specific primer sets, then real-time PCR was performed using the ABI Prism 7900HT (ABI, Applied Biosytems Corp., Foster City, CA). Differences in genes expression were compared between 4T1/vector and 4T1/Ido1− cells or between NT-5/vector and NT-5/Ido1+ cells. Analyses of the raw data were done through the Super Array Data Analysis Web Portal (Super Array Bioscience Corp.).

Enzymatic
Assay for IDO Activity. IDO activity was determined using what is described in [10,30]. Briefly, cell extracts were prepared, mixed with an equal volume of 2x reaction buffer (100 mM PBS, 40 mM ascorbate, 20 μM methylene blue, 200 μg/mL catalase, 800 μM L-Tryptophan (Affymetrix, USB Inc, Cleveland, OH), pH 6.5), and incubated at 37 • C for 30 min in order to permit IDO-mediated conversion of L-Trp into N-formylkynurenine. The reaction was stopped with 30% trichloroacetic acid, and samples were incubated at 50 • C for 30 min to hydrolyze the Nformylkynurenine produced to kynurenine. After centrifugation, supernatants were collected and mixed with Ehrlich's reagent (0.8% p-dimethylaminobenzaldehyde in acetic acid) in 96-well plates, and absorbance at 490 nm was measured on a microplate reader. IDO activity is defined as the amount of enzyme required to produce 1 nmol of kynurenine per hr per mg of protein.

In Vitro Cell Proliferation
Assay. The cells were cultured in a 96-well plate for 2 days and pulsed with 0.2 μCi/well of 3 H-thymidine overnight followed by collection of supernatant. 50 uL of supernatant was mixed with 2 mL of Cyto-Spin (MP, Biomedicals LLC, Solon, OH) and counted by liquid scintillation counter (Wallac). The tumor cells were also counted by Cellomics Array Scan HCS Reader (Thermo Fisher, Pittsburgh, PA) following 3 days in culture as described previously [31].

Flow Cytometry Analysis.
Cells were cultured in a control medium or in serum starvation conditions (RPMI with 0.1% FBS or 0.3% FBS) for 3 days followed by culture in RPMI with 10.0% FBS for 24 hours. Cells were stained with FITCconjugated antibromodeoxyuridine (BrdU) and/or propidium iodide (PI) using FITC BrdU Flow Kit (BD Biosciences, Pharmingen, San Diego, CA) according to the manufacturer's instruction. Flow cytometry was performed by CYAN ADP (Beckman Counter, Brea, CA). The data were analyzed with the Cytomation Summit v4.3 software.

In Vivo Local and Metastatic Tumors
Assays. BALB/c, FVB/N, and SCID-beige were purchased from The Jackson Laboratory (Bar Harbor, ME) and used at 8-9 weeks of age. Mice were housed in the University of Pittsburgh Cancer Institute animal facility which is accredited for animal care by the American Association of Laboratory Animal Care. Experiments were performed in accordance with the approved institutional protocol and the guidelines of the Institutional Animal Care and Use Committee. BALB/c and SCID-beige mice were inoculated s.c. with 2.5 × 10 4 or 1.0 × 10 5 of 4T1/vector cells or 4T1/Ido1− cells per mouse; FBV/N and SCID-beige mice were inoculated s.c. with 5 × 10 6 of NT-5 or NT-5/Ido1+ cells per mouse. Tumor growth was evaluated by measurement of tumor diameters 3 times a week, and the tumor volume was calculated as length × width 2 × 0.52. All data is represented as mean ± SE. Experiments were terminated when tumors reached 2.0 cm in diameter. Each group contained 10 mice. Experiments were repeated twice. Lungs were harvested when tumors reached approximately 2 cm in diameter and fixed in Bounce solution. Metastases were counted under a dissecting microscope.

Cytotoxicity
Assay. Spleen cells from BALB/c mice bearing 4T1/Ido1− or 4T1 tumors were cultured with irradiated (15,000 rad) 4T1/Ido1− or 4T1 tumor cells for 5-8 days in the presence of 300 IU/mL of IL-2 in 24-well plates at the ratio as 3 × 10 6 spleen cells: 1 × 10 6 tumor cells. The cytotoxic activity of restimulated spleen cells was tested against 51 Cr-labeled 4T1/Ido1− or 4T1 cells at the effector: target ratio as 100 : 1. Spleen cells were distributed into V-96well plates preloaded with 51 Cr-labeled tumor cells; 96-well plates containing cell mixtures were centrifuged at 2000 rpm for 2 minutes and then incubated 4 hours at 37 • C. 25 μL of supernatant was transferred into yttrium silicate scintillatorcoated microplates (LumaPlate-96, PerkinElmer) and left overnight at RT. The level of β-emission released by 51 Cr was measured in a β-counter.

Western Blot Analysis.
The cells were lysed in RIPA buffer supplemented with 1% CLAP cocktail (antipan, leupeptin, pepstatin, and chymostatin) and 1 mM PMSF. 50 ug of protein extracts were resolved on 4.5-12% of SDS-PAGE and transferred to PVDF membranes (Bio-Rad, Hercules, CA). The membranes were blocked with 5% nonfat dry milk in TBST buffer, probed with antibodies against stratifin (1 : 200, Sigma), and Atm (1 : 200, Santa Cruz Biotechnology, Inc, Santa Cruz, CA) for two hours at RT or overnight at 4 • C, and then incubated with horseradish peroxidase-labeled secondary antibody (Santa Cruz). The signals were detected by ECL (Amersham). Films were scanned and analyzed by Image-QuanT data analysis software (Molecular Dynamics).

Statistical
Analysis. Data are presented as mean ± SD. Comparisons between values were performed using a twotailed Student's t-test. For the comparison of multiple groups, a one-or two-way ANOVA test was applied. For all statistical analyses, the level of significance was set at a pro bability of P < 0.05. All experiments were repeated 2-3 times.

Establishment of Two-Paired Stable Clones (Ido1− and Ido1+) of Murine Breast Cancer Cells. Initially, we measured
Ido1 expression in a panel of mouse tumor cell lines of differing histological origin. Two mouse breast tumor cell lines, 4T1 and NT-5, which showed differential expression of Ido1, were chosen. 4T1 cells, derived from sporadic breast tumor in BALB/c mice, are highly aggressive, metastatic, and poorly immunogenic [32,33]. NT-5 breast tumor cells are derived from HER-2/neu-transgenic FBV/N mice, and they are immunogenic [34].
We found that 4T1 cells express Ido1, whereas the level of Ido1 gene expression in NT-5 cells was very low (Ido1−). Clonal variation could be attributable whenever transfection is performed on a polyclonal whole cell population. To avoid this, our transfections were designed as follows: 4T1 breast tumor cells were cloned using a standard approach for selecting single-cell clones. Ido1 expression, in these individual cell clones, was evaluated. The clone the of the 4T1 tumor cells with the highest Ido1 expression was transfected with Ido1 siRNA and a control vector to generate stable 4T1/Ido1− and 4T1/vector clones.
Similarly, NT-5 cells (expressing Ido1 at a very low level) were also cloned, and the clone of NT-5 cells, with the lowest level of Ido1 expression, was selected for transfection with cDNA encoding Ido1 or control plasmid.
As shown in Figure 1(a), very low Ido1 expression was detected by RT-PCR in 4T1/Ido1− cells compared with 4T1 and 4T1/vector cells. The quantitative RT-PCR analysis revealed 78% downregulation of Ido1 in 4T1/Ido1− cells, and no changes of Ido1 expression in 4T1/vector cells in comparison with naïve 4T1 cells (Figure 1(a)).
We transfected an Ido1 vector and scramble vector into the Ido1 low-expressing NT-5 cells and established stable  of Ido1 expression in NT-5/Ido1+ cells in comparison to NT-5/vector cells (Figure 1(c)).

Effect of Ido1 Expression on Tumor Growth and Metastasis Formation.
To determine the role of IDO1 in tumor growth, BALB/c mice were inoculated s.c. with 4T1/vector or 4T1/Ido1− tumor cells. 4T1/Ido1− tumors grew significantly (P < 0.05) slower than 4T1/vector tumors (Figure 2(a)). Mice bearing 2 cm tumors were sacrificed according to our IACUC-approved protocol, and surviving mice bearing tumors of less than 2 cm were counted at sequential time points in order to evaluate survival. At the time of complete mortality in the control group, we observed nearly 50% survival of mice bearing Ido1− tumors (Figure 2(b)).
The elevated survival of mice bearing 4T1/Ido1− tumors may be a result of a less efficient immune response against tumor cells demonstrating inhibited Ido1 expression. To assess the immune response of BALB/c mice to 4T1/vector and 4T1/Ido1− tumors, spleens of mice bearing tumors were harvested, and spleen cell suspensions were prepared. Freshly prepared spleen cells had no cytotoxic activity against 4T1 tumor cells. Therefore, spleen cells were stimulated with irradiated 4T1 tumor cells for 5-8 days in the presence of IL-2 (300 IU/mL). We observed no difference in the abilities of irradiated 4T1/Ido1− and 4T1 (Ido1+) cells to stimulate spleen cells and generate cytotoxic T cells. The cytotoxic activity of stimulated spleen cells appeared after 5 days of stimulation and became more prominent after 7-8 days. Spleen cells obtained from irradiated 4T1/Ido1 tumor cells from 4T1/Ido1− tumor bearing mice demonstrated a higher cytotoxic activity than spleen cells from mice bearing 4T1/vector tumors (Figure 2(c)). Furthermore, targeted 4T1/Ido1− tumor cells were more sensitive to the cytotoxic activity of spleen cells than 4T1/vector cells (Figure 2(c)). Spleen cells cultured in IL-2 (300 IU/mL) without irradiated tumor cells showed no cytotoxic activity (data no shown). These results indicate that inhibition of 4T1/Ido1− tumor growth in BALB/c mice could be due to an elevated immune response against these cells as well as a greater sensitivity of 4T1/Ido1− cells to immune destruction.
To further test whether 4T1/Ido1− tumor growth inhibition is immunologically mediated, immunocompetent BALB/c mice or immunodeficient SCID-beige mice (which lack T, B, and NK cells) were inoculated s.c. with 4T1/vector and 4T1/Ido1− cells. Our expectation was that if differences in Ido1+ and Ido1− tumor growth are immune modulated, then both tumors should grow similarly in SCID/beige mice. Tumor growth from Ido1− cells was inhibited in comparison with Ido1+ cells in BALB/c mice. As shown in Figure 2(d), both Ido1− and Ido1+ tumors grew faster in immunodeficient SCID/beige mice than in immunocompetent BALB/c mice. However, Ido1+ tumors grew faster than Ido1− tumors in SCID/beige mice suggesting that some Ido1-controlled nonimmunological mechanisms may be involved in tumor cell growth regulation.
To test whether spontaneous pulmonary metastasis formation is dependent on Ido1 gene expression, the number of metastases in BALB/c and SCID/beige mice was counted when tumors reached 2 cm in diameter. We found that 4T1/ Ido1− cells generated significantly less metastasis in the lungs of BALB/c mice than 4T1 cells (92 versus 396) ( Table 1). The pulmonary metastases observed in immunodeficient SCIDbeige mice bearing Ido1− tumors were higher than in the immunocompetent BALB/c mice bearing Ido1− tumors (Table 1).
To further elucidate the role of IDO1, in tumor growth and metastasis formation, we used NT-5/Ido1− cells and NT-5/Ido1+ tumor cells and FVB/N mice that are immunocompetent for NT-5 cells. It is known that NT-5 breast tumor cells, derived from HER-2/neu-transgenic FBV/N mice, are immunogenic [34]. Indeed, tumor growth was prevented when 1 × 10 6 NT-5 cells were inoculated into mice. However, inoculation of 5 × 10 6 NT-5 cells resulted in tumor development. Tumor growth in many FVB/N mice was associated with enormous necrosis which made it difficult to measure tumor size; therefore, mice were sacrificed when they showed signs of moribund, and lung metastases were counted. We found that NT-5/Ido1+ cells are more metastatic than NT-5/Ido−cells in immunocompetent FVB/N mice ( Table 1). The numbers of metastases developed from NT-5/Ido1+ cells and NT-5/Ido− cells in immunodeficient SCID/beige mice were significantly higher than in FVB/mice (Table 1). However, even in immunodeficient mice, NT-5/Ido1− cells developed less metastasis than NT-5/Ido1+ cells.
Taken together, these results indicate that Ido1 plays role in tumor growth and metastasis formation. sus S versus G2/M) and makes it possible to detect apoptotic cells with fractional DNA content. Figure 3 represents the results of the experiments in which double staining was applied. Bivariate distribution (contour map) of DNA content and BrdU incorporation in 4T1 and 4T1/Ido1− cells are shown. The 4T1/Ido1− cells growing in control media have significantly (P < 0.05) higher percentage of apoptotic cells than 4T1/Ido1+ cells. Starvation is a condition that often occurs within the tumors. We analyzed the effect of starvation conditions (low serum in culture medium) on apoptosis induction in 4T1/Ido1+ and 4T1/Ido1− cells. Here, the proportion of apoptotic 4T1/ Ido1− cells reached 86%, which is approximately 3 times higher than that observed in the 4T1/Ido1+ cell population.

Effect of Ido1 Expression on Apoptosis Changes Cell Cycle and Tumor Cell Proliferation In Vitro. The differences in
The 4T1/Ido1+ cells recover faster after starvation (Figure 3). Cell cycle phase distribution was also different in 4T1/Ido1+ and 4T1/Ido1− cell populations in control media and under starvation (Figure 3). Control cells showed a significantly higher proportion of cells in the S phase than 4T1/Ido1− cells. 4T1/Ido1+ growing in low serum demonstrated a higher proportion of cells in M and G0 phases than 4T1/Ido1− cells. We then compared the proliferation of 4T1/Ido1+ and 4T1/Ido1− cells as well as NT-5/Ido1− and NT-5/Ido1+ cells in vitro. The results of 3 H-thymidine incorporation showed a significant reduction in proliferation of Ido1− cells in comparison to Ido1+ cells (Figure 4(a)). Next, the same pairs of Ido1+/Ido1− cells were grown for 3 days and counted as described previously [31]. As shown in Figure 4(b),   Table 2). The expression of four genes (Trp63(p63), dystonin (Dst), stromal antigen1 (Stag1) and microtubule-actin crosslinking factor (Macf1)) was lower in Ido1 downregulated cells, and higher with Ido1 upregulation (Table 2). To verify whether the changes in gene expression are associated with upregulation of IDO1 protein expression in Ido1+ and Ido1− cells, we performed a western blot analysis of the cell cycle regulators ATM and stratifin (Sfn) [35,36].
Atm (ataxia telangiectasia mutated) is an important cell cycle regulation gene encoding a serine/threonine protein kinase that is recruited and activated by DNA doublestrand breaks. ATM phosphorylates several key proteins that initiate activation of the DNA damage checkpoint, leading to cell cycle arrest, DNA repair, or apoptosis [35,36]. Atm expression was significantly decreased (−18.3-fold) in 4T1/Ido1− cells and only slightly changed (1.7-fold) in NT-5/Ido1+ ( Table 2). As demonstrated in Figure 4(c), ATM is expressed in 4T1/vector (Ido1+) cells, and ATM production is dramatically reduced in Ido1− cells.
Stratifin (Sfn), also named 14-3-3 sigma, is a multifunctional protein involved in cell cycle regulation [36]. Downregulation of Ido1 in 4T1 cells is associated with upregulation of Sfn gene expression (12 fold); and upregulation of Ido1 in NT5 cells is associated with a decreased level (−7 fold) of Sfn gene expression (Table 2).
Western blot analysis demonstrated that SFN protein is increased in Ido1− cells, and it is reduced in Ido1+ cells (Figure 4(d)). Thus, the results of the PCR analysis of Atm and Sfn gene expression are consistent with the results of the western blot analysis of the ATM and SFN proteins

Discussion
We first analyzed the impact of Ido1 expression on tumor growth and spontaneous pulmonary metastasis formation and then investigated the IDO1-regulated mechanisms involved in tumorigenesis. One of the key experimental strategies to elucidate the function of a gene in vitro and in vivo is the specific inhibition or upregulation of its expression. Here, we explored both approaches. Ido1+ cells were transected with plasmid expressing shRNA for Ido1, and Ido1− cells were transfected with Ido1 cDNA. These two pairs of Ido+ and Ido− clones were evaluated throughout our investigation.
We found that 4T1/Ido1+ tumors grew faster than 4T1/Ido1− tumors in immunocompetent mice. Growth  inhibition of Ido1− tumor cells can be attributed to a more efficient immune response. This conclusion is based on the following. (1) Spleen cells from mice bearing 4T1/Ido1− tumors generated a higher level of cytotoxic T cells (CTCs). These CTCs were able to destroy 4T1/Ido1− cells and 4T1/ Ido1+ cells. However, their cytotoxic effects were higher against 4T1/Ido1− cells, most likely due to a higher sensitivity of these cells to apoptotic signaling. (2) In the absence of major immune mechanisms in SCID/beige mice, the growth of Ido1− tumors was accelerated, suggesting a role for the immune system in Ido1-dependent control of tumor growth. These findings are consistent with previous publications, demonstrating the role of IDO1 in tumor immunosuppression [5-9, 12-14, 37]. Activation of Ido1 by IFN-γ links inflammation and cancer immune surveillance [11,28,35], and an increase in IFN-γ within the tumor microenvironment also induced IDO1 expression in APCs and DCs that localize to the tumordraining lymph nodes [38]. We found that IFN-γ treatment induced Ido1 expression in Ido1+ cells and had no effect on Ido1− cells. Recently, a mouse genetic study revealed a critical role for IDO1 in supporting inflammatory skin carcinogenesis [26]. An analysis of IDO1 dysregulation in cancer showed that JAK/STAT and NF-κB signaling are also essential for IDO1 induction in oncogenically transformed skin epithelial cells [39]. The anti-inflammatory agent, ethyl pyruvate, was found to be able to block Ido1 expression, indicating a link between IDO1 and inflammatory processes [39].
For the first time, we showed that Ido1 regulates tumor metastasis via immunosuppressive mechanisms. Ido1+ tumors develop dramatically higher levels of spontaneous pulmonary metastasis in immunocompetent mice than Ido1− tumors. Levels of lung metastasis in SCID/beige mice (deficient in T cells, B cells, and NK cells) are higher than in immunocompetent mice, thus confirming an immunosuppressive role for IDO1 in spontaneous metastasis formation.
In contrast to our expectation, growth of the Ido1− tumors was slower than growth of Ido1+ tumors in immun-odeficient SCID/beige mice, suggesting that some Ido1associated nonimmunological mechanisms may be involved in tumor cell growth regulation. It is conceivable that Ido1− cells may have an increased sensitivity to apoptosis signaling or slower cell cycle progression and proliferation than Ido1+ cells. Our in vitro study revealed multiple differences between Ido1− and Ido1+ cell populations suggesting the complexity of IDO1-regulated tumor cell functions. Ido1− cells grow slower than Ido1+ cells. Similarly, the inhibition of IDO expression in rat NMU-induced breast cancer showed the reduced cell proliferation in vitro [40].
Ido1+ cells have small proportions of apoptotic cells in normal media and in starved conditions, and they recover better after starvation. In contrast, 4T1/Ido1− cells, in which the Ido1 gene was down-regulated via shRNA transfection, have dramatically increased sensitivity to apoptosis, thus suggesting an important antiapoptotic function of Ido1 in tumor cells. However, molecular pathways involved in Ido1associated apoptosis signaling remain virtually unexplored, offering new areas for investigation.
Cell cycle distributions within Ido1− and Ido1+ tumor cell populations also differ in that Ido1+ cells show a higher proportion of cells in S and M phases than Ido1− cells. Using a comparative RT-PCR analysis of cell cycle regulatory gene expression, we found that downregulation of Ido1 is associated with decreased expression of multiple genes involved in regulation of G1/S transition, S phase, G2/M transition, checkpoints, and M phase, whereas upregulation of Ido1 induces expression of these genes. Protein analysis of the products of two genes, stratifin (SFN) and ATM, in Ido1+ and Ido1− cells show that the up-and downregulation of these genes correlates with the increased/decreased production of their respective proteins. Thus, the roles of Ido1 in breast tumor growth are complex and suggest the involvement of additional nonimmunological mechanisms.

Conclusion
Our primary finding is that Ido1 expression has immunological and nonimmunological effects on breast tumor growth and spontaneous pulmonary metastasis formation. We demonstrated that IDO1 is involved in the regulation of tumor cell proliferation in vitro and in vivo. Ido1 expression also affects apoptosis in tumor cells. The involvement of IDO1 in cell cycle regulation and tumor cell stress response, in addition to the known role of IDO1 in tumor immunosuppression, make it an attractive target for antitumor therapy.