Immunodeficient hosts exhibit high acceptance of xenogeneic or neoplastic cells mainly due to lack of adaptive immunity, although it still remains to be elucidated how innate response affects the engraftment. IL-2R common γ-chain (IL-2Rγc) signaling is required for development of NK cells and a subset of dendritic cells producing IFN-γ. To better understand innate response in the absence of adaptive immunity, we examined amounts of metastatic foci in the livers after intrasplenic transfer of human colon cancer HCT116 cells into NOD/SCID versus NOD/SCID/IL-2Rγcnull (NOG) hosts. The intravital microscopic imaging of livers in the hosts depleted of NK cells and/or macrophages revealed that IL-2Rγc function critically contributes to elimination of cancer cells without the need for NK cells and macrophages. In the absence of IL-2Rγc, macrophages play a role in the defense against tumors despite the NOD Sirpa allele, which allows human CD47 to bind to the encoded signal regulatory protein α to inhibit macrophage phagocytosis of human cells. Analogous experiments using human pancreas cancer MIA PaCa-2 cells provided findings roughly similar to those from the experiments using HCT116 cells except for lack of suppression of metastases by macrophages in NOG hosts. Administration of mouse IFN-γ to NOG hosts appeared to partially compensate lack of IL-2Rγc–dependent elimination of transferred HCT116 cells. These results provide insights into the nature of innate response in the absence of adaptive immunity, aiding in developing tumor xenograft models in experimental oncology.

A number of immunodeficient mouse strains have been widely used as hosts in experimental transplantation models (1). High acceptance of xenogeneic or neoplastic cells is exhibited mainly due to lack of adaptive immunity resulting from failure in development of thymus-dependent T cells in strains bearing nu locus or in development of both T cells and B cells in strains bearing scid locus or disrupted Rag gene. Additionally, genetic introduction of impaired innate immunity onto strains lacking adaptive immunity also has been applied to the development of hosts more highly accepting xenogeneic or neoplastic cells, represented by the introduction of null gene of IL-2R common γ-chain (IL-2Rγc) or NOD background onto the strains having scid locus or disrupted Rag gene (2). IL-2Rγc signaling is required for development of NK cells and a subset of dendritic cells (DCs) producing IFN-γ represented by the CD11c+B220+CD122+ phenotype (3, 4). The allele of Sirpa gene in the NOD background mice allows human CD47 to bind to the encoded signal regulatory protein α (sirpa) to inhibit macrophage phagocytosis of human cells (5, 6). The details of how such alteration in innate immunity contributes to the engraftment are of considerable interest but still remain to be elucidated.

To better understand innate response in the absence of adaptive immunity, we examined amounts of metastatic foci in the livers after intrasplenic transfer of human colon cancer HCT116 cells to NOD/SCID versus NOD/SCID/IL-2Rγcnull (NOG) hosts using intravital microscopic imaging combined with in vivo depletion of NK cells and macrophages. The results demonstrated a line of evidence for contribution of innate effectors other than NK cells and macrophages in the presence of IL-2Rγc versus of macrophages in the absence of IL-2Rγc to elimination of human cancer cells in the NOD background hosts lacking adaptive immunity.

A human colon cancer cell line HCT116 and a human pancreas cancer cell line MIA PaCa-2 were obtained from the American Type Culture Collection (Manassas, VA). HCT116 cells were maintained in McCoy’s 5A (Sigma-Aldrich, St. Louis, MO) containing antibiotics and 10% FBS (HyClone, Logan, UT). MIA PaCa-2 cells were maintained in DMEM (Sigma-Aldrich) containing antibiotics, 10% FBS, and 2.5% horse serum. HCT116 and MIA PaCa-2 cells were incubated in a humidified (37°C, 5% CO2) incubator and passaged on reaching 80% confluence.

Venus has been developed by mutagenesis engineering as an improved version of yellow fluorescent protein, which is derived from GFP variants (7). Venus is fully and stably fluorescent due to its resistance to pH and chloride ion. The original vector containing the Venus gene sequence was gifted from Dr. Atsushi Miyawaki (Brain Science Institute, RIKEN, Saitama, Japan). The Venus gene fragments were amplified by PCR and then purified through electrophoresis. These PCR products were inserted in the pcDNA3.1/myc-His(−)A vector (Invitrogen, Carlsbad, CA) at XbaI and KpnI sites with a DNA ligation kit (Takara Bio, Shiga, Japan), followed by cloning using XL2-Blue MRF′ ultracompetent cells (Strategene, La Jolla, CA). It was confirmed by sequencing that the cloned pcDNA3.1/Venus vector plasmids keep the Venus gene sequence without any mutational alteration.

The Venus gene transduction of HCT116 cells was carried out using Lipofectamine 2000 (Invitrogen) as recommended by the manufacturer. The cells were treated with G418 (Invitrogen) at final concentration of 1.0 mg/ml for selection of positive clones 2 d after transduction and then cultured for a week. The Venus expression on selected clones was evaluated by flow cytometric analysis using Epics XL-MCL (Beckman Coulter, Brea, CA).

The in vivo experiments were performed in accordance with institutional guidelines and approved by the Animal Experimentation Committee of the Keio University and the Central Institute for Experimental Animals. We bred NOD/Shi-scid (NOD/SCID) and NOG mice and used them at the age of 9–11 wk. Liver metastases were induced by intrasplenic transfer of HCT116 cells, followed by splenectomy (8). In the case of transfer of 1 × 106 cells, the mice were subjected to the intravital microscopic bioimaging within 30 min or 1–2 wk later as described below. After the intravital microscopic studies, the mice were sacrificed and liver metastases were enumerated immediately.

The intravital observation of livers after intrasplenic transfer of fluorescently visualized HCT116 cells expressing Venus was performed with an in vivo video microscopic method as reported previously (9, 10). For each experimental group, four to five mice were examined. The mouse was placed on an inverted microscope platform, the disk scan unit system (IX81-DSU; Olympus, Tokyo, Japan), after abdominal section. For each mouse, microscopic observations on six to nine hepatic lobules were recorded using the electron multiplying CCD camera (Cascade II; Nippon Roper, Tokyo, Japan) and the MetaMorph software (MDS, Toronto, ON, Canada).

Based on the obtained bioimaging data and histological findings, terminal portal venules and centrilobular venules were identified in examined hepatic lobules. The measurement of distances between a terminal portal venule and the closest centrilobular venule was performed using the Scion image software (Scion, Frederick, MD), indicating that the average distance was 320.4 ± 66.7 μm. Considering this value, each C point was used as the center to draw a circle at the radius of 320 μm, which reflected a putative hepatic lobule. Each circle was divided into three zones, that is, the periportal zone, the middle zone, and the pericentral zone. The area of each metastatic focus of fluorescently visualized HCT116 cells expressing Venus was measured and then the percentages of tumor lesion in the periportal, middle, and pericentral zones were calculated.

Sections (3-μm-thick) were cut from paraffin-embedded tissue blocks. Immunostaining of the section with a mouse anti-HLA class I (A, B, and C) mAb (clone EMR8-5; Hokudo, Sapporo, Japan) was performed on the Bond-Max automated IHC platform (Leica Biosystems, Mount Waverley, Australia). The photographic images of the immunostained sections were prepared by Axio Imager.M1 (Carl Zeiss, Thornwood, NY) and then converted to the grayscale images using Photoshop CC 2014 (Adobe Systems, San Jose, CA). The measurement of metastatic foci of MIA PaCa-2 cells visualized immunohistochemically with HLA class I expression was carried out using the software ImageJ version 1.48 (http://imagej.nih.gov/ij/).

To deplete only NK cells, NOD/SCID mice were i.p. given 400 μl PBS containing 20 μl anti-asialo GM1 antiserum (Wako, Osaka, Japan) on days −3, −2, and −1, followed by intrasplenic transfer of human cancer cells on day 0 (3). To deplete both NK cells and macrophages, NOD/SCID mice were i.p. given 400 μl PBS containing 20 μl anti-asialo GM1 antiserum on days −3, −2, and −1, followed by treatment for macrophage depletion on day −2 as describe below and then by intrasplenic transfer of human cancer cells on day 0. To deplete macrophages, dichloromethylene diphosphonate–containing liposomes (Cl2MDP-liposomes) were prepared as described previously (11). NOD/SCID and NOG mice were i.v. given Cl2MDP-liposomes suspended in 100 μl PBS on day −2, followed by intrasplenic transfer of human cancer cells on day 0.

NOG mice were i.p. given mouse rIFN-γ at a dose of 1 × 105 units 1 d before transfer of HCT116 cells as reported previously (12).

Spleens from NOD/SCID mouse were minced and digested with 0.1% collagenase (Roche Diagnostics, Laval, QC, Canada) and DNase (1 mg/ml; Wako Pure Chemical Industries, Osaka, Japan) at 37°C for 30 min. After washing with 2% FCS in PBS, cells were stained with biotinylated mouse B220 Ab (BioLegend, San Diego, CA) and incubated with anti-biotin magnetic beads (Miltenyi Biotec, Sunnyvale, CA) to isolate the DC subpopulation. B220+ and B220 fractions were separated on a MACS column (Miltenyi Biotec) and the enriched B220+ fractions were stained with PE-labeled CD11c Ab (BioLegend). The B220+CD11c+ cells were sorted using the BD FACSAria II cell sorter (BD Biosciences, San Jose, CA). The purity level of B220+CD11c+ was 96.1%. RBCs of the B220 cell fraction were lysed in Pharm Lyse buffer (BD Biosciences) and washed with 2% FCS in PBS. The purified B220+CD11c+ cells and B220 cells were resuspended in RPMI 1640 (Corning, Tewksbury, MA), and 1 × 105 cells were transferred i.v. into NOG mice 1 d before transfer of HCT116 cells.

Sections (3 μm thick) were cut from paraffin-embedded tissue blocks. To reveal distribution of hepatic macrophages, immunostaining of liver tissues was carried out with a rat anti-mouse F4/80 mAb (Bio-Rad/AbD Serotec, Kidlington, U.K.) and a rabbit anti-mouse Ym-1 polyclonal Ab (Stemcell Technologies, Carlsbad, CA) and performed with Bond-Max automated immunostainer (Leica Biosystems). Sections were counterstained with hematoxylin.

The statistical significance of data among different experimental groups was determined by Student t test or one-way ANOVA, and a p value < 0.05 was considered significant.

As with parental HCT116 cells, HCT116 cells expressing Venus (termed HCT116/Venus cells) robustly metastasized in NOG but not in NOD/SCID hosts. Fig. 1 shows the representative results. One week after the transfer of 1 × 106 HCT116/Venus cells, amounts of metastatic foci were microscopically evident preferentially in periportal regions of the livers in NOG but not in NOD/SCID hosts; all of five NOG mice versus none of five NOD/SCID mice exhibited microscopic metastases. At that time, macroscopic metastases were not observed in any of the five NOG or NOD/SCID mice. By 2 wk after the transfer, the hepatic metastases were developed macroscopically in NOG but not in NOD/SCID hosts; all of five NOG mice versus none of five NOD/SCID mice exhibited macroscopic metastases. At that time, microscopic metastases were observed in just one of the five NOD/SCID mice, looking tiny and morphologically fragile. Even 3 wk after the transfer, only two of five NOD/SCID mice exhibited macroscopic metastases, which were observed as a small number of tiny nodules in the livers.

FIGURE 1.

Robust formation of metastatic foci of human colon cancer HCT116 cells occurs in the liver of NOG hosts. Macroscopic and microscopic (H&E staining) findings of the livers in NOD/SCID versus NOG hosts at indicated time points after transfer of 1 × 106 HCT116/Venus cells are shown. C, centrilobular venules; P, terminal portal venules; T, tumor.

FIGURE 1.

Robust formation of metastatic foci of human colon cancer HCT116 cells occurs in the liver of NOG hosts. Macroscopic and microscopic (H&E staining) findings of the livers in NOD/SCID versus NOG hosts at indicated time points after transfer of 1 × 106 HCT116/Venus cells are shown. C, centrilobular venules; P, terminal portal venules; T, tumor.

Close modal

To assess differences in events occurring immediately after intrasplenic transfer of human cancer cells between the two strains, the intravital microscopic imaging was applied. Transfer of 1 × 106 HCT116/Venus cells to each host was suitable for the bioimaging analysis. Within 30 min after the transfer, arrest of HCT116/Venus cells inside sinusoids of periportal regions of hepatic lobules was observed similarly in both strains (Fig. 2A). The distribution of arrest of HCT116/Venus seemed to overlap with preferential location of hepatic macrophages. It was of considerable interest whether macrophages participated in trapping of human cancer cells transferred intrasplenically. The in vivo macrophage-depletion experiments using Cl2MDP-liposomes in NOD/SCID and NOG hosts revealed that the cancer cell arrest inside hepatic sinusoids occurred independently of macrophages (Fig. 2B, 2C).

FIGURE 2.

Arrest of HCT116 cells occurs inside sinusoids of periportal regions of hepatic lobules immediately after intrasplenic transfer to NOD/SCID and NOG hosts. (A) Amounts and distributions of metastatic foci were intravitally evaluated within 30 min after the intrasplenic transfer to NOD/SCID and NOG mice by detecting the fluorescence of HCT116/Venus cells. (B) Representative results of immunohistochemical staining with F4/80 for detection of hepatic macrophages in NOG mice undergoing indicated treatments are shown. (C) Amounts and distributions of metastatic foci were intravitally evaluated within 30 min after the intrasplenic transfer to hosts undergoing indicted treatments. Representative results from NOG hosts are shown. Columns show mean; bars indicate SD. Outer white broken circles demarcate the periportal (P) zone, middle white broken circles demarcate the middle (M) zone, and inner white broken circles demarcate the pericentral (C) zone. (A) *p < 0.05, statistically significant difference when compared with metastatic foci located at different zones of NOD/SCID hosts; p < 0.05, statistically significant difference when compared with metastatic foci located at different zones of NOG hosts. (C) *p < 0.05, statistically significant difference when compared with metastatic foci located at different zones in NOG hosts treated with saline; p < 0.05, statistically significant difference when compared with metastatic foci located at different zones in NOG hosts treated with empty-liposomes; §p < 0.05, statistically significant difference when compared with metastatic foci located at different zones in NOG hosts treated with Cl2MDP-liposomes. C, centrilobular venules; Cl-L, treatment with Cl2MDP-liposomes; Control, treatment with saline; Empty-L, treatment with empty-liposomes; P, terminal portal venules.

FIGURE 2.

Arrest of HCT116 cells occurs inside sinusoids of periportal regions of hepatic lobules immediately after intrasplenic transfer to NOD/SCID and NOG hosts. (A) Amounts and distributions of metastatic foci were intravitally evaluated within 30 min after the intrasplenic transfer to NOD/SCID and NOG mice by detecting the fluorescence of HCT116/Venus cells. (B) Representative results of immunohistochemical staining with F4/80 for detection of hepatic macrophages in NOG mice undergoing indicated treatments are shown. (C) Amounts and distributions of metastatic foci were intravitally evaluated within 30 min after the intrasplenic transfer to hosts undergoing indicted treatments. Representative results from NOG hosts are shown. Columns show mean; bars indicate SD. Outer white broken circles demarcate the periportal (P) zone, middle white broken circles demarcate the middle (M) zone, and inner white broken circles demarcate the pericentral (C) zone. (A) *p < 0.05, statistically significant difference when compared with metastatic foci located at different zones of NOD/SCID hosts; p < 0.05, statistically significant difference when compared with metastatic foci located at different zones of NOG hosts. (C) *p < 0.05, statistically significant difference when compared with metastatic foci located at different zones in NOG hosts treated with saline; p < 0.05, statistically significant difference when compared with metastatic foci located at different zones in NOG hosts treated with empty-liposomes; §p < 0.05, statistically significant difference when compared with metastatic foci located at different zones in NOG hosts treated with Cl2MDP-liposomes. C, centrilobular venules; Cl-L, treatment with Cl2MDP-liposomes; Control, treatment with saline; Empty-L, treatment with empty-liposomes; P, terminal portal venules.

Close modal

One week after the transfer, amounts of metastatic foci were microscopically evident preferentially in periportal regions of the lobules in NOG mice but were not detected in any hepatic area of NOD/SCID mice (Figs. 1, 3A). The distribution of micrometastases was similar to that of the arrest of HCT116/Venus cells within 30 min after the transfer, suggesting that the micrometastatic foci originated from cancer cells arrested initially after the intrasplenic transfer.

FIGURE 3.

Hepatic metastatic foci appear in NOG but not in NOD/SCID hosts, distributing similarly to the initial cancer cell arrest. (A) Amounts and distributions of metastatic foci were intravitally evaluated 1 wk after the intrasplenic transfer to NOD/SCID and NOG mice by detecting the fluorescence of HCT116/Venus cells. (B) Amounts and distributions of metastatic foci were intravitally evaluated 2 wk after the intrasplenic transfer to NOG mice undergoing indicated treatments. Columns show mean; bars indicate SD. Outer white broken circles demarcate the periportal (P) zone, middle white broken circles demarcate the middle (M) zone, and inner white broken circles demarcate the pericentral (C) zone. (A) *p < 0.05, statistically significant difference between NOD/SCID versus NOG hosts. (B) *p < 0.05, statistically significant difference between NOG hosts treated with Cl2MDP-liposomes versus saline; p < 0.05, statistically significant difference between NOG hosts treated with Cl2MDP-liposomes versus empty-liposomes. C, centrilobular venules; Cl-L, treatment with Cl2MDP-liposomes; Control, treatment with saline; Empty-L, treatment with empty-liposomes; P, terminal portal venules.

FIGURE 3.

Hepatic metastatic foci appear in NOG but not in NOD/SCID hosts, distributing similarly to the initial cancer cell arrest. (A) Amounts and distributions of metastatic foci were intravitally evaluated 1 wk after the intrasplenic transfer to NOD/SCID and NOG mice by detecting the fluorescence of HCT116/Venus cells. (B) Amounts and distributions of metastatic foci were intravitally evaluated 2 wk after the intrasplenic transfer to NOG mice undergoing indicated treatments. Columns show mean; bars indicate SD. Outer white broken circles demarcate the periportal (P) zone, middle white broken circles demarcate the middle (M) zone, and inner white broken circles demarcate the pericentral (C) zone. (A) *p < 0.05, statistically significant difference between NOD/SCID versus NOG hosts. (B) *p < 0.05, statistically significant difference between NOG hosts treated with Cl2MDP-liposomes versus saline; p < 0.05, statistically significant difference between NOG hosts treated with Cl2MDP-liposomes versus empty-liposomes. C, centrilobular venules; Cl-L, treatment with Cl2MDP-liposomes; Control, treatment with saline; Empty-L, treatment with empty-liposomes; P, terminal portal venules.

Close modal

Macrophages have potentials not only to kill tumor cells but also to promote tumor progression via several mechanisms, such as proangiogenic and immunosuppressive actions, under some conditions (13). To assess how host macrophages influence the tumor formation, we compared metastatic foci of HCT116/Venus cells in NOG hosts in in vivo–depleted and nondepleted of macrophages. As shown in Fig. 3B, macrophages were found to suppress hepatic metastases in NOG mice.

Although it is thought that lack of NK cells is one of the most critical features with respect to the immunodeficiency of NOG mice, loss not only of NK cell–dependent but also of NK cell–independent mechanisms contribute to superior reconstitution of human hematopoietic tissues (3, 4). To assess the dependences of antitumor effects on NK cells in NOD/SCID hosts, in vivo depletion experiments were carried out using anti-asialo GM1 Ab treatment, analogous to the previous study (3). NK cell depletion failed to initiate hepatic metastases of HCT116 cells in any NOD/SCID hosts, revealing that the antitumor effects occurring with requiring IL-2Rγc were independent of NK cells (Fig. 4). Considering the suppressive effects of macrophages on hepatic metastases that were observed in NOG hosts as described above, metastatic foci in the livers were compared between NOD/SCID hosts depleted versus not depleted of macrophages. With or without in vivo NK cell depletion, macrophage depletion also failed to initiate hepatic metastases of HCT116 cells in contrast to the case of NOG hosts (Fig. 4). These results provided evidence that IL-2Rγc–dependent tumor elimination in NOD/SCID hosts does not require NK cells or macrophages.

FIGURE 4.

Antitumor effects occurring in NOD/SCID but not in NOG hosts do not require NK cells or macrophages. Amounts and distributions of metastatic foci were intravitally evaluated 1 wk after the intrasplenic transfer to NOD/SCID and NOG mice with or without in vivo depletion of NK cells and/or macrophages by detecting the fluorescence of HCT116/Venus cells. Columns show mean; bars indicate SD. asGM1, treatment with anti-asialo GM1 antiserum to deplete NK cells; Cl-L, treatment with Cl2MDP-liposomes to deplete macrophages. *p < 0.05, statistically significant difference when compared with each group of NOD/SCID hosts; p < 0.05, statistically significant difference between NOG hosts with versus without depletion of macrophages. C zone, pericentral zone; M zone, middle zone; P zone, periportal zone.

FIGURE 4.

Antitumor effects occurring in NOD/SCID but not in NOG hosts do not require NK cells or macrophages. Amounts and distributions of metastatic foci were intravitally evaluated 1 wk after the intrasplenic transfer to NOD/SCID and NOG mice with or without in vivo depletion of NK cells and/or macrophages by detecting the fluorescence of HCT116/Venus cells. Columns show mean; bars indicate SD. asGM1, treatment with anti-asialo GM1 antiserum to deplete NK cells; Cl-L, treatment with Cl2MDP-liposomes to deplete macrophages. *p < 0.05, statistically significant difference when compared with each group of NOD/SCID hosts; p < 0.05, statistically significant difference between NOG hosts with versus without depletion of macrophages. C zone, pericentral zone; M zone, middle zone; P zone, periportal zone.

Close modal

To examine whether antitumor effects observed are limited to HCT116 cells, analogous experiments using human pancreas cancer MIA PaCa-2 cells were carried out (Fig. 5). Transferred MIA PaCa-2 cells formed a negligible amount of metastatic foci in the livers of NOD/SCID hosts whereas robust metastases occurred in the livers of NOG hosts. Consistent with the observations in the setting of HCT116 cells, depletion of NK cells and/or macrophages did not significantly affect MIA PaCa-2 metastases in NOD/SCID hosts. Alternatively, unlike the HCT116 metastases, the MIA PaCa-2 metastases were not affected by depletion of macrophages in NOG hosts.

FIGURE 5.

Analogous experiments using MIA PaCa-2 cells provide findings roughly similar to those from experiments using HCT116 cells except for lack of suppression of metastases by macrophages in NOG hosts. (A) Amounts of metastatic foci were immunohistochemically evaluated 2 wk after the intrasplenic transfer to NOD/SCID and NOG hosts undergoing indicated treatment by detecting HLA class I expression of MIA PaCa-2 cells. (B) Representative results of photographic imaging of immunostained sections are shown. Columns show mean; bars indicate SD. asGM1, treatment with anti-asialo GM1 antiserum to deplete NK cells; Cl-L, treatment with Cl2MDP-liposomes to deplete macrophages; Control, treatment with saline. *p < 0.05, statistically significant difference when compared with each group of NOD/SCID hosts.

FIGURE 5.

Analogous experiments using MIA PaCa-2 cells provide findings roughly similar to those from experiments using HCT116 cells except for lack of suppression of metastases by macrophages in NOG hosts. (A) Amounts of metastatic foci were immunohistochemically evaluated 2 wk after the intrasplenic transfer to NOD/SCID and NOG hosts undergoing indicated treatment by detecting HLA class I expression of MIA PaCa-2 cells. (B) Representative results of photographic imaging of immunostained sections are shown. Columns show mean; bars indicate SD. asGM1, treatment with anti-asialo GM1 antiserum to deplete NK cells; Cl-L, treatment with Cl2MDP-liposomes to deplete macrophages; Control, treatment with saline. *p < 0.05, statistically significant difference when compared with each group of NOD/SCID hosts.

Close modal

To assess the potential mechanisms by which macrophages play a role in the defense against metastases in NOG but not in NOD/SCID hosts, M1/M2 macrophage polarization in the metastatic foci was analyzed by immunohistochemistry (Fig. 6). The staining of liver sections with representative markers such as inducible NO synthase, CD40, CD86, and MHC class II was inconvenient to use for specific detection of M1 macrophages in the livers because they were expressed not only by macrophages but also by nonhematopoietic cells as reported previously (1419). The staining with a representative marker Ym1 was useful to detect M2 macrophages in the liver (15). In the present study, F4/80+ cells without Ym1 positivity were regarded as M1 macrophages whereas cells exhibiting concordance of Ym1 positivity with F4/80 positivity were regarded as M2 macrophages.

FIGURE 6.

Distribution of F4/80+ and Ym1+ cells in MIA PaCa-2 metastatic foci differs from that in HCT116 metastatic foci in NOG hosts. (A) Representative results of histological examination of the livers in intact NOD/SCID versus NOG mice with indicated staining are shown. (B) Representative results of histological examination of HCT116 metastatic foci versus MIA PaCa-2 metastatic foci in the livers of NOG hosts with indicated staining are shown. Regions encircled by a broken line demarcate tumors.

FIGURE 6.

Distribution of F4/80+ and Ym1+ cells in MIA PaCa-2 metastatic foci differs from that in HCT116 metastatic foci in NOG hosts. (A) Representative results of histological examination of the livers in intact NOD/SCID versus NOG mice with indicated staining are shown. (B) Representative results of histological examination of HCT116 metastatic foci versus MIA PaCa-2 metastatic foci in the livers of NOG hosts with indicated staining are shown. Regions encircled by a broken line demarcate tumors.

Close modal

There were no obvious differences in M1/M2 polarization between intact NOD/SCID versus NOG mice, whereas remarkable differences in distribution of F4/80+ cells and Ym1+ cells were observed between HCT116 metastatic foci versus MIA PaCa-2 metastatic foci. There were few Ym1-expressing cells in contrast to a discernible number of F4/80+ cells in the HCT116 metastatic foci, roughly consistent with M1 macrophage polarization. Alternatively, there were a considerable number of Ym1-expressing cells not only with but also without F4/80 positivity in the MIA PaCa-2 metastatic foci, suggesting macrophages less polarized to M1 type than those in the HCT116 metastatic foci. The F4/80 cells expressing Ym1 were distinct from typical M2 macrophages. Some of such cells were characteristic of morphologically immature granulocytes and others were mononuclear.

The absence of a subset of DCs producing IFN-γ is also one of the important characteristics of the immunodeficiency due to lack of IL-2Rγc (3, 4). To examine whether the endogenous IFN-γ production contributes to the antitumor effects depending on IL-2Rγc, we attempted to determine the impacts of in vivo neutralization using anti–IFN-γ Abs on antitumor effects occurring in NOD/SCID hosts. However, it was hard to optimize experimental conditions in NOD/SCID mice for stable and reproducible results. Alternatively, we examined effects of mouse IFN-γ administration on metastases of HCT116 cells in NOG hosts depleted or not depleted of macrophages. As shown in Fig. 7, with or without depletion of macrophages, the treatment of NOG hosts with recombinant mouse IFN-γ resulted in a significant decrease in hepatic HCT116 tumor burden but did not achieve such full eradication of tumors as occurred in NOD/SCID hosts. Thus, lack of the IL-2Rγc–dependent elimination of HCT116 cells without requiring NK cells and macrophages appeared to be partially compensated by the mouse IFN-γ administration.

FIGURE 7.

Lack of IL-2Rγc–dependent tumor elimination appears to be partially compensated by IFN-γ administration. Amounts and distributions of metastatic foci were intravitally evaluated 1 wk after the intrasplenic transfer to NOG mice undergoing indicated treatments by detecting the fluorescence of HCT116/Venus cells. Columns show mean; bars indicate SD. *p < 0.05, statistically significant difference when compared with NOG hosts without recombinant mouse IFN-γ treatment or macrophage depletion; p < 0.05, statistically significant difference between recombinant mouse IFN-γ–treated versus nontreated NOG hosts depleted of macrophages. C, centrilobular venules; Cl-L, treatment with Cl2MDP-liposomes to deplete macrophages; C zone, pericentral zone; IFN-γ, treatment with recombinant mouse IFN-γ; M zone, middle zone; P, terminal portal venules; P zone, periportal zone.

FIGURE 7.

Lack of IL-2Rγc–dependent tumor elimination appears to be partially compensated by IFN-γ administration. Amounts and distributions of metastatic foci were intravitally evaluated 1 wk after the intrasplenic transfer to NOG mice undergoing indicated treatments by detecting the fluorescence of HCT116/Venus cells. Columns show mean; bars indicate SD. *p < 0.05, statistically significant difference when compared with NOG hosts without recombinant mouse IFN-γ treatment or macrophage depletion; p < 0.05, statistically significant difference between recombinant mouse IFN-γ–treated versus nontreated NOG hosts depleted of macrophages. C, centrilobular venules; Cl-L, treatment with Cl2MDP-liposomes to deplete macrophages; C zone, pericentral zone; IFN-γ, treatment with recombinant mouse IFN-γ; M zone, middle zone; P, terminal portal venules; P zone, periportal zone.

Close modal

To collect evidence for contribution of the IFN-γ–producing DC subset to the defense against tumors, we examined whether transfer of the DC subset from NOD/SCID mice to NOG hosts can reduce HCT116 metastases. The B220+CD11c+CD122+ phenotype reportedly represents the IFN-γ–producing DC subset, but it was difficult to stably sort high purity and a sufficient number of B220+CD11c+CD122+ cells. Alternatively, B220+CD11c+ cells were sorted from the spleens from NOD/SCID mice to prepare the DC subset used for the transfer to NOG hosts. Consistent with the idea that the DC subset but not NK cells contributes to the innate response with IL-2Rγ function in the absence of adaptive immunity, the metastases tended to be reduced by the transfer of B220+CD11c+ cells from NOD/SCID mice to NOG hosts, although it did not achieve statistical significance (p = 0.2710) (Supplemental Fig. 1). An increased number of B220+CD11c+ cells or higher purity of the IFN-γ–producing DC subset transferred to NOG hosts might provide full findings to clarify whether the DC subset plays a leading role.

The present study demonstrated a line of evidence for contribution of innate effectors other than NK cells and macrophages in the presence of IL-2Rγc versus macrophages in the absence of IL-2Rγc to eliminate human cancer cells. O’Sullivan et al. (14) reported on cancer immunoediting by the innate immune system in the absence of adaptive immunity. The report presented observations that NK cells contribute to regression of syngeneic tumors via their IFN-γ production to induce M1-type macrophages in the presence of IL-2Rγc. The apparent discrepancy from our findings revealing innate response without the need for NK cells and macrophages in the presence of IL-2Rγc seems to result from differences in immunoediting between the settings of syngeneic versus xenogeneic tumors. However, it is difficult to simply compare their findings and our findings because of the differences in experimental procedures. Notably, the in vivo depletion of NK cells was performed using anti-NK1.1 Ab in the study done by O’Sullivan et al. (14). This Ab might also deplete a subset of DCs expressing NK1.1, leading to loss of IFN-γ production by them as well as by NK cells (4). Note that observations from experiments with anti-NK1.1 Ab could not rule out the potential contribution of such DCs to innate response. Additionally, it was not ascertained in the previous study whether induced M1-type macrophages directly participate in the regression of syngeneic tumors, because functional experiments represented by in vivo depletion of macrophages that were performed in the current study were not carried out (14). At present, it appears that differences in immunoediting between the settings of syngeneic versus xenogeneic tumors still remain to be elucidated.

CD47 binds to sirpa to send a signal inhibiting phagocytosis by self-macrophages. Although mouse macrophages basically eliminate xenogeneic cells due to inability of xenogeneic CD47 to bind to mouse sirpa, human CD47 can bind to sirpa encoded by the NOD-derived allele to avoid elimination of human cells by the NOD background macrophages (5, 6). Resembling hepatic metastases of HCT116 cells transferred to NOG hosts in the present study, reconstitution of human platelets and erythrocytes in the periphery was suppressed by host macrophages in contrast to that of leukocytes in the periphery and of all lineages of hematopoietic cells in the marrow in NOG hosts in previous studies (20, 21). Susceptibilities of human cells to elimination by NOG mouse macrophages may differ by cellular origin, differentiation, and neoplasticity.

As with NK cell development, IFN-γ production by DCs requires IL-15 signaling, which involves IL-2Rγc (22, 23). As expected, NOG mice lack IFN-γ produced by DCs (3). More recently, it has been revealed that a subset of DCs producing IFN-γ, which is represented by the CD11c+B220+CD122+ phenotype, is absent from NOG mice (4). The decline in metastatic foci by mouse IFN-γ administration to NOG hosts appeared to reflect compensation for lack of the IL-2Rγc–dependent elimination of HCT116 cells via IFN-γ produced without requiring NK cells and macrophages. The IFN-γ–mediated effects to eliminate HCT116 cells should be exerted by IFN-γ–stimulated host cells owing to the absence of direct bioactivities of mouse IFN-γ on human cells (12). It is plausible that hematopoietic cells, such as granulocytes, and some nonhematopoietic cells are responsible for the antitumor effects induced by mouse IFN-γ administration to hosts lacking both macrophages and IL-2Rγc–dependent effectors. The possible mechanisms of antitumor effects triggered by IFN-γ include direct cellular attacks and indirect destruction via reactive oxygen species and NO produced by IFN-γ-stimulated cells (24). It is of considerable interest that the treatment of NOG hosts with IFN-γ did not achieve such full eradication of transferred HCT116 cells as occurred in NOD/SCID hosts even depleted of both NK cells and macrophages. This incompleteness suggests at least two possibilities. One is that a part of elimination of HCT116 cells may be mediated by IFN-γ–independent mechanisms in the presence of IL-2Rγc. Another is that innate cells producing IFN-γ dependently on IL-2Rγc function, which exist in NOD/SCID but not in NOG mice, may be also stimulated by IFN-γ to completely eradicate HCT116 cells.

IFN-γ has been reported to polarize macrophages toward M1 type (14, 15). However, there were no discernible differences in the M1/M2 macrophage ratio evaluated immunohistochemically between IFN-γ–treated versus nontreated NOG hosts undergoing HCT116 cell transfer (data not shown). Of note, the IFN-γ treatment reduced metastases in NOG hosts depleted of macrophages as well as those not depleted of macrophages. This suggests that skewing of M1/M2 macrophage ratio is less important to the mechanisms of effects of the IFN-γ treatment in NOG hosts. It is likely that M1 macrophage polarization is involved in innate response without, but not with, IL-2Rγ function.

The arrest of HCT116 cells inside sinusoids of periportal regions of hepatic lobules within 30 min after the transfer was not influenced by the presence of macrophages or IL-2Rγc–dependent effectors. The antitumor effects exerted with or without IL-2Rγc function appeared to require a priming phase after the transfer of HCT116 cells, suggesting that stimuli related to the transplantation procedures, such as debris of transferred cells, may trigger innate response to eliminate tumors (25).

The analogous experiments using MIA PaCa-2 cells also revealed that innate response to tumor cells does not require NK cells or macrophages in the presence of IL-2Rγc function, although evidence for suppression of metastases by macrophages in the absence of IL-2Rγc function was lacking, unlike the experiments using HCT116 cells. In intact livers of mice not undergoing transfer of tumor cells, M1/M2 macrophage polarization was roughly similar in NOD/SCID and NOG strains because of lacking obvious differences in the distribution of F4/80+ cells and Ym1-expressing cells in the livers. Alternatively, given that the HCT116 metastases increased by macrophage depletion as well as the discernible predominance of F4/80+ cells lacking Ym1 expression in the HCT116 metastatic foci, M1 macrophages likely contribute to the defense in NOG hosts.

A considerable number of Ym1-expressing cells in MIA PaCa-2 metastatic foci were not only localized out of distribution of F4/80+ cells but were also morphologically distinct from macrophages. Some such cells resembled immature granulocytes. Ym1 has been reportedly expressed not only by macrophages but also by immature granulocytes, monocytes, and myeloid progenitor cells (2628). Myeloid-derived suppressor cells (MDSCs), a heterogeneous population of myeloid cells, including granulocytic or monocytic characteristics, are recognized as a key player in escape from immune response and tumor progression. In our preliminary study, MDSCs defined by coexpression of Gr1 and CD11b, especially granulocytic MDSCs, which highly express Gr1, were increased in the peripheral blood of NOG mice after transfer of MIA PaCa-2 cells, compared with those after transfer of HCT116 cells and with those not undergoing any transfer (data not shown). A previous study has also shown expansion of MDSCs in NOD/SCID mice undergoing s.c. transplantation of MIA PaCa-2 cells (29). Taken together, it would be likely that non–macrophagic cells expressing Ym1 in the MIA PaCa-2 metastatic foci potentially represented MDSCs.

The lack of increase in MIA PaCa-2 metastases in NOG hosts depleted of macrophages would be explained by the idea that MIA PaCa-2 cells overcome or impair the macrophage response. Differences in the sensitivity to tumoricidal activities by macrophages between HCT116 versus MIA PaCa-2 cells still remain to be elucidated, although the predominance of Ym1-expressing cells over F4/40+ cells in the metastatic foci suggests that MDSCs and/or M2 macrophages support MIA PaCa-2 tumor progression via inhibition of M1 macrophage polarization.

Collectively, comparison of observations between HCT116 metastases versus MIA PaCa-2 metastases may be a model to understand immunoediting via modulation of innate responses with or without IL-2Rγ function in tumor-bearing hosts that lack adaptive immunity. Considering the observations from the liver metastasis models using HCT116 cells and MIA PaCa-2 cells, macrophages basically play a role in the defense against tumors in the absence of IL-2Rγc, whereas the antitumor effect can be likely affected by modulation of M1/M2 macrophage polarization by tumor cells.

In conclusion, to our knowledge the present study is the first to assess qualitative and quantitative differences in innate elimination of human cancer cells between the presence and absence of IL-2Rγc function in NOD background mice lacking adaptive immunity. The results provide insights into the nature of innate response in the absence of adaptive immunity, aiding in developing tumor xenograft models in experimental oncology.

We thank Dr. Atsushi Miyawaki (Brain Science Institute, RIKEN, Saitama, Japan) for providing the original vector containing the Venus gene sequence. We also thank Miyuki Kuronuma (Central Institute for Experimental Animals) for expert technical assistance.

This work was supported by the Keio University Global Center of Excellence Program, the Leading Project for Biosimulation, and Japan Society for the Promotion of Science Grant-in-Aid for Scientific Research (C) 25430139, all of which were funded by the Ministry of Education, Culture, Sport, Science and Technology, Japan.

The online version of this article contains supplemental material.

Abbreviations used in this article:

Cl2MDP-liposome

dichloromethylene diphosphonate–containing liposome

DC

dendritic cell

IL-2Rγc

IL-2R common γ-chain

MDSC

myeloid-derived suppressor cell

NOG

NOD/SCID/IL-2Rγcnull

sirpa

signal regulatory protein α.

1
Kelland
L. R.
2004
.
Of mice and men: values and liabilities of the athymic nude mouse model in anticancer drug development.
Eur. J. Cancer
40
:
827
836
.
2
Zhou
Q.
,
Facciponte
J.
,
Jin
M.
,
Shen
Q.
,
Lin
Q.
.
2014
.
Humanized NOD-SCID IL2rg−/− mice as a preclinical model for cancer research and its potential use for individualized cancer therapies.
Cancer Lett.
344
:
13
19
.
3
Ito
M.
,
Hiramatsu
H.
,
Kobayashi
K.
,
Suzue
K.
,
Kawahata
M.
,
Hioki
K.
,
Ueyama
Y.
,
Koyanagi
Y.
,
Sugamura
K.
,
Tsuji
K.
, et al
.
2002
.
NOD/SCID/γcnull mouse: an excellent recipient mouse model for engraftment of human cells.
Blood
100
:
3175
3182
.
4
Ito
R.
,
Katano
I.
,
Ida-Tanaka
M.
,
Kamisako
T.
,
Kawai
K.
,
Suemizu
H.
,
Aiso
S.
,
Ito
M.
.
2012
.
Efficient xenoengraftment in severe immunodeficient NOD/Shi-scid IL2rγnull mice is attributed to a lack of CD11c+B220+CD122+ cells.
J. Immunol.
189
:
4313
4320
.
5
Takenaka
K.
,
Prasolava
T. K.
,
Wang
J. C.
,
Mortin-Toth
S. M.
,
Khalouei
S.
,
Gan
O. I.
,
Dick
J. E.
,
Danska
J. S.
.
2007
.
Polymorphism in Sirpa modulates engraftment of human hematopoietic stem cells.
Nat. Immunol.
8
:
1313
1323
.
6
Yamauchi
T.
,
Takenaka
K.
,
Urata
S.
,
Shima
T.
,
Kikushige
Y.
,
Tokuyama
T.
,
Iwamoto
C.
,
Nishihara
M.
,
Iwasaki
H.
,
Miyamoto
T.
, et al
.
2013
.
Polymorphic Sirpa is the genetic determinant for NOD-based mouse lines to achieve efficient human cell engraftment.
Blood
121
:
1316
1325
.
7
Nagai
T.
,
Ibata
K.
,
Park
E. S.
,
Kubota
M.
,
Mikoshiba
K.
,
Miyawaki
A.
.
2002
.
A variant of yellow fluorescent protein with fast and efficient maturation for cell-biological applications.
Nat. Biotechnol.
20
:
87
90
.
8
Khatib
A. M.
,
Fallavollita
L.
,
Wancewicz
E. V.
,
Monia
B. P.
,
Brodt
P.
.
2002
.
Inhibition of hepatic endothelial E-selectin expression by C-raf antisense oligonucleotides blocks colorectal carcinoma liver metastasis.
Cancer Res.
62
:
5393
5398
.
9
Morris
V. L.
,
Schmidt
E. E.
,
MacDonald
I. C.
,
Groom
A. C.
,
Chambers
A. F.
.
1997
.
Sequential steps in hematogenous metastasis of cancer cells studied by in vivo videomicroscopy.
Invasion Metastasis
17
:
281
296
.
10
Kudo
A.
,
Kashiwagi
S.
,
Kajimura
M.
,
Yoshimura
Y.
,
Uchida
K.
,
Arii
S.
,
Suematsu
M.
.
2004
.
Kupffer cells alter organic anion transport through multidrug resistance protein 2 in the post-cold ischemic rat liver.
Hepatology
39
:
1099
1109
.
11
Fraser
C. C.
,
Chen
B. P.
,
Webb
S.
,
van Rooijen
N.
,
Kraal
G.
.
1995
.
Circulation of human hematopoietic cells in severe combined immunodeficient mice after Cl2MDP-liposome-mediated macrophage depletion.
Blood
86
:
183
192
.
12
Morikawa
K.
,
Fan
D.
,
Denkins
Y. M.
,
Levin
B.
,
Gutterman
J. U.
,
Walker
S. M.
,
Fidler
I. J.
.
1989
.
Mechanisms of combined effects of gamma-interferon and 5-fluorouracil on human colon cancers implanted into nude mice.
Cancer Res.
49
:
799
805
.
13
Lewis
C. E.
,
Pollard
J. W.
.
2006
.
Distinct role of macrophages in different tumor microenvironments.
Cancer Res.
66
:
605
612
.
14
O’Sullivan
T.
,
Saddawi-Konefka
R.
,
Vermi
W.
,
Koebel
C. M.
,
Arthur
C.
,
White
J. M.
,
Uppaluri
R.
,
Andrews
D. M.
,
Ngiow
S. F.
,
Teng
M. W.
, et al
.
2012
.
Cancer immunoediting by the innate immune system in the absence of adaptive immunity.
J. Exp. Med.
209
:
1869
1882
.
15
Holden
J. A.
,
Attard
T. J.
,
Laughton
K. M.
,
Mansell
A.
,
O’Brien-Simpson
N. M.
,
Reynolds
E. C.
.
2014
.
Porphyromonas gingivalis lipopolysaccharide weakly activates M1 and M2 polarized mouse macrophages but induces inflammatory cytokines.
Infect. Immun.
82
:
4190
4203
.
16
Ma
X. L.
,
Li
Y. H.
,
Gao
J. X.
,
Li
J.
,
Guo
L.
,
Wu
C. Z.
.
2008
.
Expression of inducible nitric oxide synthase in the liver is under the control of nuclear factor κB in concanavalin A-induced hepatitis.
J. Gastroenterol. Hepatol.
23
:
e231
e235
.
17
Kimura
K.
,
Nagaki
M.
,
Takai
S.
,
Satake
S.
,
Moriwaki
H.
.
2004
.
Pivotal role of nuclear factor κB signaling in anti-CD40-induced liver injury in mice.
Hepatology
40
:
1180
1189
.
18
Katz
S. C.
,
Pillarisetty
V. G.
,
Bleier
J. I.
,
Shah
A. B.
,
DeMatteo
R. P.
.
2004
.
Liver sinusoidal endothelial cells are insufficient to activate T cells.
J. Immunol.
173
:
230
235
.
19
Herkel
J.
,
Jagemann
B.
,
Wiegard
C.
,
Lazaro
J. F.
,
Lueth
S.
,
Kanzler
S.
,
Blessing
M.
,
Schmitt
E.
,
Lohse
A. W.
.
2003
.
MHC class II-expressing hepatocytes function as antigen-presenting cells and activate specific CD4 T lymphocyutes.
Hepatology
37
:
1079
1085
.
20
Hu
Z.
,
Yang
Y. G.
.
2012
.
Full reconstitution of human platelets in humanized mice after macrophage depletion.
Blood
120
:
1713
1716
.
21
Hu
Z.
,
Van Rooijen
N.
,
Yang
Y. G.
.
2011
.
Macrophages prevent human red blood cell reconstitution in immunodeficient mice.
Blood
118
:
5938
5946
.
22
Ogasawara
K.
,
Hida
S.
,
Azimi
N.
,
Tagaya
Y.
,
Sato
T.
,
Yokochi-Fukuda
T.
,
Waldmann
T. A.
,
Taniguchi
T.
,
Taki
S.
.
1998
.
Requirement for IRF-1 in the microenvironment supporting development of natural killer cells.
Nature
391
:
700
703
.
23
Ohteki
T.
,
Suzue
K.
,
Maki
C.
,
Ota
T.
,
Koyasu
S.
.
2001
.
Critical role of IL-15-IL-15R for antigen-presenting cell functions in the innate immune response.
Nat. Immunol.
2
:
1138
1143
.
24
Di Carlo
E.
,
Forni
G.
,
Lollini
P.
,
Colombo
M. P.
,
Modesti
A.
,
Musiani
P.
.
2001
.
The intriguing role of polymorphonuclear neutrophils in antitumor reactions.
Blood
97
:
339
345
.
25
He
H.
,
Stone
J. R.
,
Perkins
D. L.
.
2002
.
Analysis of robust innate immune response after transplantation in the absence of adaptive immunity.
Transplantation
73
:
853
861
.
26
Hung
S. I.
,
Chang
A. C.
,
Kato
I.
,
Chang
N. C.
.
2002
.
Transient expression of Ym1, a heparin-binding lectin, during developmental hematopoiesis and inflammation.
J. Leukoc. Biol.
72
:
72
82
.
27
Harbord
M.
,
Novelli
M.
,
Canas
B.
,
Power
D.
,
Davis
C.
,
Godovac-Zimmermann
J.
,
Roes
J.
,
Segal
A. W.
.
2002
.
Ym1 is a neutrophil granule protein that crystallizes in p47phox-deficient mice.
J. Biol. Chem.
277
:
5468
5475
.
28
Nio
J.
,
Fujimoto
W.
,
Konno
A.
,
Kon
Y.
,
Owhashi
M.
,
Iwanaga
T.
.
2004
.
Cellular expression of murine Ym1 and Ym2, chitinase family proteins, as revealed by in situ hybridization and immunohistochemistry.
Histochem. Cell Biol.
121
:
473
482
.
29
Zheng
D.
,
Chen
H.
,
Bartee
M. Y.
,
Williams
J.
,
Davids
J. A.
,
Lomas
D. A.
,
McFadden
G.
,
Lucas
A. R.
.
2013
.
Myxomaviral anti-inflammatory serpin reduces myeloid-derived suppressor cells and human pancreatic cancer cell growth in mice.
J Cancer Sci Ther
5
:
291
299
.

The authors have no financial conflicts of interest.

Supplementary data