Endogenous IL-12 triggers an antiangiogenic program in melanoma cells

Abstract

The IL12RB2 gene acts as a tumor suppressor in human B cell malignancies. Indeed, Il12rb2 knockout (KO) mice develop spontaneously B cell tumors, but also lung epithelial tumors. This latter phenotype may be related to (i) impairment of host IL-12-mediated immunosurveillance and/or (ii) IL-12 inability to inhibit directly the growth of IL-12 unresponsive malignant cells. To address this issue, we transplanted IL-12R⁺ B16 melanoma cells into syngeneic Il12rb2 KO mice with the following rationale: (i) these mice have severe defects in IFN-γ production, as well as in cytotoxic T lymphocyte and natural killer cell cytotoxicity, and (ii) they produce but do not use IL-12 that can potentially bind to and target tumor cells only. Il12rb2 KO mice displayed higher endogenous serum levels of IL-12 and developed smaller B16 tumors than WT animals. These tumors showed reduced proliferation, increased apoptosis, and defective microvessel formation related to down-regulated expression of a set of proangiogenic genes previously unrelated to IL-12. Such effects depended on direct activity of endogenous IL-12 on tumor cells in KO mice, and hydrodynamic delivered IL-12 caused further reduced tumorigenicity of B16 cells in these mice. A previously undescribed mechanism of the IL-12 antitumor activity has been here identified and characterized.

Interleukin 12 is a heterodimeric cytokine bridging innate and adaptive immunity (1). IL-12, which belongs to a family of structurally related cytokines including IL-23 and IL-27 (2, 3), is formed by the IL12p35 and IL12p40 subunits (4) and binds to the IL-12R, composed of the β1 and the β2 chains. Both chains are needed for high affinity binding of the cytokine and initiation of signal transduction (5, 6). IL12p40 associates with the p19 subunit to form IL-23, which binds to a receptor composed of IL12Rβ1 and IL-23R (2, 3). Therefore, the IL12p35 and the IL12Rβ2 subunits are unique components of IL-12 and IL-12R, respectively.

IL-12, which is produced predominantly by antigen-presenting cells, drives T helper (Th1) responses, enhances T and natural killer (NK) cell cytotoxicity, and induces IFN-γ production by T and NK cells (4, 7–9). In addition, we have shown that IL-12 contributes to the regulation of normal human B cell function through stimulation of IgM synthesis and IFN-γ production (10).

IL-12 exerts antitumor activity through IFN-γ-dependent and independent mechanisms (11–15), which include modulation of the immune system and antiangiogenesis (16–18). We have recently demonstrated that the IL12RB2 gene acts as a tumor suppressor in human chronic B cell malignancies (14) and, accordingly, that aging IL12rb2 knockout (KO) mice develop spontaneously B cell tumors. The same mice showed high incidence of lung epithelial tumors. This latter finding may be related to reduced immune surveillance, because of their deficient T and NK cell-mediated cytotoxicity (19), Th2-biased T cell differentiation (20), and/or lack of IL-12-mediated direct inhibition of tumor growth. To analyze these different possibilities, we have investigated the tumorigenicity of a transplanted IL12RB2⁺ syngeneic tumor in IL12rb2 KO mice. This experimental model system allows one to assess the effects of endogenous IL-12, not used by the host cells, on tumor growth in the absence of IL-12-driven immune modulation.

Results

Endogenous IL-12 Dampens Tumorigenicity of B16 Melanoma Cells Expressing IL-12R.

We first investigated expression of the IL12rb2 gene by RT-PCR in a panel of tumor cell lines with C57BL/6J genetic background. These tumor cell lines included MCA38 colon carcinoma cells, T1525 prostate cancer cells, B16 and B16F10 melanoma cells, and TC1 lung carcinoma cells. All of these cell lines tested positive for IL12rb2 expression (data not shown). We next focused on B16 melanoma cells and showed that they expressed surface IL12Rβ1 chain (Fig. 1A Left). Fig. 1A Right shows detection of IL12rb2 transcript in B16 cells. These results indicated that the latter cells expressed the complete IL-12 receptor.

Fig. 1.

Characterization and tumorigenicity of B16 cells and IL12p70 serum level in KO mice. (A) Expression of IL12rb1 and b2 chains in B16 melanoma cells. (Left) IL12rb1 surface expression in B16 cells, as assessed by flow cytometry. Open profile: IL12rb1 staining; dark profile: isotype matched mAb staining. (Right) IL12rb2 expression in B16 cells, as assessed by RT-PCR. MW, molecular weight; NC, negative control (water in the place of cDNA); PC, positive control (3T3 cell line); B16 cells, B16 melanoma cells (B16). On the right, the expected mw of the amplified band is shown. (B) Quantitative determination of IL-12p70 concentrations in sera from 10 KO mice, as assessed by ELISA. (C) Volume of s.c. tumors grown in KO and WT animals 17 days (Left) and 7 days (Right) after B16 cell inoculation. The differences in size between tumors removed from KO and WT mice either 17 days or 7 days after B16 cell inoculation were evaluated by Mann–Whitney U test. Boxes indicate values between the 25^th and 75^th percentiles, whisker lines represent highest and lowest values for each group. Horizontal lines represent median values.

Next IL12p70 serum level was measured in KO and WT animals. Fig. 2B shows that IL12p70 concentration in sera from ten KO animals ranged from 15.6 to 46.6 pg/ml. In contrast, IL12p70 was not detected in sera from WT mice.

Fig. 2.

Histological feature and vascularity of B16 tumors in KO and WT mice. (A) Histological feature of B16 tumors grown after 7 days in WT (Aa and Ac) and KO (Ab and Ad) mice. s.c. injection of B16 cells in both WT (Aa) and KO (Ab) mice gives rise to solid tumor masses showing a monomorphous pattern of small round neoplastic cells. However, differently from tumors developed in WT mice, those from KO mice (Ab) are frequently altered by extensive areas of ischemic-hemorrhagic necrosis (N). This condition is associated with a deficiency in microvessel network (Ad) which, by contrast, seems to be well developed in tumors from WT mice (Ac). (Magnification: ×200 in Aa–Ad.) (B) Vascular network in B16 tumors grown after 7 days in WT and KO mice. The numerous microvessels supporting tumor developed in WT mice are endowed with complete and robust basement membrane, evidenced by laminin immunostaining (Ba), and with pericyte covering, evidenced by α-sma immunostaining (Bc). By contrast, in tumor from KO mice, both laminin deposition (Bb) and pericyte covering (Bd) are dramatically compromised. Furthermore, the production and perivascular accumulation of tenascin-C is prominent in tumor from WT mice (Be) and deficient in tumor from KO mice (Bf).

Fourteen KO and 14 WT C57BL/6J mice were injected s.c. with B16 melanoma cells and killed after 17 days. As shown in Fig. 1C Left, tumors from KO mice were significantly smaller than those from WT mice (n = 14 for both groups; WT, median 4055.5; range 1528–6500; KO, median 1914.5; range 100-2120; P = 0.001).

Next, histological and immunohistochemical studies were performed in tumors removed from 12 KO and 12 WT mice 7 days after B16 cell inoculation. Seven day tumors from KO animals were already significantly smaller than those from WT animals (Fig. 1C Right, n = 12 for both groups; WT, median 179.5; range 69–440; KO, median 55; range 20–205; P = 0.0014).

In contrast to tumors from WT mice (Fig. 2Aa), tumors from KO mice were characterized by a wide focus of ischemic-coagulative necrosis (Fig. 2Ab). Microvessel density was also significantly reduced in tumors from KO mice (Fig. 2Bd and Table 1) vs. WT mice (Fig. 2Ac and Table 1), as assessed by CD31 immunostaining. Finally, tumors from KO animals showed a significantly lower proliferation index (Table 1), as assessed by proliferating cell nuclear antigen (PCNA) immunoreactivity, and a significant increase in the number of apoptotic cells (Table 1), as assessed by the ApopTag assay. Analysis of tumor associated leukocytes showed a scanty infiltrate, mainly comprised of macrophages, at the edge of tumors from both WT and KO mice, without any appreciable difference.

Table 1.

Immunohistochemical analysis of tumors developed 7 days after B16 cell injection in WT or KO mice

	WT	KO
Microvessel count	16.5 ± 2.2	12.2 ± 1.8*
Proliferation index, %	71.8 ± 9.4	49.5 ± 7.2*
Apoptotic index, %	10.7 ± 2.3	15.1 ± 3.0*

Cell and microvessel count was performed as reported in Methods. Results are expressed as mean ± SD of positive cells or microvessels per field evaluated on formalin-fixed or cryostat (microvessels) sections by immunohistochemistry.

*Values significantly different (P ≤ 0.005) from corresponding values in tumors developed in WT mice.

To gain more insight into the mechanisms underlying the decreased microvessel density in B16 tumors from KO mice, we investigated expression of three angiogenesis-related markers, i.e., laminin, α-smooth muscle actin (α-sma) and tenascin-C (21, 22). Fig. 2B shows that the defective neovascularization of B16 tumors implanted in KO mice was associated with a deficiency in mature, laminin lined (Fig. 2 Ba and Bb), pericyte (αsma⁺ cells) covered (Fig. 2 Bc and Bd) microvessels, and a marked reduction in perivascular and stromal storage of tenascin–C (Fig. 2 Be and Bf).

The IFN-γ inducible antiangiogenic chemokines CXCL9, CXCL10, and CXCL11 as well as iNOS, IFN-γ, and TNF-α were not detected by specific immunohistochemical stainings in either group of mice. Thus, the decreased microvessel density in tumors from KO mice was apparently unrelated to the typical antiangiogenic pathway initiated in the tumor microenvironment by IL-12.

Direct Activity of Murine Recombinant (mr) IL-12 on B16 Melanoma Cells in Vitro.

Four IL12rb2⁺ and 4 IL12rb2⁻ clones were generated from the B16 cell line by limiting dilution (Fig. 3A) and used for further experiments. IL12RB2 expression was checked every three days in all of these cultured clones and remained unaltered until at least 3 weeks. The IL12rb2 transcripts reappeared in IL12rb2⁻ clones starting from day 24 in culture.

Fig. 3.

In vitro characterization of IL12rb2⁺ and IL12rb2⁻ B16 clones. (A) Expression of IL12rb2 mRNA in B16 cells, and in IL12rb2⁺ and IL12rb2⁻ B16 clones, as assessed by RT-PCR. MW, molecular weight; NC, negative control (water in the place of cDNA); PC, positive control (3T3 cell line); four different IL12rb2⁺ B16 clones (clones 1 to 4) and four different IL12rb2⁻ B16 clones (from clones 5 to 8). On the right, the expected mw of the amplified band is shown. (B) Angiogenic activity of supernatants from one representative IL12rb2⁺ (clone 1) and one IL12rb2⁻ clone (clone 7) cultured in the presence or absence of mrIL12. CAMs treated with sponges loaded with the conditioned medium from the IL12rb2⁺ clone were surrounded by allantoic vessels developing radially toward the implant in a “spoked-wheel” pattern (Ba). When medium from IL12rb2⁺ clone cultured with mrIL12 was tested, a significant reduction of the angiogenic response was evident (Bb). Supernatants from IL12rb2⁻ B16 clone cultured in the presence (Bd) or absence (Bc) of mrIL12 elicited an angiogenic response superimposable to that induced by medium from the untreated IL-12rb2⁺ B16 clone (see Ba). One representative experiment is shown. (Original magnification: ×50.) (C) Volume of s.c. tumor grown in KO animals 12 days after IL12rb2⁺ or IL12rb2⁻ B16 clone inoculation. (Left) Differences in size of tumors formed by each individual IL12rb2⁺ and IL12rb2⁻ B16 clone is shown. Five animals were inoculated with each clone. Whisker lines represent highest and lowest values for each group. Horizontal lines represent median values. (Right) Differences between tumors formed by the four pooled IL12rb2⁺ and the four pooled IL12rb2⁻ B16 clones in KO animals (Left) were evaluated by Mann–Whitney U test. Boxes indicate values between the 25th and 75th percentiles, whisker lines represent highest and lowest values for each group. Horizontal lines represent median values. (D) Mouse Angiogenesis RT-PCR-Array on tumors explanted from KO animals 12 days after IL12rb2⁺ and IL12Rb2⁻ B16 clone inoculation. Columns represent fold differences of individual gene expression between tumors formed by the IL12rb2⁺ clone 1 and IL12rb2⁻ clone 5. Upper columns (>1.00) represent genes whose expression is up-regulated, lower columns (<1.00) genes whose expression is down-regulated. One experiment representative of the two performed with similar results using two different IL12rb2⁺ clones and two different IL12rb2⁻ clones is shown.

Next, the angiogenic activity of B16 cell clones was investigated in the chorioallantoic membrane (CAM) assay by using supernatants from two IL12Rb2⁺ and two IL12rb2⁻ B16 cell clones, cultured in the presence or absence of mrIL12.

CAM treated with sponges loaded with VEGF (positive control) or with medium from IL12rbB2⁺ B16 clones were surrounded by allantoic vessels developing radially toward the implant in a “spoked-wheel” pattern. In the representative experiment shown in Fig. 3Ba, the mean number of vessels formed in the presence of medium from the IL12rb2⁺ B16 clone 1 was 28 ± 4 whereas that formed in the presence of VEGF was 30 ± 5 (data not shown). No vascular reaction was detected around the sponges upon exposure to mrIL12 diluted in medium at the same final concentration used to treat tumor cells (mean number of vessels = 7 ± 3 in the presence or absence of mrIL12, data not shown). When the medium from IL12rb2⁺ B16 cell clones that had been incubated with mrIL12 was tested in the CAM assay, a significant (P < 0.001) reduction of the angiogenic response was appreciable (mean number of vessels = 13 ± 3), as compared with positive control (one representative experiment is shown in Fig. 3Bb). The medium from IL12rb2⁻ B16 cell clones contained an angiogenic activity (mean number of vessels = 26 ± 4) (Fig. 3Bc) similar to that of untreated IL12Rb2⁺ B16 clones (see Fig. 3Ba). This activity was left unaltered by preincubation of the IL12rb2⁻ B16 clones with mrIL12 (mean number of vessels = 25 ± 3) (one representative experiment performed is shown in Fig. 3Bd).

Taken together, these results demonstrated unambiguously that IL-12 affected in vitro angiogenic activity of IL12rb2⁺ B16 melanoma cells through direct interaction with IL-12R.

IL12Rβ2⁺ Clones Are Significantly Less Tumorigenic than IL12rb2⁻ B16 clones in KO Animals.

Next, we tested the tumorigenicity of four IL12rb2⁺ and four IL12rb2⁻ B16 clones in KO mice (Fig. 3C). Eight groups (one group for each clone) of 5 KO mice were injected s.c. and killed after 12 days. Tumors were removed, measured, and checked for the IL12rb2 expression by RT-PCR. At this time point, all of the IL12rb2⁻ clones tested negative and all IL12rb2⁺ tested positive. Fig. 3C Left shows that tumors grown after IL12rb2⁺ B16 cell clone inoculation were smaller than those formed by IL12rb2⁻ clones (n = 5 for each clone; clone 1, median 319.5; range 173–512; clone 2, median 228.5; range 54–1149; clone 3, median 531.5; range 334-1230; clone 4, median 614.5; range 286–867; clone 5, median 758; range 663-1277; clone 6, median 712.5; range 371-2277; clone 7, median 1433.5; range 878-2759; clone 8, median 1334; range 831-1893). Differences between tumors formed by the four IL12rb2⁺ and the four IL12rb2⁻ B16 clones in KO animals were statistically significant, as evaluated by Mann–Whitney U test (Fig. 3C Right, n = 20 for both groups; IL12rb2⁺ clones, median 382.5; IL12rb2⁻; B16 clones, median 949.5; P = 0.0054).

Tumors Formed by IL12rb2⁺ B16 Clones in KO Mice Display Down-Regulated Expression of Multiple Proangiogenic Genes.

The angiogenic phenotype of tumors formed in KO mice by two of the above IL12rb2⁺ and two of the above IL12rb2⁻ B16 cell clones (see previous paragraph) was next investigated by using a PCR array kit.

The representative experiment (of the two performed) shown in Fig. 3D demonstrates that the expression of different proangiogenic genes was significantly down-regulated and that of few antiangiogenic genes was significantly up-regulated in IL12rb2⁺ vs. IL12rb2⁻ B16 tumors. Thus, expression of cadherin-5, laminin, ephrin A1, ephrin B2, ephrin receptor B4, FGF6, FGFR3, and neuropilin, whose relationship with IL-12 has never been reported, was decreased. These molecules are expressed during tumor neo-vascularization and promote organization, survival or migration of endothelial cells (23–31). Other, IL-12-related proangiogenic molecules down-regulated in IL12rb2⁺ vs. IL12rb2⁻ tumors were angiopoietin 1 and 2, eotaxin, procollagen type XVIIIα1, cox1, MMP9, and sphyngosin kinase 1 (32–38). Finally, IFN-γ and epiregulin, a major regulator of vascular smooth muscle cell dedifferentiation (39, 40), were found to be up-regulated 4.5- and 3.5-fold, respectively, in IL12rb2⁺ vs. IL12rb2⁻ tumors grown in KO mice (Fig. 3D). These findings support the hypothesis that endogenous IL-12 inhibited directly the growth of IL12rb2⁺ tumors in KO mice through activation in tumor cells of a previously undescribed antiangiogenic program.

IL12rb2⁻ B16 Melanoma Cells Are Equally Tumorigenic in KO and WT Mice.

Next, two IL12rb2⁻ B16 clones were injected in KO and WT mice, and animals were monitored for tumor growth. Tumors removed from KO and WT mice 7 days (Fig. 4A) or 17 days (data not shown) after tumor cell inoculation did not differ significantly in size (Fig. 4A) or histological and immunohistochemical features (Fig. 4B). Thus, no evident areas of necrosis were detected in tumors formed by the IL12rb2⁻ clones in both WT and KO mice (Fig. 4 Ba and Bb) and no defects in microvessel formation were observed in the same tumors, as demonstrated by laminin staining (Fig. 4 Bc and Bd).

Fig. 4.

Size and immunohistological features of tumors formed in WT vs. KO mice by an IL12rb2⁻ and an IL12rb2⁺ B16 clone 7 days after inoculation. (A) The difference in size between IL12rb2⁺clone 1 tumors from KO and WT mice was statistically significant (Mann–Whitney U test). In contrast, tumors formed in the two groups of mice by the IL12rb2⁻ B16 clone 7 were of similar size. Boxes indicate values between the 25th and 75th percentiles, whisker lines represent highest and lowest values for each group. Horizontal lines represent median values. (B) Morphological features of tumors grown 7 days after IL12rb2⁻ B16 clone 7 injection in WT and KO mice. Both tumors grown in WT (Ba) and in KO (Bb) mice show frequent mitosis (arrows) without signs of ischemia or necrosis. Their histological pattern is similar to that of tumors produced by B16 cell injection (see Fig. 1). A vigorous and mature microvessel network supports tumor growth in both WT (Bc) and KO (Bd) mice, as revealed by laminin immunostaining. (Magnification: ×400 in B.)

In contrast, the difference in size between IL12rb2⁺clone 1 tumors from KO and WT mice was statistically significant (n = 7 for both groups; IL12rb2⁺clone 1 in WT, median 179; range 85–262; IL12rβ2⁺clone 1 in KO, median 50, range 20–142; P = 0.0001).

Hydrodynamic Delivered IL-12 Further Decreases the Tumorigenicity of IL12rb2⁺, but Not IL12rb2⁻, B16 Cells in KO Mice.

The above data demonstrated a direct antitumor effect of endogenous IL-12 on IL12rb2⁺, but not IL12rb2⁻, B16 tumors in KO mice. The following experiments were carried out to investigate whether increased endogenous levels of IL-12 in these mice would translate into increased inhibition of IL12rb2⁺, but not IL12rb2⁻, tumor growth.

To this end, we have used a hydrodynamic gene delivery system allowing us to express biologically active IL-12 protein in vivo at high concentration (22). KO mice were inoculated with pcDEF/CMV-GFP (control) or with pcDEF/CMV-GFP-IL12 and after 2 h were injected with IL12rb2⁺ or IL12Rb2⁻ B16 clones. In sera from IL12rb2⁺ tumor transplanted KO mice, IL-12 plasmid administration was accompanied by rapid production of IL-12 protein, which was detected at high level after 6 h (range from 560 to 1,087 ng/ml, n = 6) and decreased progressively reaching the lowest concentration on day 7 (range 43.5 to 324.6 ng/ml, n = 6) (Fig. 5A Left). In contrast, KO mice inoculated with IL-12 plasmid and transplanted with IL12rb2⁻ B16 clones (Fig. 5A Right) showed high serum levels of IL-12 protein after 6 h (range 1275–1865 ng/ml, n = 3) that remained roughly stable until day 7. KO mice injected with the pcDEF/CMV-GFP control plasmid and transplanted with IL12rb2⁺ (n = 6) or IL12rb2⁻ (n = 6) clones showed serum IL-12 levels ranging from 13 to 39 pg/ml at any time tested and overlapping with those detected in untreated KO mice (see Fig. 1B). Altogether, these results suggested that IL-12 was sequestered by IL12rb2⁺ but not IL12rb2⁻ B16 tumors in KO mice.

Fig. 5.

IL12p70 serum concentration and tumorigenicity of IL12rb2⁺ and IL12rb2⁻ B16 clones in KO mice injected with IL-12p70 containing plasmid. (A) IL12p70 serum concentration was evaluated 6 h, 3 days, and 7 days after inoculation of IL-12p70 plasmid and B16 clones, as assessed by ELISA. (Left) IL-12p70 serum levels from three different animals injected first with IL-12p70 plasmid and 2 h later with IL-12rb2⁺ clone 1 (▵, ○, □) or IL-12rb2⁺ clone 2 (•, *, ■), respectively. (Right) IL12p70 serum levels from six different animals injected first with IL12p70 plasmid and 2 h later with the IL12rb2⁻ B16 clone 5 (black symbols) or IL12rb2⁻ B16 clone 7 (white symbols). (B) Tumors grown 7 days after inoculation of the IL12rb2⁺ clones 1 and 2 in IL12p70 delivered (IL-12) KO mice were significantly smaller than those formed by the same clones in animals injected with control plasmid (GFP) (Mann–Whitney U test). Boxes indicate values between the 25th and 75th percentiles, whisker lines represent highest and lowest values for each group. Horizontal lines represent median values. In contrast, no differences in size were observed between tumors formed by the IL12rRb2⁻ clone 6 in KO mice inoculated with IL-12 or control plasmid.

KO mice treated with hydrodynamic delivered IL-12 showed significantly smaller IL12rb2⁺ B16 tumors (n = 7 for each clone; clone 1, median 26.5; range 25.5–35; clone 2, median 23.5; range 14.5–33.4) than animals treated with control plasmid (n = 7 for each clone; clone 1, median 97.6; range 35–154; clone 2, median 88.6; range 70.4–172.5) (Fig. 5B). Differences were statistically significant (P = 0.0286 for clone 1, and 0.0031 for clone 2). In contrast, IL12rb2⁻ B16 tumors grown in IL-12 plasmid or control plasmid treated KO mice did not differ significantly in size (Fig. 5B).

Taken together, these experiments provide unambiguous proof of the concept that IL-12 acts directly on IL12rb2⁺ B16 tumors to restrain their growth.

Discussion

In this study, we demonstrate that the in vivo growth of B16 melanoma tumor was significantly reduced in Il12rb2 KO mice in comparison with their WT counterparts. The former mice have severe defects in IFN-γ production, as well as in cytotoxic T lymphocyte (CTL) and NK cytotoxicity (8, 19) and, as shown here, produce but do not use IL-12, that can bind to and target tumor cells only.

The following findings support the latter statement: (i) IL12rb2⁺ B16 clones were significantly less tumorigenic than IL12rβ2⁻ clones in KO mice, (ii) experiments with hydrodynamic IL-12 delivery showed that the growth of IL12rb2⁺ B16 tumors only was further reduced in IL-12 plasmid in comparison with control plasmid-treated KO mice, indicating a dose dependent potentiation of tumor inhibition by IL-12, and (iii) IL12rb2⁻ clones were equally tumorigenic in KO and WT mice.

The mechanisms of IL-12-mediated antitumor activity depend not only on activation of innate and adaptive effector immune mechanisms, but also on inhibition of angiogenesis (16–18). The latter is mediated by induction of IFN-γ, which in turn leads to release of the antiangiogenic chemokines CXCL9, CXCL10, and CXCL11. An additional antiangiogenic mechanism operated by IL-12 is down-regulation of VEGF and FGF-2 production, which is partly IFN-γ dependent (17, 41, 42).

Here we show that, upon in vitro incubation of IL12Rb2⁺ B16 cells with exogenous IL-12, supernatants from the latter cells contained lower angiogenic activity than IL12rb2⁻ B16 cells. Furthermore, histological and immunohistochemical studies of B16 melanomas from KO mice demonstrated defective microvessel formation, together with decreased proliferation and enhanced apoptosis of tumor cells. These results suggested that, in KO mice, tumor targeted endogenous IL-12 inhibited the growth of malignant cells through different mechanisms.

To gain more insight into IL-12 driven angiogenesis inhibition in our experimental model, we first investigated by immunohistochemistry expression of IFN-γ, TNF-α, the antiangiogenic chemokines CXCL9, CXCL10, CXCL11, and iNOS in tumor tissue sections from KO and WT mice. None of these proteins was detected in any sample from either group.

Angiogenesis PCR array analysis of tumors formed in KO mice by IL12rb2⁺ vs. IL12rb2⁻ B16 clones demonstrated a significant down-regulation in the former tumors of various proangiogenic factors rather than an up-regulation of antiangiogenic molecules. Two groups of proangiogenic factors were decreased in IL12rb2⁺ tumors, the first included molecules that have never been associated before to IL-12 (cadherin-5, laminin, ephrin A1, ephrin B2, ephrin receptor B4, FGF6, FGFR3, and neuropilin), the second encompassed molecules already related to IL-12 (angiopoietin 1 and 2, eotaxin, procollagen type XVIIIα1, cox1, MMP9, and sphyngosin kinase 1) (25–42).

Consistent with molecular analysis, immunohistochemical studies showed impaired expression of laminin, and in addition of tenascin-C and α-sma, in B16 tumors from KO mice. Laminin and tenascin-C are detected at the site of active neovascularization, and enhance the proangiogenic effects of various growth factors (21, 22). α-Sma is a specific marker for pericytes, which form the outer layer of capillary vessels in physiological conditions.

Although the antiangiogenic effects in KO mice were initiated by binding of endogenous IL-12 to IL12R⁺ tumor cells, the possibility that the latter cells recruited other cell types from the tumor microenvironment in the antiangiogenic program cannot be ruled out. However, the minimal infiltrates detected in tumors from both KO and WT mice were comprised of macrophages, but not of CTL or NK cells, which are functionally defective in KO mice (19).

Our data, together with some previously reported suggestive findings (43–45), delineate a scenario whereby two categories of malignancies can be envisaged, i.e., tumors characterized by silencing of the IL12rb2 gene (e.g., human B cell lymphoproliferative disorders), and tumors expressing the complete IL-12R, that may represent a potential therapeutic target. In this study, the latter scenario was a common feature because complete IL-12R was detected in different epithelial tumors on a C57BL/6J background.

In conclusion, a reappraisal of IL-12 as therapeutic agent for IL-12R expressing human malignancies is needed, based upon combination of the well known immunomodulatory effects of the cytokine with its direct activity on tumor cells. A caveat may come from the toxicity that IL-12 has shown in numerous clinical trials (12), but this drawback can be circumvented by targeting directly the cytokine to tumor cells. Finally, our experimental model suggests that low concentrations of tumor targeted IL-12 may be sufficient to inhibit tumor growth and angiogenesis.

Methods

Cell Culture, Antibodies, Reagents, and Flow Cytometry.

B16 cells were cultured in RPMI medium 1640 with 10% FCS (Seromed-BiochromKG, Berlin, Germany).

mrIL12 (Peprotec EC, London, U.K.) was used in vitro at the concentration of 2 ng/ml. Flow cytometry experiments were performed by using a PE-conjugated anti-mouse IL12rb1 monoclonal antibody (Ab) (BD PharMingen, San Jose, CA). PE-conjugated, isotype-matched monoclonal Ab of irrelevant specificity (Caltag, Burlingame, CA) was used as control. Cells were scored by using a FACSCalibur analyzer (BD) and data processed by using CellQuest software (BD) (20).

IL12p70 ELISA.

Peripheral blood was collected from the retro-orbital sinus and centrifuged at 500 × g for 10 min at 4°C. Serum was isolated and stored at −20°C. IL12p70 ELISA was performed by using mouse IL12p70 Quantikine kit (R&D Systems, Abington, U.K.) following the manufacturer's protocol.

B16 Subcloning.

B16 melanoma cells were seeded in 96-well U-bottom plates at a concentration of 1 cell per well in 0.2 ml of complete medium. After 9 days, clones were checked for IL12rb2 mRNA expression by RT-PCR every 3 days by using the conditions and primers published elsewhere (20).

Mice Studies.

Eight- to 10-week-old KO C57BL/6J (The Jackson Laboratory, Bar Harbor, ME) and WT C57BL/6J mice (Harlan Italy, San Pietro al Natisone, Italy) were housed under specific pathogen-free conditions. All procedures involving animals were performed in the respect of the National and International current regulations (D.l.vo 27/01/1992, n.116, European Economic Community Council Directive 86/609, OJL 358, Dec. 1, 1987).

Groups of 7–12 KO or WT animals, were injected s.c. with 5 × 10⁶ B16 cells, IL12rb2⁺ or IL12rb2⁻ clones and killed by excess ethyl-ether anesthesia at different time points (7, 9, 12, or 17 days). Tumors were removed, measured as described (14), and processed for histological and immunohistochemical analyses.

Tissue Studies.

Tissue samples were fixed in 4% neutral buffered formalin, embedded in paraffin, sectioned at 4 μm, and stained with H&E. For immunohistochemistry, formalin-fixed, paraffin-embedded (PCNA and ApopTag) or cryostat sections were immunostained with the following anti-mouse Ab: CD11c (Chemicon International, Temecula, CA); CD11b, CD8, and CD4 (Sera-lab, Crawley Down, U.K.); anti-GR1 (granulocytes) (ATCC) and CD31 (BD); anti-NK (Wako Chemicals GmbH, Neuss, Germany), anti-laminin (Biogenex, San Ramon, CA), anti-tenascin-C (Abcam, Cambridge, U.K.), anti-CXCL10 (PeproTech), anti-CXCL9 (R&D Systems), anti-TNF-α (Immuno Kontact, Frankfurt, Germany), anti-CXCL11, anti-iNOS (Transduction Laboratories, Lexington, KY), anti-IFN-γ (Santa Cruz Biotechnology Inc., Santa Cruz, CA), anti-PCNA (Dako Cytomation, Glostrup, Denmark), and anti-αsma (Dako) antibodies.

The rate of proliferating cells, as assessed by immunoreactivity for PCNA, and of apoptotic cells, as assessed by the ApopTag assay, were determined as described (14). Microvessels were counted in 10 randomly chosen fields under a microscope ×400 field (×40 objective and ×10 ocular lens; 0.180 mm² per field). The rates of proliferating cells, as assessed by immunoreactivity for PCNA, and of apoptotic cells, as assessed by the Apoptag assay, were determined by counting the number of positive cells/number of total cells in the viable neoplastic tissue excluding areas of tissue necrosis under a microscope ×600 field (×60 objective and ×10 ocular lens; 0.120 mm² per field).

CAM Assay.

Fertilized White Leghorn chicken eggs (20 per group) were incubated at 37°C at constant humidity. On day 3, a square window was opened in the shell, and 2–3 ml of albumen was removed to allow detachment of the developing CAM. The window was sealed with a glass, and the eggs were returned to the incubator. On day 8, eggs were treated with 1 mm³ sterilized gelatin sponges (Gelfoam Upjohn, Kalamazoo, MI) placed on top of the growing CAM, as reported (46), and loaded with 1 μl of PBS (negative control); 1 μl of PBS with 250 ng VEGF (R&D Systems) as positive control; 1 μl of medium from IL-12rb2⁺ clones cultured 48 h with or without mrIL12; 1 μl of medium from IL12rb2⁻ B16 clones cultured 48 h with or without mrIL12; 1 μl of medium containing mrIL12. All supernatants were tested in triplicate and means ± SD were calculated. CAM were examined daily until day 12 and photographed in ovo with a stereomicroscope equipped with a camera and image analyzer system (Olympus Italia, Italy). On day 12, the angiogenic response was evaluated by the image analyzer system as the number of vessels converging toward the sponges.

Angiogenesis PCR-Array.

RNA was extracted from tumors removed from KO animals after 12 days injection of two IL12rb2⁺ and two IL12rb2⁻ B16 cell clones, by using TRIzol from Invitrogen (Carlsbad, CA) and retrotranscribed by RT-PCR kit (Clontech, Mountain View, CA). Genomic DNA contamination was eliminated by Dnase treatment by using Rneasy Micro Kit (Qiagen GmbH, Hilden, Germany) and IL-12Rb2 expression was tested by PCR before starting PCR-Array procedure. Mouse Angiogenesis RT² Profiler PCR Array and RT² Real-Timer SyBR Green/ROX PCR Mix were purchased from SuperArray Bioscience Corporation (Frederick, MD). PCR was performed on ABI Prism 7700 Sequence Detector (Applied Biosystems). For data analysis the ΔΔCt method was used; for each gene fold-changes were calculated as difference in gene expression between tumor formed by IL12rb2⁺ and that formed by IL12rb2⁻ B16 clone. A positive value indicates gene up-regulation and a negative value indicates gene down-regulation.

Hydrodynamic Gene Delivery.

Injection of the plasmid containing GFP and murine IL12p70 or GFP alone was performed as described (47). Five micrograms of DNA were diluted in 1.6 ml of sterilized 0.9% NaCl solution and injected into KO mice through their tail vein 5 s, 6 h, 3 days, and 7 days after plasmid inoculation blood was collected from animals and sera subjected to IL12p70 ELISA. B16 clones were injected into the same mice 2 h after plasmid inoculation.

Statistical Methods.

Quantitative studies of stained sections were performed independently by three pathologists in a blind fashion. Differences in the number of PCNA positive cells, apoptotic cells and tumor microvessels in immunohistological studies were evaluated by Student's t test. Differences in tumor volume were calculated by using Mann–Whitney U test comparing two independent samples, with 99% confidence interval. All statistical tests were two tailed. P < 0.05 was considered statistically significant.

Abbreviations

Th

T helper

PCNA

proliferating cell nuclear antigen

CAM

chorioallantoic membrane

NK

natural killer

KO

knockout

α-sma

α-smooth muscle actin

CTL

cytotoxic T lymphocyte.