Uroepithelial cells can directly respond to Mycobacterium bovis bacillus Calmette-Guérin through Toll-like receptor signalling


Jun Miyazaki, Department of Urology, Institute of Clinical Medicine, University of Tsukuba, 1-1-1 Tennodai, Tsukuba-City, Ibaraki 305–8575, Japan. e-mail: kontama@v004.vaio.ne.jp



To investigate, in a human urinary tract cell line, the interaction of Toll-like receptor (TLR) signals with cytoplasmic adapter proteins MyD88 and bacillus Calmette-Guérin (BCG), and evaluate the epithelial cytokine response to BCG infection. Intravesical BCG therapy is effective against carcinoma in situ and as prophylaxis for recurrence, but although immunological mechanisms have been assumed, the mechanisms of the antitumour effects of BCG have not been completely elucidated.


The cell line was first screened for TLR expression by reverse transcription-polymerase chain reaction (RT-PCR). Total RNA was isolated from a human urinary cell line, Hu35E6E7, and cDNA synthesised. PCR was used to measure the gene expression of TLR-2, -3, -4, -5, -9, MyD88, MD-2, CD14 and interleukin-8 and -6. The Hu35E6E7 cell line was cultured in keratinocyte serum-free medium, and BCG was added to the cell culture. After Hu35E6E7 cells were stimulated by BCG for various periods, the total RNA of the cells was extracted. Quantitative real-time PCR was conducted for MyD88 using appropriate probes, and the expression of MyD88 analysed. The cell supernatant was collected, and the levels of interferon-γ, tumour necrosis factor-α, interleukin-2, -12, -4, -6, -10, -8 and -1β were assayed using an enzyme-linked immunosorbent assay.


Uroepithelial cells expressed TLR-2, -3, -4 and -9, and MyD88, MD2, CD14, interleukin-6 and -8 were also detected. At 3, 6, 9 and 12 h after adding BCG, quantitative PCR assay showed that the expression of MyD88 was maximal at 6 h. The presence of BCG stimulated the release only of interleukin-6 and -8 from Hu35E6E7 cells after 6 h. By contrast, interferon-γ, tumour necrosis factor-α, interleukin-2, -12, -4, -10 and -1β were not detected in the culture supernatant.


These results show that uroepithelial cells, but not immune cells, responded directly to BCG through TLR signalling. Further investigation is needed to determine the role of cytokines released from uroepithelial cells after BCG infection.


Toll-like receptor


pathogen-associated molecular pattern


glyceraldehyde phosphate dehydrogenase






Toll/interleukin-1 receptor (domain).


Intravesical Mycobacterium bovis BCG therapy is effective against carcinoma in situ and as prophylaxis for recurrence of bladder cancer [1]. In addition to the direct antitumour effect, it is widely recognized that intravesical BCG therapy is more potent in preventing tumour recurrence than intravesical chemotherapy [2]. Although intravesical BCG therapy is effective, it is not free from serious side-effects, e.g. high fever, granulomatous prostatitis, pneumonitis, hepatitis, and BCG sepsis [3]. Therefore, a greater understanding of the immunological mechanisms involved in BCG immunotherapy is needed for developing better and safer immunotherapy [4]. Several immunological mechanisms have been proposed [5], but the mechanisms of the antitumour effects of BCG have not been completely elucidated [6].

Recently, several investigators reported that human Toll-like receptors (TLRs) TLR-2 and -4 are involved in Mycobacterium tuberculosis-mediated intracellular signalling in vitro [7]. These receptors are termed pattern-recognition receptors because they recognize repetitive patterns, i.e. pathogen-associated molecular patterns (PAMPs), present on diverse microbes including Gram-positive and Gram-negative bacteria, fungi, and mycobacteria. Tsuji et al. [7] showed that both TLR-2 and TLR-4 act as receptors of the cell-wall skeleton of BCG by using murine macrophage cell lines lacking TLR-2 or TLR-4. Therefore, it appears that several kinds of human TLRs are closely involved in the immune response to Mycobacterium tuberculosis through interaction with PAMPs presented in Mycobacterium tuberculosis.

While the overall consequences of BCG binding to the uroepithelial surface are poorly understood, it is clear that BCG interaction with uroepithelial cells in the epithelium is active rather than passive. BCG adherence induces the expression of genes by bladder tumour cells, specifically interleukin-6 [8]. Chen et al. showed that BCG adherence stimulates tumour cell expression of interleukin-6 through an immediate-early pathway requiring the signal transducers nuclear factor NF-κB and AP-1 [9].

Epithelial linings are often exposed to microbes, and the epithelial cell itself is active in mucosal immunity on bacterial attachment [10]. This might occur through adhesion-mediated signalling by receptor interaction or through an attachment-facilitated presentation of conserved microbial patterns to TLRs on epithelial cells. Based on this background, in the present study we used a human urinary tract cell line, Hu35E6E7, to investigate the interaction of TLR signals with an adapter protein MyD88 and BCG, and also evaluated the epithelial cytokine response to BCG infection.


Lyophilized preparations of BCG, Tokyo 172 (Japan BCG Laboratory, Tokyo, Japan), containing 1.2–1.4 × 108 colony-forming units/mg, were used. Hu35E6E7 is an immortalized human uroepithelial cell line that was established from a normal ureteric epithelium after infection with a retroviral vector carrying human papilloma virus oncoproteins E6 and E7 [11]. The cells were kindly provided by Dr R. Oyasu, Department of Pathology, Northwestern University Medical School, USA. The cells were grown in keratinocyte-serum-free medium (Invitrogen Japan, Tokyo) supplemented with 100 µg/mL of streptomycin and 100 U/mL of penicillin (Invitrogen), and incubated in a humidified atmosphere of 95% air and 5% CO2 at 37 °C.

For the adherence assay, Hu35E6E7 cells were incubated in the above medium in a 25-cm2 culture flask (Asahi Techno Glass Co., Japan) at 37 °C in 5% CO2. About 2.8 × 108 units/mg BCG was added to each flask in a multiplicity of infection of 1 : 100. The culture plates were incubated for various periods in a 37 °C incubator with 5% CO2. At the end of infection, cells were washed three times with PBS.

TLR expression was first screened using RT-PCR; uroepithelial cells were cultured as described above, and total RNA was isolated from Hu35E6E7 cells by the guanidine thiocyanate-phenol-chloroform method (Isogen, Nippon Gene Co., Tokyo, Japan). RNA samples were quantified by spectrophotometric analysis, and cDNA synthesised according to the manufacturer's instructions (Invitrogen) with 1 µg of total RNA using Super ScriptTM First-Strand Synthesis System for RT-PCR. PCR was used to measure the gene expression of TLR-2, -3, -4, -5, -9, MyD88, MD-2, CD14, and interleukin-8 and -6. All primers were reported previously [12,13]. The cDNA reactions were diluted 1 : 10, and 2 µL of the diluted cDNA reaction was added to 18 µL Light Cycler PCR containing 0.5 µm of each primer

Quantitative real-time PCR was conducted for MyD88 and glyceraldehyde phosphate dehydrogenase (GAPDH) using TaqMan probes (Applied Biosystems, Foster City, CA, USA); MyD88 and GAPDH primers were used in TaqMan Gene Expression Assays Hs00182082_m1 and Hs99999905_m1. A total of 2 µL of cDNA was produced following the TaqMan assay reagents protocol in an ABI Prism 7000 system, used to analyse the standards and to quantify samples. Quantitative real-time PCR was performed in at least three independent experiments.

To measure T-helper (Th) 1 and Th2 cytokines, concentrations of nine cytokines, including interferon-γ, TNF-α, and interleukin-2, -12, -4, -6, -10, -8, and -1β were measured using the Human Th1/Th2 Cytokine Cytometric Bead Array Kit II (BD PharMingen, San Diego, CA, USA).


To determine whether uroepithelial cells can respond directly to BCG through TLR signalling, we first examined the TLR expression profile of Hu35E6E7. TLR-2, -3 and -4 were detected previously in uroepithelial cells [14]. We also examined the expression of TLR-5 and -9, both known to be closely related to bacterial ligands.

Total RNA from the Hu35E6E7 cell line was analysed using RT-PCR; in the present study there was expression of TLR-2, -3 and -4 (Fig. 1). The expression profile is consistent with that of T-24 bladder cancer cells as reported by Backhed et al. [14]. Hu35E6E7 cells also expressed TLR-9, the TLR known to respond to bacterial ligands. To our knowledge, there is no previous report of TLR-9 expression in uroepithelial cells. MyD88, MD2 and CD14, all key molecules in TLR signalling, were also expressed (Fig. 1). Finally, two cytokines, interleukin-6 and -8, were also expressed. After infection by BCG Tokyo 172 strain, the cellular response of Hu35E6E7 uroepithelial cells was assessed at 3, 6, 9 and 12 h (Fig. 2). The expression of MyD88 increased gradually and reached a maximum after 6 h, suggesting that the TLR gene in the uroepithelial cells responded to BCG. We also investigated the cytokine response of Hu35E6E7 cells; the culture supernatants were analysed for the release of nine cytokines. The presence of BCG stimulated the release of only interleukin-6 and -8 from cells after 6 h (Fig. 3). By contrast, interferon-γ, TNF-α, interleukin-2, -12, -4, -10 and -1β were not detected in the culture supernatant.

Figure 1.

In vitro expression of TLR, MyD88, MD2, CD14, interleukin-6 and -8 in uroepithelial cells. Total RNA from uroepithelial cells was used in a PCR assay to assess the expression pattern of TLRs in vitro. TLR-2, -3, -4 and -9 were expressed, but TLR-5 was not. MyD88, MD2 and CD14, all key molecules in TLR signalling, were also expressed. Two kinds of cytokine gene, interleukin-6 and -8, were also detected. M, molecular marker φX174/Hae III digest.

Figure 2.

MyD88 mRNA was induced by M. bovis BCG after 6 h.

Figure 3.

The interleukin-6 and -8 response of uroepithelial cells to M. bovis BCG infection, as assessed by analysis of culture supernatants for the release of the nine cytokines, using an ELISA. The presence of BCG stimulated the release only of interleukin-6 and -8 from cells after 6 h; the other cytokines were not stimulated.


Several types of immune cells, normal uroepithelial cells, and bladder cancer cells are thought to participate in the complex immune mechanism of intravesical BCG therapy. In the present study we used an immortalised normal human uroepithelial cell line to test whether uroepithelial cells, but not immune cells, can respond directly to BCG in vitro. To our knowledge, no previous reports have addressed this issue, probably because of the cost and instability of primarily cultured normal human uroepithelial cells. However, several investigators reported the expression of TLR family genes in urological cancer cell lines. Backhed et al. [14] used RT-PCR to show that the human bladder cancer cell line T-24 expresses TLR-2, -3 and -4, and the human kidney cancer cell line A498 expresses TLR-2, -3 and -5. In response to lipopolysaccharide (LPS) stimulation, T-24 cells released interleukin-6 and -8, whereas A498 cells did not. The different TLR expression profiles, especially the absence of TLR-4 in A498 cells, might explain the different cytokine responses of A498 cells and T-24 cells to LPS [14].

In the present study, expression of TLR-2, -3 and -4 in immortalised normal human uroepithelial cells from the Hu35E6E7 cell line, was detected by PCR. Hu35E6E7 cells also expressed TLR-9, an essential element in CpG DNA-mediated cellular activation [15]. Bacterial DNA triggers cells of the innate immune system to proliferate and become functionally active [16], and this stimulation is mediated by unmethylated CpG motifs that are present at high frequency in bacterial DNA, but which are rare in mammalian DNA. Thus, the uroepithelial cell response to CpG DNA mediated by TLR-9 might be an important antitumour mechanism of BCG [17]. The present study also showed secretion of MD-2 and CD14; MD-2 interacts with the extracellular domain of TLR-4 and significantly enhances host cell responses to LPS [18], while CD14 interacts with TLR-4, the signalling component of the LPS receptor. Samuelsson et al. [19] reported that CD14 is important in uroepithelial cell responses to LPS or Escherichia coli.

Recently, several investigators have focused on TLRs, a family of transmembrane receptors homologous to the Drosophila Toll protein, as the immunomodulatory molecules involved in the reaction to BCG [20]. The cytoplasmic portion of TLRs, which is very similar to that of the interleukin-1 receptor (R) family, is now termed the Toll/interleukin-1R (TIR) domain. Despite this similarity, the extracellular portions of both receptors are structurally unrelated. The interleukin-1R is characterized by an Ig-like domain, whereas TLRs have leucine-rich repeats in the extracellular domain. TLR ligand is recognised through an interaction of the extracellular leucine-rich domain of the TLR with its PAMP. Both TLR and interleukin-1R interact with an adapter protein MyD88 in their TIR domain. When stimulated, MyD88 recruits interleukin-1R-associated kinase to the receptor; this kinase is activated by phosphorylation and then associates with TRAF6, leading to the activation of NF-κB and the production of cytokines such as the neutrophil chemoattractant interleukin-8, -6, -12, TNF, and other inflammatory cytokines [21].

Cytokines are early markers of the epithelial response to infection and play a key role in the innate defence [22]. Urinary cytokine levels are high in patients with UTI [23], and epithelial cells have been identified as early producers of cytokines in a murine UTI model [24], but the epithelial cytokine response of the human mucosa in situ has not been investigated. One remarkable effect of instilling BCG into the bladder is a transient secretion of several cytokines in voided urine [25], including a large burst of urinary interferon-γ, along with other Th1 cytokines (interleukin-12 and -2), a common feature in BCG responders [25]. By contrast, higher levels of Th2 cytokines interleukin-10 and/or -6 appear to be associated with BCG failure [25]. Correspondingly, in mouse models, interferon-γ and interleukin-12, but not -10 or -4, are suggested to be required for the immunotherapeutic control of orthotopic bladder cancer [25]. These observations suggest that effective BCG therapy of bladder cancer requires proper activation of the Th1 immune pathway [25], which is thought to be a result of a complex interaction of the uroepithelial cells and the immune cells to BCG. However, the release of these cytokines is the result of a complex immune response between BCG and various kinds of immune cells, and uroepithelial cells. In the present study, Hu35E6E7 cells, with no immune cells, released significant interleukin-6 and -8 within 6 h of BCG exposure. More importantly, quantitative PCR clearly showed that the expression of MyD88 was maximal at 6 h after BCG exposure. Taken together, these results suggest that the uroepithelial cells can directly recognize BCG through TLRs and MyD88. Subsequently, the uroepithelial cells release the inflammatory cytokines interleukin-6 and -8.

Interleukin-6 is thought to be a major source of early Th1/Th2 control during CD4+ T-cell activation, as it contributes to the promotion of the Th2 response and simultaneously inhibits Th1 polarization. Interleukin-6 activates interleukin-4 production by CD4+ T cells and their differentiation into Th2 effector cells. Moreover, it inhibits Th1 differentiation by interfering with interferon-γ signalling and the development of Th1 cells [26]. The large interleukin-6 response during BCG therapy and the virtual absence of urinary interleukin-4 seem to negatively regulate the Th1 response in BCG recognition. Interleukin-8 is a pro-inflammatory cytokine with strong chemotactic properties, attracting T lymphocytes and neutrophils. Interleukin-8 is induced rapidly, after the first instillation, during BCG therapy [27]. Dendritic cells, macrophages, and several other cells, including TCC cell lines, are known to produce interleukin-8. Saint et al. [28] showed that leukocyturia appears to correlate with both the efficacy and tolerability in intravesical BCG induction and maintenance therapy for superficial bladder cancer. Secretion of interleukin-8 may indicate the importance of an interaction between BCG and normal uroepithelial cells in initiating the immune response involved in intravesical BCG therapy.

In conclusion, the present results show that uroepithelial cells, but not immune cells, directly respond to BCG through TLR signalling. The present findings cannot directly translate into clinical practice, but recent evidence suggests that some TLRs play a role in the pathogenesis of infectious and/or inflammatory diseases [29]. Some epidemiological investigations show that TLR-2 gene polymorphism may yield important information for assessing risk profiles of susceptibility to Mycobacterium leprae infection [30]. Furthermore, various therapeutic approaches targeting TLR signalling are now being investigated. Therefore, TLR signalling is a possible molecular target for patient selection of BCG therapy or risk assessment, and treating adverse events of the therapy. Further investigation is needed to determine the role of cytokines released from uroepithelial cells after BCG infection.


None declared.