Bacillus Calmette–Guerin (BCG) immunotherapy for bladder cancer: Current understanding and perspectives on engineered BCG vaccine

Abstract

Since the first report in 1976, accumulated clinical evidence has supported intravesical Bacillus Calmette–Guerin (BCG) therapy as one of the standard methods of management of intermediate‐ and high‐risk non‐muscle invasive bladder cancer. Despite its efficacy, intravesical BCG therapy is associated with a variety of adverse events (AEs), most of which are tolerable or controllable with supportive care. However, some patients receiving intravesical BCG therapy may experience uncommon but severe AEs, leading to cessation of BCG therapy. Not all, but most severe AEs result from either local or systemic infection with live BCG. Intravesical instillation of BCG elicits multiple immune reactions, although the precise immunological mechanism of BCG therapy is not clear. It is convenient to separate the complex reactions into the following three categories: infection of urothelial cells or bladder cancer cells, induction of immune reactions, and induction of antitumor effects. Recently, our knowledge about each category has increased. Based on this understanding, predictors of the efficacy of intravesical BCG therapy, such as urinary cytokine measurement and cytokine gene polymorphism, have been investigated. Recently, preclinical studies using a novel engineered mycobacterium vaccine have been conducted to overcome the limitations of BCG therapy. One approach is Th1 cytokine‐expressing recombinant forms of BCG; another approach is development of non‐live bacterial agents to avoid AEs due to live BCG infection. We also briefly describe our approach using an octaarginine‐modified liposome‐incorporating BCG cell wall component to develop future substitutes for live BCG. (Cancer Sci 2013; 104: 22–27)

Intravesical Bacillus Calmette–Guerin (BCG) therapy is undoubtedly the most successful immunotherapy for cancer.1 In 1976, Morales et al.2 reported the prophylactic effect of topical application of live BCG to the bladder to prevent bladder cancer recurrence. In the original protocol, 120 mg BCG diluted with 50 mL saline was instilled into the bladder via a urethral catheter with an indwelling time of 2 h and was repeated once a week for 6 weeks. Since then, intravesical BCG therapy has been widely used and is recognized as a standard method of management of intermediate‐ and high‐risk non‐muscle invasive bladder cancer, although the optimal schedule and duration are still being tested. Despite its efficacy, intravesical BCG therapy is not free from severe and possibly fatal complications due to the use of live BCG. However, the precise immunological mechanism of BCG therapy is not clear, and understanding the immune responses induced by BCG is needed to develop a more active and less toxic immunotherapy. In this review, we summarize basic and clinical information about intravesical BCG therapy, and also describe the prospect of an engineered mycobacterium vaccine.

Current clinical use of intravesical BCG therapy

At initial diagnosis, approximately 70% of bladder cancer patients are diagnosed with non‐muscle invasive bladder cancer (NMIBC), which include tumor stages Ta (low‐grade noninvasive), T1 (invasion of lamina propria), and carcinoma in situ (CIS0).3 The remaining patients have muscle‐invasive tumor (stage T2 or higher), which usually requires radical cystectomy. In contrast, most visible NMIBC lesions (Ta or T1 tumor) can be removed by endoscopic surgery, i.e., transurethral resection (TUR). For complete resection, a second TUR within 6 weeks after the initial resection is advised for T1 tumor.4 One problem in the management of NMIBC is the high intravesical recurrence rates ranging from 30% to nearly 80%, depending on the risk profile. Several mechanisms for intravesical recurrence have been proposed including microscopic persistence of tumor, cancer cell implantation, and new tumor formation. More importantly, NMIBC may progress to muscle‐invasive cancer during repeated episodes of intravesical recurrence. Urothelial CIS, unlike CIS in other organs, has high malignant potential. CIS has over 50% risk of progression to muscle‐invasive cancer.5

Intravesical therapy is used for prophylaxis against intravesical recurrence after TUR and also for treating CIS. In intravesical chemotherapy, high‐molecular‐weight anticancer agents such as mitomycin C or doxorubicin are used to avoid systemic toxicity from absorption.6 Generally, intravesical BCG therapy is more effective than intravesical chemotherapy. In addition to effectively preventing recurrence, intravesical BCG therapy shows a potent direct effect against bladder cancer. Akaza et al.7, 8 reported high complete response rates against existing CIS and also papillary bladder cancer. However, intravesical BCG therapy is associated with higher toxicity than intravesical chemotherapy. Therefore, recent guidelines recommend risk‐adapted management rather than routine use of BCG. The risk stratification criteria differ in each guideline; however, there is agreement that intravesical BCG is preferred therapy for the high risk group.9 Table 1 shows the risk stratification and treatment recommendations of the International Bladder Cancer Group.9

Table 1.

Risk stratification and treatment recommendation for NMIBT

	Definition	Treatment recommendation
Low risk	Solitary, primary low‐grade Ta tumors	TURBT+ an intermediate single intravesical chemotherapy
Intermediate risk	Multiple or recurrent low‐grade tumors	TURBT+ intravesical BCG with induction and maintenance or TURBT+ intravesical chemotherapy
High risk	Any combination of T1 and/or G3 and/or CIS	TURBT+ intravesical BCG with induction and maintenance
		Immediate radical cystectomy should be considered for high grade, multiple T1 tumors; T1 tumors located at a site difficult to resect; residual T1 tumors on resection; high‐grade tumors with CIS

Originally, BCG was administered each week for 6 or 8 weeks, and this induction therapy was used as a standard treatment. Recently, to improve the efficacy of BCG therapy, many different maintenance schedules have been reported, ranging from a total of four instillations in 1 year, to 27 instillations over 3 years.10 However, the optimal number of induction instillations and the optimal frequency and duration of maintenance instillations remain unknown. Among them, the largest randomized clinical trial (RCT), by the Southwest Oncology Group (SWOG), has shown a possible benefit to addition of maintenance therapy to induction therapy. The maintenance therapy used consisted of intravesical BCG each week for 3 weeks for 3, 6, 12, 18, 24, 30, or 36 months.11 Hinotsu et al.,12 using the smoothed hazards analysis of recurrence, demonstrated a persistent preventive effect for up to approximately 2 years after TUR in patients receiving maintenance therapy. In addition to the RCTs, a series of meta‐analyses revealed superior results in studies that provided BCG maintenance therapy for 1 year or more.13, 14

Despite its efficacy, intravesical BCG therapy is associated with a variety of adverse events (AEs), as shown in Table 2. In intravesical BCG therapy, systemic AEs in addition to local AEs are observed, which is in sharp contrast to the observation that systemic AEs are rarely seen in intravesical chemotherapy.15 The most frequent AEs are urinary frequency and irritative bladder symptoms, hematuria, low grade fever, and flu‐like symptoms. Most of them are tolerable or controllable with supportive care, but the incidences and severity of AEs are generally increased in maintenance therapy. Accordingly, in the original SWOG maintenance protocol, many patients dropped out, largely due to AEs, and only 16% of patients completed the planned maintenance therapy.11 Intravesical BCG therapy may be complicated by uncommon but severe AEs (Table 2).16, 17 Most but not all severe AEs are due to either local or systemic infection with live BCG.16, 17 When severe AEs developed, patients were obliged to discontinue BCG therapy at that point in time. The prompt initiation of antituberculous therapy is recommended in most of these cases. The incidence of BCG sepsis, the most serious and possibly fatal AE, is estimated to be 0.4%. More than 10 deaths due to BCG sepsis have been reported since 2006.18 Because BCG sepsis is due to systemic absorption of live BCG, a history of TUR within 2 weeks, traumatic catheterization, and presence of macroscopic hematuria are major contraindications for intravesical BCG therapy.

Table 2.

Adverse effects of intravesical Bacillus Calmette–Guerin (BCG) therapy

Local	Systemic
Frequent adverse effets
Urinary frequency	Fever
irritative bladder symptoms	Flu‐like symptoms
Hematuria	General malaise
Infrequent severe adverse effects (complication)
Granulomatous prostatitis	BCG sepsis
Granulomatous epididymoorchitis	Allergic reaction
Contracted bladder	Hepatitis
	Pneumonitis

Current understanding of immunological mechanisms of intravesical BCG therapy

Intravesical instillation of BCG results in multiple immune reactions. Although the precise immunological mechanism of BCG therapy is not clear, it is convenient for understanding to separate the complex reactions into the following three categories: infection of urothelial cells or bladder cancer cells, induction of immune responses, and induction of antitumor effects.

Infection of urothelial cells or bladder cancer cells by BCG

Akaza et al.19 showed that the coexistence of BCG with tumor cells activated local immunity in a syngenic mouse subcutaneous tumor model. In intravesical therapy, it is also generally assumed that the attachment of intraluminal BCG to urothelial cells or bladder cancer cells is the first step in BCG‐induced immune responses. The normal urothelium is covered with glycosaminoglycans (GAGs) which, because of their negative charge, are thought to protect bladder urothelium from BCG and other bacteria, whose surfaces are highly negatively charged.20 This may explain why a high dose of BCG is needed for intravesical therapy. Besides this physicochemical interaction, intravesical BCG attachment is also mediated by fibronectin, a glycoprotein of the extracellular matrix, which is distributed in normal and malignant urothelium. The fibronectin‐mediated BCG attachment is considered to be a crucial step because an anti‐fibronectin antibody inhibited the antitumor activity of BCG.21 After attachment, BCG enters urothelial cells and cancer cells. In an animal model, BCG was detected within umbrella cells and underlying urothelial cells as soon as 24 h after instillation into the mouse bladder.22 Cellular internalization of BCG was also confirmed in cultured human bladder cancer cell lines.23, 24 It is well known that internalization of bacteria including BCG modulates the antigen processing of antigen‐presenting cells (APCs).25 Interestingly, BCG can also enhance the surface expression of histocompatibility complex (MHC) class II and CD1 of bladder cancer cells.26 Lattime et al.27 reported that murine bladder cancer cells present the antigen to BCG‐specific CD4⁺ T cells. These observations suggest that not only professional APCs but also BCG‐infected urothelial cells and bladder cancer cells take part in initiation of the early immune response. At present, the fate of BCG internalized into the urothelium is not fully understood, but, in the human bladder, BCG DNA was detectable in the bladder wall more than 2 years after intravesical BCG instillation.28 Also, long‐lasting MHC class II expression was observed in the urothelium after BCG treatment.29

Induction of immune responses

As shown in Table 3, various types of cytokines are detectable in the urine after BCG instillation. The cytokines in the urine indicate an immunological response specific for BCG instillation, although not completely. Most of the cytokines are detected after the second BCG instillation, and their concentrations increase after BCG instillations.30, 31 Macrophages and activated lymphocytes are thought to be major sources of these cytokines. In contrast, some cytokines such as interleukin (IL)‐1, IL‐6, and IL‐8 are detected after the first BCG instillation.32, 33 Miyazaki et al.34 reported that cultured urothelial cells release IL‐6 and IL‐8 directly in response to BCG. Although this urothelial cell reaction is not specific to BCG, urothelial cells are supposed to be involved in the early phase of mucosal cytokine network induction by BCG. In addition to Th1 and Th2 cytokines, recently, several investigations showed that IL‐17 plays a crucial role in generation of Th1‐cell responses in BCG vaccination.35 At present, clinical data about IL‐17 is limited, but in a mouse model, increases in IL‐17 family genes and urinary IL‐17 concentrations are reported.22 Recently, Takeuchi et al.36 reported that IL‐17‐producing T cells induced recruitment of neutrophils to the bladder.

Table 3.

Immunological mechanisms of intravesical Bacillus Calmette–Guerin (BCG) instillation

Infection by BCG of urothelial cells or bladder cancer cells	Attachment of BCG to urothelium and cancer cell
	Internalization of BCG in urothelium and cancer cell
	Increased expression of antigen‐presenting molecules
Induction of immune responses	Th1 cytokine (NF‐γ, IL‐2, TNF, IL‐12)
	Th2 cytokine (IL‐6, IL‐10)
	Others (IL‐8, IL‐17)
Induction of antitumor effects	Cytotoxic T lymphocyte
	NK cell
	Neutrophil
	γδ T cell
	Macrophages

Induction of antitumor effects

The antitumor effect of BCG has been recognized as dependent on T‐cells. In particular, the role of Th1 cell‐mediated immunity including CD4+ T cells and CD8+ cytotoxic T lymphocytes (CTLs) is well known. Ratlieff et al.37 showed that the athymic nude mouse did not undergo BCG‐induced tumor rejection. However, the actual antitumor effector mechanism is still unclear. In addition to the acquired immunity, the innate immune response plays a role in the antitumor effect of BCG. Brandau et al.38 reported that BCG therapy was completely ineffective in natural killer (NK)‐cell deficient beige mice and in mice treated with an anti‐NK cell monoclonal antibody. More recently, several investigations demonstrated the antitumor effect of neutrophils, which comprise the major cell subset in the leukocyturia observed after BCG instillation, or TNF‐related apoptosis‐inducing ligand (TRAIL) produced by neutrophils.39 Macrophages and other types of innate immune cells are also reported to be involved in the antitumor effect of BCG.40, 41

Predictors of efficacy of intravesical BCG therapy

Despite its efficacy, 30–50% of patients ultimately fail to respond to intravesical BCG therapy. Therefore, it is useful to identify the predictors for BCG response. Analysis of data from four RCTs showed that female sex, history of recurrence, multiplicity, and the presence of associated CIS are significant independent predictors of recurrence.42 In addition to these conventional clinical parameters, the usefulness of various immunological parameters has been tested. Among the cytokines detected in urine, IL‐2, a Th1 cytokine closely related to CTL response, has been most extensively studied. Several studies showed that a high level of urinary IL‐2 was associated with favorable clinical outcome.43, 44 Urinary IL‐8, a potent chemoattractant for neutrophils, was reported to be a possible predictor of response.45, 46 Thalmann et al.46 reported that early elevation of urinary IL‐8 and IL‐18 after BCG therapy predicted recurrence‐free survival. IL‐18 is one of the cytokines that is known to play a role in the generation of protective immunity to mycobacterium. These observations suggest the usefulness of measuring urinary cytokines, but cytokine monitoring in clinical practice is still investigational. Another expectation for cytokine measuring is that it will help to establish the optimal frequency and duration of maintenance therapy. Several investigations showed that urinary IL‐2 peaked earlier during maintenance therapy than during the induction cycle. This may reflect a booster effect of maintenance BCG.47, 48 However, Saint et al. reported that favorable IL‐2 production gradually switched to production of IL‐10, a Th2 cytokine, during subsequent maintenance BCG instillations.43 This switch of urinary cytokine profile might be associated with AEs. Currently, available data about urinary cytokine profile during maintenance therapy is limited; further study with a large number of patients is needed to demonstrate the validity.

As another promising predictor, cytokine gene polymorphism has been investigated. Although some results are still contradictory, polymorphisms of IL‐6, IL‐8, and TNF‐α might be a candidate.49 More recently, Wei et al.50 reported that polymorphism of oxidative stress genes in NMIBC patients might affect response to BCG therapy.

Prospect of engineered mycobacterium vaccine

Intravesical BCG therapy is not free from treatment failure, and AEs include rare but possibly fatal complications. Therefore, major clinical efforts have been made to develop more active and less toxic modes of immunotherapy; these include use of a combination of interferon, reduction of BCG dose, and prophylactic administration of tuberculostatic agents. Although some strategies show promising results, further clinical evaluation is needed.

Based on recent increases in understanding of BCG‐induced immune responses, a variety of preclinical studies have been conducted to overcome the limitations of BCG therapy. One approach is the generation of Th1 cytokine‐expressing recombinant forms of BCG (rBCG). In animal models, some cytokine‐expressing rBCGs showed promising results against malignant melanoma and breast cancer; however, studies using intravesical models of bladder cancer are still limited.51

Another approach is the development of non‐live bacterial agents such as killed mycobacterium or the mycobacterium extracts to avoid AEs caused by using live BCG. Although the BCG cell wall (BCG‐CW) has long been investigated for this purpose, its clinical use is limited because of difficulties relating to solubility and stability. The BCW‐CW consists of highly characteristic hydrophobic molecules, including mycoloyl glycolipids, mannose‐containing lipoglycans, and the CW skeleton (BCG‐CWS), most of which stimulate Th‐1 type immune responses through production of TNF‐α, IL‐12, and interferon‐γ in experimental animal systems.52, 53, 54 Despite the immunotherapeutic potential, BCG‐CW is hampered by unfavorable physicochemical characteristics. In addition to its hydrophobic properties, the negative surface charge causes poor cellular association.55 As described earlier, the urothelium is covered with negatively charged GAGs. Therefore, the BCG‐CW and cell membrane electronically repel each other, which obstructs internalization into urothelial cells and cancer cells, a critical step for evoking the BCG‐induced immune reaction. Furthermore, because urothelial cancer cells are not phagocytic, transportation of a non‐infectious molecule such as BCG‐CW (but not live BCG) across the cell membrane is performed passively by endocytosis. Due to its hydrophobic properties, BCG‐CW tends to form relatively large agglutinated clumps in an aqueous environment, which also obstructs its transportation by endocytosis.

Therefore, we hypothesized that the immune response would be evoked if we could successfully deliver BCG‐CW into the cytoplasm of murine bladder cancer cells by using an appropriate vector. We have applied octaarginine‐modified liposomes (R8‐liposomes) as a vector to transport BCG‐CW into the cytoplasm effectively. R8‐liposomes were developed to transfer highly negative charged DNA molecules into the cytoplasm by macropinocytosis.56, 57, 58 R8‐liposomes resemble an envelope‐type virus and their surface‐modification is anchored by R8, a characteristic and efficient cell‐penetrating peptide.59 We have previously reported that a R8‐liposome‐incorporating BCG cell wall (R8‐liposome‐BCG‐CW) successfully attached to the surface of a mouse bladder tumor cell, MBT‐2 cells, and were efficiently internalized into the cytoplasm within 1 h of co‐incubation.60 As shown in Figure 1, internalized BCG‐CW was then distributed to the lysosome of the MBT‐2 cells. In a syngenic subcutaneous tumor model, R8‐liposome‐BCG‐CW has been shown to completely inhibit the growth of MBT‐2 tumors in vivo. We also examined the activity of R8‐liposomes incorporating the BCG cell wall skeleton (R8‐liposome‐BCG‐CWS) against human bladder cancer cell lines.57 The cell wall skeleton (CWS), a component of BCG‐CW, is more suitable for manufacturing with good quality. When co‐cultured with lymphokine‐activated killer (LAK) cells, in which NK cells are a major subpopulation, R8‐liposome‐BCG‐CWS can directly enhance the susceptibility of cancer cells to LAK cells through up‐regulation of natural‐killer group 2, member D (NKG2D) ligands.61 NKG2D is known to be an important activating receptor present on the surface of NK cells. Like R8‐liposome‐BCG‐CW, R8‐liposome‐BCG‐CWS showed a potent anti‐tumor effect in vivo. In an N‐butyl‐N‐(4‐hydroxibutyl)nitrosamine (BBN)‐induced rat bladder cancer model, rats receiving R8‐liposome‐BCG‐CWS intravesically showed a significantly reduced number of tumors, especially those due to simple hyperplasia.62 It is hoped that R8‐liposome‐BCGCWS, a non‐live bacterial agent, will contribute to providing a more active and less toxic tool as a future substitute for live BCG.

Figure 1.

The fate of internalized Bacillus Calmette–Guerin cell wall (BCG‐CW) was tracked. MBT‐2 cells were incubated with sulforhodamine‐labeled R8‐liposome BCG‐CW (green) (A) at 37°C for 1 h. Endosomes and lysosomes of MBT‐2 cells (red) were stained with 75 nm Lysotracker Red for 30 min before visualization (B). The cell nuclei (blue) were stained with Hoechst 33342. BCG‐CW in the cytoplasm was co‐localized with endosome/lysosome compartments (C).

Conclusion

Since the first report in 1976, the accumulated clinical evidence supports the use of intravesical BCG therapy as one of the standard methods of managing intermediate‐ and high‐risk NMIBC. Although the precise immunological mechanism of BCG therapy is still unclear, our understanding is increasing of reactions induced by BCG complexes, including infection of urothelial cells or bladder cancer cells, induction of immune responses, and induction of antitumor effects. A number of clinical and preclinical studies have been conducted to overcome the limitations of intravesical BCG therapy. Among them, development of a non‐live bacterial agent is one of the most promising candidates for a future substitute for live BCG.

Disclosure Statement

The authors have no conflict of interest.

(Cancer Sci, doi: 10.1111/cas.12075, 2012)