Low-dose paclitaxel improves the therapeutic efficacy of recombinant adenovirus encoding CCL21 chemokine against murine cancer.

Secondary lymphoid tissue chemokine (SLC/CCL21), one of the CC chemokines, exerts potent antitumor immunity by co-localizing T cells and dendritic cells at the tumor site and is currently tested against human solid tumors. Here, we investigated whether the combination of recombinant adenovirus encoding murine CCL21 (Ad-mCCL21) with low-dose paclitaxel would improve therapeutic efficacy against murine cancer. Immunocompetent mice bearing B16-F10 melanoma or 4T1 breast carcinoma were treated with either Ad-mCCL21, paclitaxel, or both agents together. Our results showed that Ad-mCCL21 + low-dose paclitaxel more effectively reduced the growth of tumors as compared with either treatment alone and significantly prolonged survival time of the tumor-bearing animals. These antitumor effects of the combined therapy were linked to altered cytokine network at the tumor site, enhanced apoptosis of tumor cells, and decreased formation of new vessels in tumors. Importantly, the combined therapy elicited a strong therapeutic antitumor immunity, which could be partly abrogated by the depletion of CD4(+) or CD8(+) T lymphocytes. Collectively, these preclinical evaluations may provide a combined strategy for antitumor immunity and should be considered for testing in clinical trials.

Induction of an effective antitumor immune response is a complex process requiring coordinate interaction of different subsets of effector cells, including antigen-presenting cells, lymphocytes, and natural killer (NK) cells.(1–3) However, tumor cells often express limited MHC antigens and lack co-stimulatory molecules; they are ineffective antigen-presenting cells.(4) In addition, many tumor cells produce immunosuppressive cytokines that promote escape from immune surveillance.(5) Thus, enlisting the host response to recognize poorly immunogenic tumors, and altering the immunologic intratumoral microenvironment may be beneficial for the treatment of cancer.

Chemokines are a family of structurally related proteins that activate inflammatory responses, regulate leukocyte trafficking, and participate in many other pleiotropic functions, including regulation of tumor growth.(6,7) The CC chemokine CCL21, also known as secondary lymphoid chemokine or C6kine, is mainly and constitutively expressed in high endothelial venules and in T-cell zones of spleen, lymph nodes, and Peyer's patches.(8,9) Previous studies have established that CCL21 exerts potent chemotactic activity on T cells, B cells, and dendritic cells (DCs).(10–12) The capacity to co-localize DCs and lymphocytes makes CCL21 a good therapeutic candidate against cancer. Accordingly, intratumoral treatment with recombinant CCL21 or CCL21-gene-modified DCs would lead to immune-mediated tumor growth inhibition.(13–17) However, for the multifaceted nature of the tumor microenvironment, the antitumoral response of CCL21 is not sufficiently robust, which suggests that further treatment may require supplemental drugs.

Paclitaxel, a prototypic taxane compound, which exerts its cytotoxic effect by inducing G2/M phase arrest and apoptosis in tumor cells, is one of the most active agents used for the clinical treatment of cancer.(18,19) However, dose-dependent toxicity and drug resistance greatly limit the clinical application of paclitaxel. One promising strategy to reduce the toxicity of chemotherapy is frequent protracted use of low-dose conventional cytotoxic drugs on a metronomic schedule.(20,21) Several recent studies from other groups have indicated that low-dose paclitaxel might alter the intratumoral immunologic microenvironment, including cytokine network and immunosuppressive activity of tumor cells, which suggests that low-dose paclitaxel may improve the outcome of immunotherapy with less toxicity.(22,23) Thus, we wondered whether it would improve the therapeutic and primary immunologic effects of the antitumor by combining low-dose paclitaxel with CCL21. The present article showed that this combined therapy would induce superior antitumor activity.

Materials and Methods

Cell lines and animals

Human cervical carcinoma cell line Hela, human embryo kidney cell line 293, murine melanoma cell line B16-F10, and murine breast carcinoma cell line 4T1 were obtained from the ATCC (Manassas, VA, USA). B16-F10 cells were cultured in DMEM, and 4T1 cells were cultured in RPMI-1640, each supplemented with 10% FBS. Female C57BL/6 and BALB/c mice, 6–8 weeks old, were purchased from the Laboratory Animal Center of Sichuan University (Sichuan, China). All studies involving mice were carried out in conformity with institutional guidelines concerning animal use and care.

Construction of recombinant adenoviral vector encoding murine CCL21

The nucleotide sequence encoding the mature region of murine CCL21 cDNA was PCR amplified from pORF5-mExodus2 (InvivoGen, San Diego, CA, USA) and cloned into pENTR11 vector (Invitrogen, Carlsbad, CA, USA). The resulting plasmid was recombined with destination vector according to the manufacturer's instructions. The viral particles (Ad-mCCL21) were produced and amplified in 293 cells, purified by two-step CsCl gradient ultracentrifugation. The virus titer was quantified using a standard TCID50 assay.(24) The recombinant adenovirus containing the LacZ gene (Ad-LacZ) served as a control.

In vitro infection with recombinant adenovirus

Hela cells were infected at an MOI of 100 (108 p.f.u. per 106 cells in 1 mL complete media) with Ad-mCCL21, Ad-LacZ, or no transduction. Forty-eight hours after infection, cells were harvested and analyzed by Western blotting using a rabbit anti-murine mCCL21 monoclonal IgG antibody (Peprotech EC, London, UK). To detect the supernatant mCCL21 concentrations, 4T1 cells were infected at MOI ranging from 0.1 to 100 with Ad-mCCL21 or Ad-LacZ. Supernatants were harvested 48 h after infection and the mCCL21 concentrations were analyzed with an ELISA kit (Boster, Wuhan, China).

Functional assay of virally generated mCCL21

Biological activity of virally generated mCCL21 was evaluated by chemotaxis as previously described.(25) Briefly, 50 μL 4T1 cell culture supernatant was added to the bottom wells of a 96-well Boyden microchamber's chemotaxis plate (Neuroprobe, Gaithersburg, MD, USA). Then 50 μL lymphocyte cells (2 × 106 cells/mL migration medium) were added to the upper compartment of the migration chamber (5-μm pore size, polycarbonate filters, polyvinylpyrrolidone-free polycarbonate membranes). For the blocking experiments, cell culture supernatant was added with 30 μg/mL rabbit anti-mCCL21 antibody (Peprotech EC) or control rabbit IgG (Sigma-Aldrich, St Louis, MO, USA). The chemotaxis plate was incubated at 37°C for 4 h and transferred to 4°C for 10 min. Then the membrane was fixed in 70% methanol and stained with Wright's stain for 5 min. The number of cells that migrated to the lower surface was counted on six randomly selected high power fields. Commercially murine CCL21 (Peprotech EC) with biological activity served as the standard control.

In vivo production of mCCL21

To assess the duration and amount of mCCL21 expression, mice bearing 4T1 breast carcinoma were intratumorally injected with 5 × 108 p.f.u. Ad-mCCL21 or Ad-LacZ, or no virus. On days 0, 2, 4, 6, and 8, mice were euthanized with carbon dioxide, and the tumor tissue were harvested, cut into small pieces, homogenized, and centrifuged at 12 000g for 30 min. The concentrations of mCCL21 in the supernatants were determined by ELISA as described previously.(26) Tumor-derived mCCL21 concentrations were corrected for total protein by Lowry protein assay (Thermo Fisher Scienfic, Rockford, IL, USA), and the results were expressed as pg/mg total protein.

Tumor growth inhibition study

B16-F10 melanoma and 4T1 breast carcinoma models were established in immunocompetent C57BL/6 and BALB/c female mice at 6–8 weeks of age, respectively, as previously described.(26) Mice were randomly divided into five groups when the size of tumors reached approximately 30 mm3 (about 6 days after tumor cell inoculation), and received the corresponding treatment: normal saline (NS, 100 μL); paclitaxel (Pac, 40 μg/mouse, 100 μL); Ad-mCCL21 (5 × 108 p.f.u., 100 μL); Ad-LacZ+Pac (Ad-LacZ, 5 × 108 p.f.u., 100 μL; paclitaxel, 40 μg/mouse, 100 μL); or Ad-mCCL21+Pac (Ad-mCCL21, 5 × 108 p.f.u., 100 μL; paclitaxel, 40 μg/mouse, 100 μL). Normal saline and the viral particles were injected at several points into the tumors on days 6, 11, 16, and 21 after tumor cell inoculation. Paclitaxel was delivered intrapleurally on the same schedule as above. The experiment for the observation of tumor growth and survival advantage included 10 mice per group. Tumor dimensions were measured with calipers every 3 days, and tumor volumes were calculated according to the formula V = π/6 × length (mm) × width (mm) × width (mm), where length is the longest dimension and width is the shortest dimension. For the histological analysis, the mice were killed by cervical dislocation on day 4 after the completion of treatment as described above. Autopsy was carried out to determine the number and diameter of the metastatic nodules. Tumors and tissues were excised and fixed in 10% neutral buffered formalin solution or frozen at −80°C.(25)

Histological analysis

Tissues from each group were embedded in paraffin. Sections 3–5-μm thick were stained with H&E. Expression of CCL21 was determined by immunostaining with an anti-mCCL21 mAb (Peprotech EC). To determine whether the antitumor effects of the combined therapy were involved in the inhibition of angiogenesis, neovascularization in the tumor tissues was determined by immunohistological analysis with an anti-CD31 antibody (Thermo Fisher Scienfic, Rockford, IL, USA). Vessel density was determined by counting the number of microvessels per high-power field in the sections, as described.(4)

Immunofluorescence staining was used to determine the infiltration of immune cells in the tumor tissue. Tumors were snap-frozen and 8-μm sections were prepared in Tissue Tek (Sakura Finetek) for immunofluorescence analysis. Anti-CD4 (FITC conjugate; eBioscience) and anti-CD8 (Cy5PE conjugate; eBioscience) mAbs were used to determine the T lymphocytes, whereas anti-CD11c (Cy5PE conjugate; BD Bioscience) mAbs were used to detect dendritic cells. Fluorescence was visualized, and images were captured with fluorescence microscopy (Leica).

Quantitative assessment of apoptosis

Paraffin sections were prepared as described above. The presence of apoptotic cells within the tumor sections was determined using a TUNEL Kit (DeadEnd Fluorometric TUNEL System; Promega) following the manufacturer's protocol. In tissue sections, apoptotic cells were counted under a microscope (×200) in five randomly selected fields. The apoptotic index was calculated as a ratio of the total tumor cell number in each high-powered field.

Cytokine-specific ELISA

Mice bearing 4T1 breast carcinoma were treated with various therapies stated above, and on day 4 after the completion of treatment, cytokine protein concentrations from tumor nodules were determined by ELISA as described previously.(14) Tumor tissues were harvested, cut into small pieces, homogenized, and centrifuged at 12 000g for 30 min. Cytokines (granulocyte–macrophage colony-stimulating factor [GM-CSF], γ-interferon [IFN-γ], monokine induced by interferon-gamma [MIG], interferon-gamma-induced protein 10 [IP-10], interleukin-12 [IL-12], transforming growth factor-β [TGF-β], IL-10, and vascular endothelial growth factor [VEGF]) were determined by ELISA in homogenized tumors directly. Tumor-derived cytokine concentrations were corrected for total protein by Lowry protein assay (Thermo), and the results were expressed as pg/mg total protein.

γ-Interferon enzyme-linked immunospot assay

To evaluate the immune specificity of the treatments, IFN-γ enzyme-linked immunospot (ELISPOT) assay was carried out to determine the frequency of T lymphocytes producing IFN-γ in response to specific tumors. On day 4 after cessation of treatment, splenic T lymphocytes from mice bearing 4T1 breast carcinoma were purified with Nylon Fiber Column T (L-Type; Wako). T lymphocytes were co-incubated with either 4T1 cells or non-specific syngeneic CT26 cells at a lymphocyte effector-to-stimulator ratio of 5:1 for 24 h. Spots were quantified with the use of computer-assisted video image analysis as described previously.(27)

In vitro cytotoxicity assay

The 51Cr release assay was carried out to evaluate the adaptive cellular immune response as described previously.(26) Briefly, 4T1 breast carcinoma-bearing animals received therapy as described above. On day 4 after the complement of the treatment, splenic T lymphocytes were isolated from single-cell suspensions with Nylon Fiber Column T (L-Type; Wako) as CTL effector cells. The 4T1 breast carcinoma cells were used as target cells by incubation at 37°C with 51Cr for 2 h. Effector and target cells were seeded into the 96-well microtiter plate at different effector/target (E/T) ratios, and incubated for 4 h at 37°C. The supernatant was harvested, and the CTL activity was calculated by the following formula: %cytotoxicity = [(experimental release − spontaneous release)/(maximum release − spontaneous release)] × 100.

In vivo depletion of immune cell subsets

Immune cell subsets were depleted as described previously.(28) Briefly, mice bearing 4T1 breast carcinoma received i.p. injections of 500 μg of either the anti-CD4 (clone GK1.5, rat IgG), anti-CD8 (clone 2.43, rat IgG), anti-NK (clone PK136) mAb, or isotype controls, 1 day before the combined treatment and then twice per week for 3 weeks. These hybridomas were obtained from ATCC. Tumors were excised and weighed on day 7 after the completion of treatment.

Statistical analysis

Data were assayed by anova and an unpaired Student's t-test. Kaplan–Meier curves were established for each group of animals, and survival curves were compared by the log–rank test. Differences between means or ranks, as appropriate, were considered significant when yielding a P < 0.05.

Results

In vitro and in vivo expression and biological activity of murine CCL21

We first cloned the murine CCL21 gene into an adenoviral shuttle plasmid and generated recombinant Ad-mCCL21. The expression of mCCL21-infected Hela cells was confirmed by Western blot analysis (Fig. 1a). Murine 4T1 breast carcinoma cells were infected with Ad-mCCL21 or Ad-LacZ. A dose-dependent elevation of supernatant mCCL21 concentrations was observed in Ad-mCCL21-infected cells, up to 31.9 ± 2.7 ng/mL at a MOI of 100 (Fig. 1b). The in vitro biological activity of mCCL21 was tested using a chemotaxis assay. Commercially available recombinant murine CCL21 from Peprotech was used as the standard control. The results suggested that conditioned medium from Ad-mCCL21-infected 4T1 cells has a superior capacity of chemotaxis compared with purchased mCCL21 in vitro (Fig. 1c).

Fig 1

In vitro and in vivo expression and biological activity of murine CCL21. (a) Western blot analysis of murine CCL21 expression after infection of Hela cells. (b) 4T1 cells were infected with Ad-mCCL21 or Ad-LacZ at varying MOIs, yielding dose-dependent supernatant mCCL21 concentrations up to 31.9 ± 2.7 ng/mL after Ad-mCCL21 infection. (c) Chemotactic assay of conditioned medium from infected 4T1 cells. Supernatants from Ad-mCCL21 infected 4T1 cells (MOI = 100) clearly attracted lymphocytes, with a superior capacity for chemotaxis compared with purchased mCCL21 in vitro. (d) Tumor mCCL21 levels 0–8 days after a single intratumoral injection of Ad-mCCL21. Levels were 4220.1 ± 303.5 pg/mg at 4 days and remained significantly higher than controls at day 8. Bars, ±SD. *P < 0.01, significantly different from the controls.

In vivo expression of mCCL21 was assessed in the 4T1 tumor tissues after intratumoral delivery of Ad-mCCL21, Ad-LacZ, or sterile NS. Peak expression of mCCL21 was observed 4 days after injection, at which time the mCCL21 concentrations were 4220.1 ± 303.5 pg/mg. On day 8, intratumoral levels of mCCL21 remained significantly elevated (1050.7 ± 95.3 pg/mg vs 291.4 ± 29.0 pg/mg in Ad-LacZ-treated animals; P < 0.01) (Fig. 1d).

Tumor growth inhibition

We evaluated the antitumor activity of the combination therapy in the immunocompetent C57BL/6 and BALB/c mice bearing B16-F10 melanoma and 4T1 breast carcinoma, respectively. The results showed that Ad-mCCL21 or low-dose paclitaxel alone inhibited tumor growth to a certain extent. However, combined treatment with Ad-mCCL21 and low-dose paclitaxel had a superior antitumor effect, resulting in more than 66% and 78% inhibition in tumor volume compared with the NS group in B16-F10 and 4T1 (P < 0.01, 4 days after the completion of treatment; Fig. 2a). In the two tumor models, control animals that received NS treatment survived 44.1 and 27.7 days on average, respectively. In contrast, the combined treatment resulted in a significant twofold increase in life span (P < 0.01, log–rank test; Fig. 2b). Additionally, the appearance of lung metastases in the 4T1 model was significantly delayed by treatment with Ad-mCCL21 or Ad-mCCL21+Pac in comparison with NS control (data not shown).

Fig 2

Tumor suppression and survival advantage in mice. Immunocompetent C57BL/6 and BALB/c mice bearing B16-F10 melanoma and 4T1 breast carcinoma were treated with normal saline (NS), paclitaxel (Pac), Ad-mCCL21, Ad-LacZ+Pac, or Ad-mCCL21+Pac. (a) Suppression of s.c. tumor growth in mice. Ad-mCCL21+Pac treatment resulting in significant tumor growth inhibition versus NS controls (P < 0.01), Pac (P < 0.01), Ad-mCCL21 (P < 0.05), or Ad-LacZ+Pac (P < 0.01) from day 20 after initiation of treatment. Points, average tumor volume; bars, ±SD. (b) A significant increase was observed in the survival rates of Ad-mCCL21+Pac treatment mice compared with the other groups (P < 0.01, by log–rank test).

Inhibition of tumor angiogenesis

Angiogenesis within tumor tissues was estimated in sections by staining with an antibody reactive to CD31. The most highly vascularized area of each tumor was identified on low power and five high-powered fields were counted in this area of greatest vessel density. The combination of Ad-mCCL21 and paclitaxel apparently reduced the number of vessels compared with the control groups, including NS, paclitaxel, or Ad-null+Pac. The reduced vessel density could also be observed in the Ad-mCCL21-treated group. No statistically significant difference was found between paclitaxel and Ad-null+Pac groups (Fig. 3a).

Fig 3

Histochemical analysis of tumors. (a) Tumor angiogenesis was determined by staining paraffin-embedded sections with anti-CD31 antibody. The number of vessels per ×200 field was counted. Occasionally, isolated microvessels could be found in Ad-mCCL21 + low-dose paclitaxel (Ad-mCCL21+Pac)-treated tumors. Ad-mCCL21, Pac, or Ad-LacZ treatment also has an angiostatic effect. At the same magnification, the well-formed capillaries surrounding nests of tumor cells could be found in the normal saline (NS) control group. (b) Apoptotic tumor cells within tumor tissues were elevated by TUNEL assays. The apoptotic index was calculated as a ratio of the apoptotic cell number to the total cell number in each field. Sequential analysis showed that Ad-mCCL21+Pac treatment resulted in significant increment of apoptotic index versus NS control. Pac, Ad-mCCL21, and Ad-LacZ+Pac also efficiently induced the apoptosis of tumor cells. The columns in all graphs correspond to the labeled columns in the picture. Bars, ±SD. *P < 0.01 relative to NS control. ▲P < 0.01 relative to Pac, Ad-mCCL21, or Ad-LacZ+Pac group. MVD, microvessel density, #P < 0.05 relative to NS control.

Therapeutic effect on apoptosis

To explore the role of combined therapy on apoptosis of tumor cells, tumors resected 4 days after the completion of treatment were subjected to TUNEL assays. Paclitaxel, Ad-mCCL21, or Ad-null+Pac treatment affected the apoptotic rate of tumor cells, whereas an apparent increase in the number of apoptotic cells was observed within the tumors of Ad-mCCL21+Pac treated group (Fig. 3b). The apoptotic index (AI) was defined as follows: AI (%) = 100×apoptotic cells/total tumor cells.

Contribution of cytokines in tumor growth delay

To determine the relative contribution of cytokines to the tumor growth delay observed in the combined treated group, we measured the cytokine production from tumor sites following therapy. The tumor tissues were evaluated in the presence of GM-CSF, IFN-γ, MIG, IP-10, IL-12, TGF-β, IL-10, and VEGF by ELISA. Compared with the controls, animals receiving Ad-mCCL21 had a modest but significant increase in type 1 cytokine (IFN-γ and IL-12) and antiangiogenic chemokines (MIG and IP-10) and a decrease in the immunosuppressive mediators (TGF-β) at the tumor sites. However, as was evident for tumor reduction, the combined treatment most effectively altered the cytokine network at the tumor microenvironment. Compared with the NS group, combined treatment significantly reduced the expression of IL-10 (2.7-fold; P < 0.01), TGF-β (2.0-fold; P < 0.01), and VEGF (3.1-fold; P < 0.01). This was coupled with an increase in GM-CSF (5.4-fold; P < 0.01), IFN-γ (4.2-fold; P < 0.01), MIG (2.0-fold; P < 0.01), IP-10 (2.4-fold; P < 0.01), and IL-12 (2.9-fold; P < 0.01) within the tumor sites (Fig. 4).

Fig 4

Analysis of cytokines in the tumor site. Mice bearing 4T1 breast carcinoma were treated with normal saline (NS), paclitaxel (Pac), Ad-mCCL21, Ad-LacZ+Pac, or Ad-mCCL21+Pac. On day 4 after the completion of treatment, tumor tissues were harvested, cut into small pieces, homogenized, and centrifuged at 12 000g for 30 min. Tumors were evaluated for the presence of granulocyte–macrophage colony-stimulating factor (GM-CSF), γ-interferon (IFN-γ), monokine induced by interferon-gamma (MIG), interferon-gamma-induced protein 10 (IP-10), interleukin-12 (IL-12), transforming growth factor-β (TGF-β), IL-10, and vascular endothelial growth factor (VEGF) by ELISA. Cytokine concentrations were corrected for total protein by Lowry protein assay. Results are expressed as pg/mg total protein. Compared with tumor tissues from the NS control group, mice treated with Ad-mCCL21+Pac had significant increase in GM-CSF, IFN-γ, MIG, IP-10, and IL-12 (a) but a decrease in TGF-β (b), IL-10 (c), and VEGF (d). Bars, ±SD. *P < 0.01 relative to NS control. #P < 0.05 relative to NS control. ▲P < 0.01 relative to Pac, Ad-mCCL21, or Ad-LacZ+Pac group. ▵P < 0.01 relative to Pac or Ad-LacZ+Pac group.

Expression of mCCL21 within tumor results in recruitment immune cell subsets

Immunohistochemical staining was done to detect intratumoral expression of mCCL21. The results indicated that mCCL21 can be effectively expressed and secreted from tumor cells in the tumor tissues. No significantly different distribution of CCL21 was found in the tissues of Ad-mCCL21- and Ad-mCCL21+Pac-treated groups in the present study. In order to test whether the combined therapy elevated the infiltration of immune cells, tumors resected 4 days after the completion of treatment were subjected to immunofluorescence assay. The infiltration of lymphocytes and dendritic cells was apparently found in the tumor tissues of both Ad-mCCL21 alone and Ad-mCCL21 in conjunction with paclitaxel groups. Paclitaxel or Ad-null+Pac treatment also affected the infiltration of immune cells (Fig. 5).

Fig 5

Expression of at the tumor site attracts the infiltration of immune cells. (a) mCCL21 immunohistochemical peroxidase staining was done on paraffin-embedded tumor sections obtained from each of the five therapy groups: normal saline (NS), paclitaxel (Pac), Ad-mCCL21, Ad-LacZ+Pac, or Ad-mCCL21+Pac. Significant expression of mCCL21 could be detected in both Ad-mCCL21 and Ad-mCCL21+Pac groups. Representative images of immunofluorescence stained with anti-CD4 (b), anti-CD8 (c), and anti-CD11c (d) antibodies. Significant infiltration of CD4-, CD8-, and CD11c-positive cells in the tumor tissues indicate that Ad-mCCL21 or Ad-mCCL21+Pac treatment would lead to the recruitment and maturation of immune cells. The infiltration of different immune cell subsets was microscopically counted at five randomly chosen high-power fields (×200). The columns in all graphs correspond to the labeled columns in the picture. Bars, ±SD. *P < 0.01 relative to NS control. ▲P < 0.01 relative to Pac, Ad-mCCL21, or Ad-LacZ+Pac group. ▵P < 0.01 relative to Pac or Ad-LacZ+Pac group.

Combined therapy induces specific T-cell responses

To determine whether the therapy induces specific T-cell responses, IFN-γ ELISPOT assays were carried out. The combination of mCCL21 and paclitaxel had significantly greater frequency of specific T cells releasing IFN-γ when restimulated with 4T1 cells (P < 0.01 relative to NS control). AdmCCL21, paclitaxel, or Ad-LacZ+Pac treatment also induced significant specific T-cell responses to 4T1 cells as compared to the NS group (P < 0.01). The minimal responses to control CT26 cells could be observed in the AdmCCL21, paclitaxel, Ad-LacZ+Pac, and Ad-mCCL21+Pac treatment group (Fig. 6a).

Fig 6

Specific antitumor responses were enhanced following treatment with Ad-mCCL21 + low-dose paclitaxel (Ad-mCCL21+Pac) therapy. (a) Enzyme-linked immunospot assay. Mouse γ-interferon-specific assay was done, and spots were quantified with an Immunospot Image Analyzer. Compared with treatment with normal saline (NS), paclitaxel (Pac), Ad-mCCL21, or Ad-LacZ+Pac, T lymphocytes from the Ad-mCCL21+Pac-treated group had significantly greater frequency of specific T cells releasing γ-interferon when restimulated with 4T1 cells. There were minimal responses to the control CT26 cells. Bars, ±SD. *P < 0.01 relative to NS control. (b) CTL-mediated cytotoxicity in vitro. The cytotoxic activity of splenic lymphocytes was measured in 4-h 51Cr-release assay. The purified T lymphocytes were added to 51Cr-labeled 4T1 cells immediately after isolation from spleens. T cells derived from the Ad-mCCL21+Pac-treated mice showed higher cytotoxicity against 4T1 cells than those from the Ad-mCCL21 (P < 0.05), Ad-LacZ+Pac (P < 0.01), Pac (P < 0.01), or NS group (P < 0.01). (c) Abrogation of CTL-mediated cytotoxicity in vivo. Depletion of CD4+ or CD8+ T lymphocytes or natural killer cells by corresponding mAbs. Depletion of CD4+ or CD8+ T lymphocytes significantly impaired the antitumor activity of Ad-mCCL21+Pac therapy, the depletion of natural killer cells showed partly abrogation. Bars, ±SD. *P < 0.01 relative to Ad-mCCL21+Pac therapy. #P < 0.05 relative to Ad-mCCL21+Pac therapy.

Effect of combination therapy on cytotoxic T-lymphocyte activities

Determination of CTL activity provided evidence for the involvement of an immunological mechanism in the antitumor effects of the treatments. As shown in Figure 6(b), splenic T lymphocytes isolated from mice (4T1 model) treated with Ad-mCCL21 or Ad-mCCL21+Pac were able to lyse specific target cells (4T1 cells) in an E/T-dependent manner with statistical significance at E/T ratios ≥10:1. These results suggested that Ad-mCCL21 alone or Ad-mCCL21 plus paclitaxel could efficiently induce tumor-specific CTL response in vivo.

Function of T-cell subsets in antitumor activity

To explore the roles of immune cell subsets in this antitumor activity elicited by the combined therapy, we depleted CD4+ or CD8+ T lymphocytes or NK cells through injection of the corresponding antibodies. As shown in Figure 6(c), depletion of CD4+ or CD8+ T lymphocytes could significantly decrease the antitumor activity of the combined therapy, whereas the depletion of NK cells slightly abrogate the antitumor activity. These findings indicate both CD4+ and CD8+ T lymphocytes were involved in this therapy-induced antitumor activity.

Observation of potential toxicity

To evaluate the possible adverse effects of the combined treatment, Ad-mCCL21+Pac-treated animals without tumor burden were particularly investigated for potential toxicity for more than 4 months. No adverse consequences were indicated in gross measures, such as weight loss, ruffling of fur, behavior, feeding, or toxic death. No pathologic changes in liver, heart, lung, kidney, spleen, brain, or bone marrow were found by microscopic examination (data not shown). Our results suggested that low-dose paclitaxel in combination with Ad-mCCL21 was concurrently practicable with less host toxicity.

Discussion

In these studies we tested the hypothesis that the combination of low-dose paclitaxel with recombinant adenovirus expressing mCCL21 would improve therapeutic antitumor efficacy. This proved to be the case, as we found that the combined therapy limits the growth and progression of murine cancer more effectively, thereby improving tumor-related cachexia and prolonging survival. Our results indicate that the effects of the combination therapy are associated with increased infiltration of immune cells, altered the immunosuppressive microenvironment, improved tumor cell apoptosis, and reduced new vessel formation in the tumor tissues. We further show that the antitumor activity of the combined treatment could be partly abrogated by the depletion of CD4+ or CD8+ T lymphocytes, which suggested these antitumor immune responses are dependent on both CD4+ helper and CD8+ cytotoxic T cells. Importantly, the enhancement of antitumor activity occurred without an ensuing increase in host toxicity or signs of acquired drug resistance during the course of treatment.

Tumor cells, together with tumor-infiltrating stromal cells, produce a variety of growth factors and cytokines with autocrine and paracrine functions, which enables them to progress to a more aggressive phenotype and escape the surveillance of the immune system.(22) Thus, altered cytokine networks in the tumor microenvironment may contribute to the dysregulation of cellular functions in cancer cells. It has been shown that significant change in cytokine profiles can be observed in tumor sites after intratumoral delivery of CCL21.(13–17) Our findings that the cytokine production from tumor tissues (GM-CSF, IFN-γ, MIG, IP-10, IL-12, TGF-β, IL-10, and VEGF) was significantly altered by Ad-mCCL21 or Ad-mCCL21+Pac therapy are in agreement with previous reports. Moreover, the combined treatment synergistically reduced the immunosuppressive molecule IL-10 and angiogenesis factor VEGF, and increased GM-CSF, IFN-γ, MIG, IP-10, and IL-12 within the tumor sites. Thus, the combined therapy may benefit from the synergistic altering cytokines by improving antigen presentation, CTL generation, and angiogenesis limitation.

In the tumor microenvironment, VEGF can be produced by various cell types, including tumor cells, inflammatory and stromal cells, platelets and vascular cells.(29) Overexpressing TGF-β can be observed in many cancer cells (including B16-F10 and 4T1), and the elevation of TGF-β is correlated with tumor progression, invasion, metastases, and poor prognosis.(30,31) It has been reported that the full potency of CCL21-mediated antitumor response required the induction of IFN-γ, MIG, and IP10 in concert,(32) which is in agreement with our findings (Fig. 4a). Both MIG and IP10 are potent inhibitors of angiogenesis that can limit new vessel growth by binding to CXCR3 on endothelial cells and triggering endothelial cell anoikis, thus downregulating the expression of VEGF by endothelial cells.(33,34) Moreover, reducing microvessel density would lead to the observed increase in both tumor cell apoptosis and necrosis, and block the tumor cells to express VEGF and TGF-β. Transforming growth factor-β is one of the immune inhibitory cytokines that may potently suppress Ag presentation and antagonize CTL generation and macrophage activities, thus enabling the tumor to escape immune detection.(13) The decrease in TGF-β at the tumor site after CCL21 treatment may have contributed to an increase in antitumor immunity. Thus, we may speculate that the decreasing expression of TGF-β at the tumor site is not only a result but also a reason of the decrease of tumor volume. Further studies will be necessary to delineate the exact mechanism by which CCL21 suppresses the expression of cytokines.

Angiogenesis, a key regulatory factor in tumor growth and metastasis, is tightly regulated by many positive and negative factors, and antiangiogenic therapy has become a potential antitumor strategy.(25,35–37) Here we showed that combined treatment with Ad-mCCL21 and paclitaxel led to enhanced antiangiogenic effectiveness as compared with either treatment alone. The effectiveness of this approach is likely attributable to the cascade of events that are initiated within the tumor microenvironment by delivery of Ad-mCCL21 and low-dose paclitaxel. At the outset, although CCL21 has been identified as a ligand for CCR7, recent studies have shown that mCCL21 can signal through mCXCR3 and share some of the activities of MIG and IP10, including their antiangiogenic activity.(15,38,39) In this regard, Ad-mCCL21 treatment may directly inhibit the formation of neoangiogenesis in the tumor bed. In addition, a significant increase in the expression of angiostatic factors MIG and IP10 can be observed in the tumor tissues of the Ad-mCCL21+Pac-treated group. Hence, the antiangiogenic activities of the combined therapy may in part due to the increase in antiangiogenic proteins MIG and IP10. Furthermore, growing evidence suggests that frequent administration of low doses of chemotherapeutic agents could inhibit angiogenesis by targeting tumor-associated endothelial cells.(40–44) Thus, the combined therapy could maximize the growth-limiting effects on the tumor vasculature. The enhanced antiangiogenic activities of the combined therapy would lead to observed increase in both tumor cell apoptosis and necrosis. This may, in turn, improve the presentation of tumor antigens from the dying tumor cells to T lymphocytes and thus support antitumor immunity.

The breaking of immune tolerance against tumor-associated antigens and induction of autoimmunity against tumors should be a useful approach for the treatment of tumors.(45–47) Although it is common to consider conventional cytotoxic chemotherapy as an immunosuppressive modality, more recent studies have indicated that some cytotoxic drugs have the ability to disrupt pathways of immune suppression and immune tolerance in a manner that depends on the drug type, its dose, and the treatment schedule in relation to the immune-based intervention.(48) Clinical evidence also suggests that standard dose chemotherapy can affect the human immune system in both negative and positive ways.(49) In this context, the treatment time of standard dose chemotherapy drugs should be carefully dissected for integrating with immune-based therapies, to achieve the greatest long-term clinical benefit for cancer patients. Alternatively, low dose chemotherapy may more efficiently augment antitumor immunity by a variety of mechanisms.(50,51) It has been shown that chemotherapeutic agents in non-cytotoxic concentrations are capable of triggering the release of danger signals from dying tumor cells, increasing the immunogenicity of tumor cells, upregulating the ability of DCs to present Ags to Ag-specific T cells, and altering the immune tolerance of the tumor environment.(22,23,48,52,53) The data in the present study showed that combined treatment with Ad-mCCL21 and paclitaxel led to enhanced antitumor growth and the increased infiltration of CD4+, CD8+, and CD11c+ cells in the tumor tissues compared with the Pac, Ad-mCCL21, Ad-LacZ+Pac, or no-treatment groups. In addition, the depletion of CD8+ or CD4+ lymphocytes showed partial abrogation of the antitumor activity of the combination therapy in vivo, which indicates the antitumor immunity of the combination therapy is dependent on both CD4+ helper and CD8+ cytotoxic T cells.

In conclusion, we showed that the combination of CCL21 with low-dose paclitaxel significantly enhanced the antitumor activity compared with either treatment alone. These enhanced antitumor efficacies functioned by altering the intratumoral microenvironment, limiting tumor angiogenesis, inducing tumor cell apoptosis, and infiltration of lymphocytes. The present findings may prove useful in further explorations of the potential application of this combined approach in the treatment of malignant cancer.