This article is mentioned in:

Introduction

The global prevalence of bladder cancer is estimated at more than 1 million and is steadily increasing. At diagnosis, 20–25% of cases are non-muscle invasive bladder cancer (NMIBC), and there is a high rate of tumor recurrence and progression even after local surgical therapy. Thus, numerous patients require lifelong follow-up examinations that include additional prophylactic treatments in the event of recurrence. Intravesical Bacillus Calmette-Guérin (BCG), a live attenuated Mycobacterium bovis vaccine widely used to induce immunity against tuberculosis, is currently the most common therapy employed for NMIBC, best known as the most effective agent for the treatment of high-grade superficial bladder cancer. As an adjunct to transurethral resection, BCG is the treatment of choice for urothelial carcinoma in situ (CIS) and for recurrent or multi-focal Ta and high grade T1 bladder lesions (1). In patients for whom radical cystectomy is not performed, BCG is currently the first treatment option for high-risk NMIBC and CIS. BCG therapy achieves 50–60% effectiveness against small residual tumors, and a 70–75% complete response rate for CIS (2). Recent trials indicate that immunotherapy with BCG is superior to chemotherapy in patients with intermediate- to high-risk for recurrence (3).

Despite therapeutic application since 1976, the mechanism of BCG action against bladder cancer remains unknown, yet it is assumed that BCG-induced antitumor activity is dominated by local non-specific immunological reactions reflecting the activity of immunocompetent T cells (4). BCG accumulates near and adheres to the bladder wall by binding to fibronectin (5). After passage through the GAG layer, BCG is internalized and processed by professional APC and tumor cells, which then locally activate numerous lymphocytes, macrophages, pleomorphic mononuclear and NK cells (6,7). Th1-polarized cell-mediated immunity appears to play a key role during BCG immunotherapy and IFN-γ-producing, NK and CD8+ in addition to CD4+ T cells are major cellular mediators of this antitumor action (3).

However, a high percentage of patients fail initial BCG therapy, and 40–50% of BCG responders develop recurrent tumors within the first 5 years (8). In addition, BCG, a viable living organism, causes infections. Unfortunately, up to 90% of patients experience side-effects ranging from bothersome cystitis in the majority of patients to life-threatening complications such as sepsis in rare cases (9). Where BCG is ineffective, treatment schedules consisting of viable BCG and human lL-2 or other Th1 cytokines are proven to be effective (10). Induction of Th1 immunity is required for successful BCG immunotherapy of bladder cancer. The Th1 cytokine hIFN-α has been found to be safe and tolerable when administered intravesically, alone or in combination with BCG, in numerous controlled studies (11,12). Additionally administration of IFN-α2b, both alone and in combination with BCG, has been reported to achieve improved clinical efficacy (13). The side-effect profile of combination therapies is similar to BCG monotherapy, and one phase III study has recommended combination therapy with BCG and IFN-α in BCG non-responders or relapsers (14). Although BCG and IFN-α combination therapy may benefit patients with high-risk disease or carcinoma in situ, the efficacy of cytokine perfusion is limited by the high cost, short half-life and water solubility of cytokines, which are readily lost to urine (15).

BCG has also been used as a live vehicle to deliver multiple pathogen antigens due to the high immunogenicity and low toxicity of this organism (16). We thus sought to assess the efficacy of administration of a previously described genetically engineered recombinant hIFN-α2b-secreting BCG (rBCG) (17) in a murine model of bladder cancer.

Materials and methods

BCG strains and culture

The Mycobacterium bovis (M. bovis) BCG Danish 2 strain was purchased from the Shanghai Institute of Biological Products. Recombinant hIFN-α2b-BCG was constructed in-house. The hIFN-α2b fragment was directionally cloned into the shuttle plasmid pMAO-4 to form a recombinant plasmid phIFN-α-2B, which was extracted prior to enrichment in DH5α-E. coli and electrically transduced into BCG as previously described (17) (Fig. 1). BCG was recovered on 7H10 solid media, and cultured in Middlebrock 7H9 broth media (both from Difco Laboratories, USA) supplemented with 0.05% (by vol) glycerol, 10% (by vol) albumin-dextrose-catalase (ADC) and 0.05% (by vol) Tween-80, at 37°C and 150 rpm until the log-stationary phase. rBCG was supplemented with 15 μg/ml kanamycin (Sigma-Aldrich, St. Louis, MO, USA), and then washed and dissolved in phosphate-buffered saline (PBS). When A60 = 1, BCG/rBCG in the present study was ~1.4×108 CFU/ml. Exogenous recombinant IFN-α2b (Sigma-Aldrich) was applied by the same amount of rBCG in the 7-day culture supernatant: ~1,000 IU in ~3×108 CFU/ml.

Figure 1 Construction strategy of recombinant human interferon-α-2b Danish 2 BCG. (A) Plasmid profile of shuttle expression vector, (B) phIFN-α-2B. The recombinant plasmid phIFN-α-2B was produced from the shuttle plasmid pMAO-4 by cloning in the hIFN-α2b, enriched in DH5α-E. coli prior to extraction, and then electrically transduced into BCG. Four weeks after electrotransformation, kanamycin selection produced a single colony confirmed as rBCG by sequencing and positive acid-fast staining of classic Mycobacterium (C, oil immersion magnification, x1,000). BCG, Bacillus Calmette-Guérin; rBCG, recombinant hIFN-α2b-secreting BCG.

Bladder cancer cell line and culture

The mouse bladder tumor cell line MB49, originating from the C57BL/6 mouse, was provided by Luo Yi of the Department of Urology, University of Iowa, USA. Cells were grown in RPMI-1640 medium (Gibco, USA) supplemented with 10% fetal bovine serum, at 37°C in 5% CO2 in a humidified incubator. Cells in the logarithmic growth phase were seeded at 1×106/ml, and at 90% confluency were co-cultured with BCG or rBCG at an MOI of 1:1 at 37°C in 5% CO2 and saturated humidity, for 24, 48 and 72 h. PBS was used in place of BGC as a control.

Light and electron microscopy

MB49 cells were observed under an inverted microscope and by transmission electron microscopy, as previously described (18). Ultrathin sections were stained with 0.5% uranyl acetate and 0.04% lead citrate and observed under a transmission electron microscope JEM-2000EX (Jeol, Ltd., Tokyo, Japan) at 80 kV.

CCK-8 tumor cell viability assay

Cellular viability was assessed using a colorimetric cell counting kit, the Cell Counting Kit-8 (CCK-8; Dojindo, Japan), according to the manufacturer’s instructions. MB49 cells (1×105 cells/well) we r e seeded in 96-well plates. At 90% confluency, rBCG, BCG, BCG+hIFN-α2b or PBS were added in triplicate to each well at an MOI of 10:1, and BCG (rBCG, BCG+hIFN-α2b) alone in medium was used as the blank control. CCK-8 solution (10 μl) was added to each well, and after 4 h the absorbance was read on a microplate reader at 450 nm. Cell growth inhibition rate (%) = (1-Abstest/AbsPBS) x 100%.

Apoptosis assay

Cells were seeded on coverslips in a 6-well plate, cultured in 3 ml RPMI-1640. When cells adhered, BCG or rBCG was added at an MOI of 1:1 and co-cultured for 24, 48 and 72 h. After 4% paraformaldehyde fixation, 10 μl of acridine orange (AO) (AppliChem, Gatersleben, Germany) dye mix (100 mg/l) was added to each well, and after 5 min, the slides were rinsed twice in PBS with 1% hydrochloric acid for 5 sec, then decolorized in CaCl2 for 3–5 min. Staining with Hoechst 33258 (Sigma-Aldrich) was carried out according to the manufacturer’s instructions, and the cells were observed under a fluorescence microscope (Nikon 80i; Nikon, Japan) (excitation 340 nm for Hoechst 33258, 502 nm for AO).

Cells grown in 6-well plates (2×105 cells/well) were incubated in the presence or absence of inducers, and then harvested by centrifugation at 4°C and 1,000 x g for 5 min after trypsinization, and then rinsed twice with PBS. Cells were suspended in 100 μl binding buffer, and then stained in triplicate with the fluorescein isothiocyanate (FITC) Annexin V apoptosis detection kit (BD Biosciences, Bedford, MA, USA) at room temperature for 15 min. Thereafter, the flow cytometric analysis of cells was performed with BD FACSVantage SE cytometer (BD Biosciences) within 1 h.

FCM analysis of MHC-I expression

Cells were incubated with 5 μl FITC-anti-mouse MHC-I (H-2Kb) Ab (BD Biosciences) for 20 min. After washing with buffer [PBS containing 10 mmol/l ethylenediaminetetraacetic acid (EDTA) and 0.1% sodium azide], the cells were analyzed using the BD FACSVantage SE cytometer.

Murine orthotopic bladder cancer model and BCG treatment

We used a well-defined murine syngeneic orthotopic MB49 bladder cancer model to evaluate the role of rBCG in vivo (19). In brief, 6- to 8-week-old SPF female C57BL/6 mice (Animal Laboratory Center, Beijing Medical University, China) upon approval by the Animal Use and Care Committee of Tianjin Medical University, were divided into 4 groups (n=15) and one group of controls (n=6). The 4 groups were catheterized to receive an intravesical inoculate of 105 MB49 bladder tumor cells on day 0 following a 5-sec treatment of the bladder wall with 30 μl of 0.2 M silver nitrite. On days 1, 8, 15 and 22 following tumor implantation, the mice were treated intravesically with 100 μl of either 3×108 CFU/ml BCG, 3×108 CFU/ml rBCG, 3×108 CFU/ml BCG+hIFN-α2b or PBS. Exogenous hIFN-α2b was administered at the concentration achieved on day 7 of culture for the same amount of rBCG.

Survival was recorded daily for 6 weeks. Bladders were weighed to obtain the tumor growth inhibition rate (%) = (1 - weighttreated group/weightcontrol group) x 100%. Twelve hours after the last perfusion, blood was taken from 6 mice in each group through the inner canthus venous plexus.

Hematoxylin and eosin (H&E) and auramine O staining

The bladder, liver, spleen, kidney, lung and heart of mice sacrificed on day 15 following the last perfusion and those that died during the observation period were immediately removed. Tissues were fixed in 4% paraformaldehyde, paraffin-embedded, and then H&E stained for pathological observation under a light microscope. For observation of mycobacteria in tissue, auramine O was applied for staining as follows. The paraffin section was dewaxed to water, dyed by auramine O solution for 5 min, decolorized with 3% hydrochloric acid alcohol for 1–2 min, then redyed 1–2 min in hematoxylin.

Immunohistochemistry

Mouse bladders were resected, shock frozen in liquid nitrogen and stored at -80°C. Immuno-histochemical staining of 5-Am cryostat sections with peroxidase-conjugated secondary antibodies was carried out as previously described (20) with mouse CD3, CD20 and Gr1; and mouse Fas (both from BD Biosciences, Heidelberg, Germany). The number of infiltrating immune cells, expressed as average number/high-power field (HP), was determined by examining 10 randomly selected non-overlapping microscopic fields at x400 magnification. To semi-quantify the cellular Fas expression, each slide was scored on the basis of percentage and intensity of stained tumor cells as previously described (21), as follows: no staining, 0; <20%, 1; 20–75%, 2; and >75%, 3. The intensity of staining was graded on the following scale: negative, 0; low, 1; moderate, 2; and strong, 3. The product of the scores for the intensity and positive rate of staining was the total score. In the present study, a final total score >2 for Fas expression was defined as high-expression. All cell counting results were verified between the manual and automated methods and are expressed as the viable cell numbers for interpretation. Scoring was carried out in a blinded manner by two independent individuals.

FCM analysis of T lymphocyte populations in peripheral blood

Whole blood samples were treated with EDTA, according to the manufacturer’s instructions. Whole blood (100 μl) was incubated with directly conjugated fluorescent CD3/CD4/CD8 antibodies for 30 min in the dark at room temperature, and then red cells were lysed using FACS Lyse (both from BD Biosciences). Stained cells were washed and fixed in PBS with 1% formaldehyde, and samples were acquired on FACSVantage. Fluorescence minus one gating techniques were employed to establish staining thresholds and aid gating of T cell subpopulations.

Detection of cytokine secretion by ELISA

Heparinized whole blood supernatants (50 μl) were harvested, and mTNF-α and mIL-12 levels were assayed by ELISA (mouse TNF-α and IL-12 ELISA kit; Biosource, France) according to the manufacturer’s instructions. The signal was detected using the BioTek Epoch microarray reader, and the results were analyzed using Gen5 software.

Statistical analyses

SPSS 18.0 software was used for data analysis. Results displayed represent the mean ± SD. One-way ANOVA followed by the Newman-Keuls post-hoc and Fisher exact test probability were used to compare values and rates between groups, respectively. ANOVA was employed for pairwise comparison of repeated measurements. Survival of the mice was evaluated using Kaplan-Meier plots and the logrank test. Statistical analyses are given as two-sided p-values. p<0.05 was considered to indicate a statistically significant result, and p<0.01 indicated extreme significance.

Results

rBCG affects bladder cancer cell morphology

Untreated MB49 cells were mostly round or of irregular shape, began to adhere to the well wall after 1–2 h culture, and extended cytoplasmic protrusions were observed after 12 h. MB49 cells demonstrated a high nuclear/cytoplasmic ratio, uneven chromatin and the mitotic phase was occasionally observed (Fig. 2A and D).

Figure 2 MB49 cell morphology: light microscope (scale bar, 50 μm, A-C; magnification, x200) and TEM (D-F). (A) Intact cells in the control group were generally bulky and wall-attached. (B) Cells treated with rBCG for 48 h shrunk, clumped and detached from the wall. (C) After 72 h, cellular morphology was abnormal and cell death was observed. (D) Intact cells contained ribosomes, mitochondria and endoplasmic reticulum in plasma, and protuberances and microvilli were present on the cell surface (scale bar, 2 μm; TEM magnification, x4,000). (E) After 48 h, the cell size was reduced, cytoplasm was condensed and bulges rather than microvilli were visible on the cell surface (scale bar, 1 μm; TEM magnification, x5,000). (F) After 72 h, irregular nuclei and large vacuoles were observed in the condensed cytoplasm (scale bar, 1 μm; TEM magnification, x5,000). TEM, transmission electron microscopy.

During co-culture with BCG, rBCG or BCG+hIFN-α2b cellular refraction decreased, cytoplasmic protrusions gradually disappeared, soma became stubby and the cytoplasm became granular (Fig. 2B and E). After 48 h, the cells had clearly shrunk, protuberances disappeared, cells rounded up and proliferation slowed (Fig. 2B and E). Some cells grew in suspension and formed clumps surrounded by bacteria. After 72 h, the cells were granular, sparsely distributed, and some exhibited vacuolization, blebbing and lytic necrosis (Fig. 2C and F). The less damaged cells showed mitochondrial swelling, large vacuolar degeneration, dissolved nuclear chromatin, cytoplasm necrosis and surface microvilli decrement (Fig. 2F).

rBCG inhibits tumor cell growth and induces apoptosis

MB49 cells incubated with rBCG grew less quickly than the controls incubated with PBS (Fig. 3). After 72 h, rBCG achieved a significantly higher inhibitory ratio (86.37±3.67%) than BGC (53.43±1.84%) or BCG+hIFN-α2b (56.03±2.79%) (all p<0.05) (Fig. 3).

Figure 3 Growth inhibitory ratio of MB49 cells incubated with BCG, rBCG or BCG+IFN. MB49 cells were seeded at an initial density of 1×106 cells/ml, and incubated for 24, 48 and 72 h with an MOI of 10:1 BCG and/or BCG+IFN. The inhibitory ratio was determined by CCK-8. *p<0.05 vs. rBCG at 72 h.

Intact MB49 cells stained with AO adhered and grew vigorously. The nuclei emitted yellow or green fluorescence, and the cytoplasm emitted orange red fluorescence, indicating a high karyoplasmic ratio and cells were spindle shaped (Fig. 4A). MB49 cells incubated for 24 h with rBCG, BCG or BCG+hIFN-α2b were not obviously affected, yet after 48 h, the cells clumped, cytoplasmic protuberances gradually disappeared, the karyoplasmic ratio decreased, and vesicular bulging membranes and apoptotic bodies were observed (Fig. 4B). After 72 h, the cells no longer adhered firmly, and the cell number and karyoplasmic ratio were reduced (Fig. 4C and D). Hoechst 33258 staining revealed apoptotic bodies at 48 and 72 h (Fig. 4E–H).

Figure 4 Fluorescence imaging of bladder tumor cells stained with (A-D) AO and (E-H) Hoechst 33258. (A) Intact MB49 cells were spindle-shaped and nuclei fluoresced yellow/green and the cytoplasm emitted orange/red fluorescence, indicating a high karyoplasmic ratio (FSM, x100). (B) After 48 h in the presence of rBCG, the cells became rounded and clumped, cytoplasmic protuberances disappeared, the karyoplasmic ratio increased and vesicular bulging membrane and apoptotic bodies were observed (FSM, x200), (C) After 72 h, significantly fewer cells were observed (FSM, x100). (D) Apoptotic cells forming typical petal-like structures (FSM, x400). (E) Intact cells showed faint blue luminescence under UV (FSM, x200). (F) Luminous apoptotic bodies were observed (FSM, x200). (G and H) Cells incubated with rBCG for 48 and 72 h (FSM, x40).

The rate of apoptosis was 2.31±1.02% in the control cells, but after 24 h the presence of BGC, rBCG and BCG+hIFN-α2b significantly increased the rate of apoptosis to 7.9±0.97% (p<0.05), 19.92±0.77% (p<0.01) and 20.11±0.74% (p<0.01), respectively (Fig. 5).

Figure 5 MB49 cell apoptosis. (A) Representative images of Annexin V/PI staining of MB49 cells incubated alone or with BCG, rBCG or rBCG+IFN. The upper right quadrant (PI+/Annexin V+) displays dead and late apoptotic cells, the left lower quadrant displays living cells (PI−/Annexin V−), the lower right quadrant (PI−/Annexin V+) displays early apoptotic cells. (B) MB49 cell apoptosis and death were detected by confocal laser scanning microscopy. Dead cells were stained internally with PI and apoptotic cell membranes were stained with Annexin V (magnification, x400). (C) Flow cytometry showing MB49 cell apoptosis in the control, BCG, rBCG and BCG+IFN groups. Data are shown as the mean ± SD. n=4; *p<0.05; **p<0.01 vs. control group.

rBCG promotes MHC-I expression on MB49 cells

MB49 cells were incubated with MHC-I-directed antibody and staining was detected by flow cytometry. A 24-h incubation with BCG did not significantly influence the fraction of MHC-I+ cells (2.89±0.24%). However, rBCG and BCG+hIFN-α2b increased the fraction of MHC-I+ cells to 17.18±0.88% (p<0.05) and 25.13±1.42% (p<0.01), respectively (Fig. 6).

Figure 6 MHC-I expression on tumor cells detected by FCM. (A) Representative images of MHC-I expression in control cells, and cells incubated with BCG, rBCG or rBCG+hIFN-α2b for 24 h. (B) Comparison of MHC-I expression in control cells, and cells incubated with BCG, rBCG or rBCG+IFN for 24 h (mean ± SD). n=3; *p<0.05; **p<0.01 vs. control group. FCM, flow cytometry.

rBCG inhibits tumor growth and promotes survival in a mouse model of orthotopic bladder cancer

After 2 weeks of modeling, C57BL/6 mice exhibited different degrees of hema-turia. After 6 weeks, several mice developed a hypogastric mass of 0.5–2 cm upon palpation. Mice administered rBCG (n=5) survived for significantly longer than mice administered BCG (n=7) (p<0.001), yet did not survive significantly longer than mice administered BCG+hIFN-α2b (n=8) (Fig. 7A).

Figure 7 Tumor growth and survival in the C57BL/6 mice and microscopic morphology of the organ tissues. (A) Survival of mice following tumor establishment. Black triangles indicate that the survival of mice administered rBCG was higher than mice administered BCG (p<0.05), and did not differ significantly from the mice administered BCG+IFN. (B) Average bladder weight (mean ± SD, n=4). *p<0.001 vs. the TP group. TP group, tumor-bearing mice administered PBS. (C) Liver, (D) lung, (E) heart, (F) kidney and (G) spleen of the tumor-bearing mice administered rBCG all showed no obvious abnormal, no tumor metastasis and miliary pattern. (H) Individual mouse spleen showed hyperplasia (H, magnification, x100; C-G, magnification, x200).

The average bladder weight was significantly lower in mice administered rBCG (143.6±1.6 mg) than that in mice administered PBS (n=14) (251.5±2.2 mg, p<0.001, Fig. 7B). Average bladder weight was reduced by 39.9% in mice administered BCG, 42.9% in mice administered rBCG and 40.2% in mice administered BCG+hIFN-α2b.

The bladder, liver, spleen, kidney, lung and heart of the mice sacrificed following the last perfusion and those that died during the observation period indicated no tumor metastasis and miliary nodules. No significant difference in pathological features was observed among the mice treated with BCG, rBCG or BCG+hIFN-α2b (Fig. 7C-H).

rBCG induces immunity in the bladder

The pathological morphology of the bladder tissue indicated diffuse infiltration of tumor cells with high-grade malignancy (Fig. 8A and B). Administration of BCG, rBCG and BCG+hIFN-α2b induced bladder inflammation. Migrating inflammatory cells were observed in the submucosa. The local inflammatory reaction in response to BCG was characterized by an initial increase in blood flow, enhanced vascular permeability characterized by edema, and an influx of effector cells. Vasodilation was evident and leukocytes populated the submucosal layer (Fig. 8C). After administration of BCG into the mouse bladder, auramine O staining-positive bacteria were detected within and underlying urothelial cells, indicating that BCG was taken up by the epithelium (Fig. 8D).

Figure 8 H&E staining and immunocyte content of the bladder and cancer tissues in the C57BL/6 mice (A-C, magnification, x200; F, magnification, x100). (A) Untreated bladder exhibited normal morphology. (B) Tumor tissue formed by seeded MB49 cells did not affect normal bladder structure; tumor cells were diffusely infiltrated. (C) After BCG administration, submucosal edema appeared, vasodilation was evident and inflammatory cells infiltrated the submucosal layer. (D) After 24 h of BCG administration, auramine O-positive bacteria were detected within and underlying the urothelial cells. The red frame is the merge of the bright and fluorescent field 2 h after BCG administration, indicating BCG adhesion to the intimal. The white frame displays BCG taken up by the epithelium. (E) IHC image of CD20, CD3 and Gr1 staining in the rBCG-treated bladder (magnification, x400). (F) Comparison of immunocyte content in the tumor tissues. Data are shown as means ± SD (n/10 HP). *p<0.05; ***p<0.001 vs. the TP group. TP group, tumor-bearing mice administered PBS; BI group, BCG+IFN group. H&E, hematoxylin and eosin. PMN, polymorphonuclear leukocytes; M-M, monocytes/macrophages.

Immunohistochemical staining indicated infiltration of CD3+ lymphocytes, CD20+ monocytes and Gr1+ polymorpho-nuclear leukocytes (PMNs) (Fig. 8E and F). PMN, monocyte and T lymphocyte infiltration increased significantly in the treated groups, compared with the infiltration in the control group (all p≤0.05). Monocyte and T cell counts were significantly higher in mice administered rBCG than these counts in the mice administered BCG or BCG+hIFN-α2b (both p=0.000). However, PMN counts did not differ significantly between the treated animals.

rBCG increases the expression of Fas

Fas expression was initially low in the bladder tumor tissues. Administration of BCG, rBCG and BCG+hIFN-α2b significantly increased Fas expression (p=0.000), but the intensity of Fas staining did not differ significantly between the three treated groups (Fig. 9 and Table I).

Figure 9 Fas immunohistochemical staining in the C57BL/6 mouse bladder cancer tissues. Fas was detected only in the cytoplasm of the cells. (A) Fas-negative tissue showed cytoplasm and cell membrane stained blue or light blue. (B and C) Positive staining was indicated by brown uniform fine particles located in the cytoplasm and on the cell membrane. (A) Negative control (isotype-matched control; magnification, x200). (B) Mouse bladder tumors had low Fas expression (magnification, x100). (C) The tumor tissue treated by BCG showed positive expression of Fas (magnification, x100).

Table I Fas expression in the tumors (cases) according to the IHC results.

Table I

Fas expression in the tumors (cases) according to the IHC results.

Groups	Low expression	High expression
TP	14	1
BCGa	4	11
rBCGa	1	13
BIa	4	10

a p<0.001 vs. TP group. C57BL/6 mice were administered BCG, rBCG and rBCG+IFN, as presented in Fig. 7.

rBCG increases the ratio of CD4+/CD8+ in peripheral blood

In comparison to normal mice, the tumor-bearing mice had depressed levels of peripheral blood CD4+ cells and lower CD4+/CD8+ ratios. However, administration of BCG, rBCG or BCG+hIFN-α2b elevated CD4+ cell counts and CD4+/CD8+ ratios to near-normal levels (Fig. 10A-C) (all p<0.01). However, no significant difference in these values was detected between the three treatment groups.

Figure 10 T lymphocyte subsets and cytokines in peripheral blood of C57BL/6 mice administered BCG, rBCG or BCG+IFN as analyzed by FCM. (A) Representative images of CD3+, CD3+CD8+ and CD3+CD8+ lymphocytes in the peripheral blood of mice administered BCG, rBCG or BCG+IFN, and (B) the percentage of CD3+, CD3+CD8+ and CD3+CD8+ lymphocytes in the peripheral blood of mice administered BCG, rBCG or BCG+IFN (mean ± SD) (n=6). *p<0.05 vs. normal. (C) CD4+/CD8+ ratio, mean ± SD (n=6). **p<0.01 vs. normal. (D) The level of mTNF-α and mIL-12 in the peripheral blood of the C57BL/6 mice administered BCG, rBCG or BCG+IFN, mean ± SD (n=6, pg/ml). *p<0.001 vs. normal. N, normal group; TP group, tumor-bearing mice administered PBS; BI group, BCG+IFN group

rBCG increases the level of TNF-a and IL-12

rBCG administration increased circulating levels of mTNF-α and mIL-12 to 02.33±11.00 and 854.46±4.56 pg/ml, respectively (Fig. 10D). However the circulating levels of these cytokines did not differ significantly among mice administered rBCG, BCG or BCG+hIFN-α2b.

Discussion

BCG prevents bladder cancer-related metastasis and decreases bladder cancer-associated mortality. Although systemic reactions have been reported, the likely mechanism of BCG action involves local inflammation (1). We sought to further characterize the mechanism of action of BCG in a bladder cancer cell line and an orthotopic murine bladder cancer model.

Due to the involvement of the host immune system in BCG efficacy, immunodeficient mice are not suitable for investigation of its mechanism of action. We instead established an orthotopic bladder tumor model in which the mouse bladder tumor cell line MB49 was implanted into the C57BL/6 mouse bladder. C57BL/6 mice were chosen, as they are widely used, permitting direct comparison with previously established clinical baselines (22). Following chemical injury and cell transplantation (19), tumor cell proliferation was observed, yet no tumor metastasis and miliary nodules were found in all groups.

A wide range of rBCG vaccine candidates expressing bacterial, viral, parasitic antigens have previously been developed, and rBCG strains secreting mouse and human cytokines, primarily Th1 cytokines (e.g., IL-2, IL-18, IFN-γ and IFN-α), have previously been investigated (10). We studied the influence of hIFN-α2b-rBCG on tumor growth in vitro and both tumor growth and the systemic and local immune response in vivo. We used the BCG Shanghai substrain, derived from the Danish 2 strain that has been used for tuberculosis prevention and immune-modulation universally in China. We engineered a strain of BCG Danish 2 to secrete high levels of recombinant hIFN-α2b.

Direct effects on tumor cells

We found that both BCG and rBCG inhibited growth of a mouse bladder cancer cell line, inducing morphological changes and apoptosis, while rBCG was significantly more potent. IFN induces tumor cell apoptosis by promoting protooncogenes and TNF-α receptor expression, or by inhibiting intracellular proteins (23,24). Following administration to mice, we observed BCG adherence to the bladder intima prior to bladder pathologic changes, a process that has previously been reported to involve non-specific, physicochemical and specific receptor-ligand-mediated events (7,25). Phospholipids, lipids, wax D, bacterial proteins and lipopolysaccharides in BCG are all strongly immunogenic (26–28), and BCG cell wall protein and Arabian-polysaccharide pathogen-associated molecular patterns induce expression of lysosomal membrane protein and apoptosis in host cells (27–29). A BCG cell wall glycolipid, trehalose dimycolate, has also been reported to damage host cells by attacking the mitochondrial membrane, affecting cellular respiration and energy metabolism, destroying microsomal enzymes and inducing programmed cell death (30–33). Furthermore, internalized BCG increased production of intracellular cytotoxic nitric oxide, which at a high concentration, causes DNA damage, and cytostatic and cytotoxic effects (7). BCG has also been reported to induce upregulation of certain surface molecules and cytokines in epithelial cells (34,35), and activation of the signal transduction pathways involving activator protein 1 and NFκB (36,37).

We also observed cytoplasmic Fas in mouse bladder tumor cells, and administration of rBCG and BCG upregulated Fas expression on the surface of the tumor cells. Fas typically is highly expressed in rapidly proliferating cells and injured tissues, and the triggering of Fas by its ligand induces apop-tosis in target cells (38). Expression of Fas can be upregulated by IL-2 and IFNs (39), but we observed no significant difference between Fas expression in the rBCG- and BCG-treated cells, potentially due to the species specificity of human IFN.

Immunomodulation

BGC has been hypothesized to ameliorate the aberrant imbalance of T helper and cytotoxic T cell subsets observed in the bladder during proliferation of cancer cells (3). Systemic immunity is induced by administration of BGC in patients with bladder cancer and a cytokine profile typical of active tuberculosis can be observed (40). In our mouse model, the circulating levels of CD3+ and CD4+ cells, and the CD4+/CD8+ ratio was lower in the tumor-bearing mice than in the normal animals. This immune response was likely caused by activation and proliferation of antigen-specific T lymphocytes, followed by immune tolerance or apoptosis, and depletion of T lymphocytes by activation-induced cell death. Administration of both BCG and rBCG induced lymphocyte proliferation, and regulated lymphocyte subsets, adjusting cellular immune function. CD4+ and CD3+ cell counts recovered to the levels observed in healthy control animals, yet the level of CD8+ cells was not significantly influenced. The CD4+/CD8+ ratio was also increased, but no significant difference was observed between the effect of BCG and rBCG.

Local immune responses

A small number of local lymphocytes are usually observed in the tumor tissue of patients with bladder tumors, mainly distributed in the submucosa (41). Administration of BCG in our mouse model of bladder cancer caused extensive local inflammation in the bladder wall. Expression of CD3, CD20 and Gr1 in the tumors was negative or weak, yet administration of BCG or rBCG induced influx of PMNs and other inflammatory cells. BCG appears to activate local non-specific Th1 type cells and cytokines (42). We observed that BCG and rBCG induced acute inflammation of the bladder, characterized by a strong vascular component and edema, followed by a gradual influx of mono-cytes/macrophages, T lymphocytes and NK cells, which form chronic granuloma-like cellular infiltrates in the suburothelial st roma (3).

In vitro, BCG also induced MHC-I expression on mouse MB49 cells, as previously reported by Shankaran et al (43), and we found that rBCG did so more potently than BCG. Tumor cells often downregulate MHC-I to avoid immune surveillance (44), yet IFN is well known to induce MHC-I (45). Increased expression of MHC-I on the tumor cell surface likely enhances the immunogenicity of these cells.

Cytokine release

Patients with bladder tumors often exhibit a marked polarization towards expression of Th2 type cyto-kines while expression of Th1 cytokines are suppressed (46). Administration of BCG adjusted the imbalance of Th1 cyto-kines (IL-2, IL-12, IFN-γ and TNF-α) leading to detection of these cytokines in the urine of BCG-treated patients (7). The induction of proinfammatory cytokines, specifically IFN-γ, TNF-α and IL-2, are crucial for the cytotoxic effect of live BCG-activated cells (47). BCG-activated lymphocytes and macrophages are the most likely sources of these cytokines, but at present, other cellular sources such as urothelial cells cannot be ruled out (7,46). We found that administration of BCG and rBCG promoted Th1 type immunity in the tumor-bearing mice. In vitro, IFN-α2b enhances the BCG induction of Th1 immune responses in human PBMCs (48), yet this effect was not observed in the peripheral blood of our experimental animals. T cell-mediated cell lysis and release of regulatory cytokines have been shown to represent late acquired immune events of the antitumor effector phase (49), to induce tumor apoptosis and prevent tumor growth. For example, IL-12 promotes the activation of effector cells and induces IFN-γ, IL-2 and TNF-α (50). TNF-α kills tumor cells directly and also induces apoptosis by the Fas/FasL-mediated path or others, which all could be strengthened by IFN.

In conclusion, our findings suggest that the therapeutic effects of BCG can be attributed in part to the rapid accumulation of antigen-presenting, and activated immune cells responsible for the production of a multiphasic immune response as demonstrated by the presence of the Th1 or Th2 phenotype. IFN directly inhibits proliferation and angio-genesis (51), and further immunomodulation of tumor MHC expression (52), could be advantageous in bladder cancer treatment. According to our results rBCG also exhibited antitumor activity in the bladder of mice transplanted with a bladder cancer cell line. Mice administered rBCG survived longer than those administered BCG plus endogenous IFN. Although the capacity of rBCG to inhibit tumor growth was not sup er ior t o t hat of BCG, t he in fluence of t he secreted cytokine may have been limited by species specificity. As a novel immune-modulatory agent for the treatment of human bladder cancer, rBCG has excellent prospects for development, but more in-depth exploration of its function and mechanisms is needed.

Acknowledgments

The present study was supported by the National Natural Science Foundation of China (no. 81402095). We thank Professor M.A. O’Donnell and Professor Yi Luo for kindly providing plasmid pMAO-4 to construct the recombinant hIFN-α2b-secreting BCG.

References