Bacteria and cancer: cause, coincidence or cure? A review

Salmonella typhi and gallbladder cancer

Worldwide annual incidence of gallbladder cancer (GC) is 17 million cases with high incidence rates in certain populations. The malignancy is usually associated with gallstone disease, late diagnosis, unsatisfactory treatment, and poor prognosis. The five-year survival rate is approximately 32 percent for lesions confined to the gallbladder mucosa and one-year survival rate of 10 percent for more advanced stages [26]. Over 90 percent of gallbladder carcinomas are adenocarcinoma [27] involving gallstones in 78% – 85% of cases [26].

There are several risk factors for gallbladder cancer. The main associated risk factors include cholelithiasis (especially untreated chronic symptomatic gallstones), obesity, reproductive factors, environmental exposure to certain chemicals, congenital developmental abnormalities of the pancreatic bile-duct junction and chronic infections of the gallbladder [26, 28]. The interplay of genetic susceptibility, lifestyle factors and infections in gallbladder carcinogenesis is still poorly understood [29], however a link has been specifically proposed between chronic bacterial infections of the gallbladder and Salmonella typhi [26].

The strongest epidemiological evidence of bacterial oncogenic potential, aside of Helicobacter pylori, concerns S. typhi. Infection with this bacterium of typhoid, can lead to chronic bacterial carriage in the gallbladder [30]. Recent epidemiological studies have shown that those who become carriers of S. typhi have 8.47 times the increased risk of developing carcinoma of the gallbladder compared with people who have had acute typhoid and have cleared the infection [26]. These findings agreed with earlier investigations by Welton et al. [31] and Caygill [30].

A case-control study by Welton et al [31] compared those who experienced acute infection with S. typhi to those who subsequently became chronic carriers following the 1922 typhoid outbreak in New York. Carriers were six times more likely to die of hepatobiliary carcinoma than matched controls. Additional evidence was found in an analysis of the 1964 typhoid outbreak in Aberdeen [30]. Their findings also suggested a strong association between chronic carrier status and hepatobiliary carcinoma. These studies also agreed, people who contracted typhoid but did not become carriers were not at higher risk of cancer [8, 26, 30, 31].

The highest incidence of gallbladder cancer (GC) in the world is among populations of the Andean area, North American Indians, and Mexican Americans. In Europe, the highest rates are found in Poland, the Czech Republic and Slovakia. The high rates observed in Latin America are primarily in populations with high levels of Indian mixture [32]. This evidence supports the notion that increased susceptibility to gallbladder cancer depends on genetic factors that predispose people to gallbladder cancer either as primary factors, or secondarily as promoters by favoring the development of cholesterol gallstones. The highest mortality rates are in South America, (3.5–15.5 per 100,000) and among Mexican Americans [26]. Incidence rates of GC in various ethnic groups in the USA confirmed the worldwide pattern, as GC was substantially more frequent among Hispanic than non-Hispanic white women and men. Interestingly, compared to non-Hispanic whites an excess of GC was also reported among American Indians in New Mexico, in agreement with the excess in incidence rates reported for American Indians and Alaskan Natives [33]. The malignancy is 3 times higher, however, among women than men in all populations [26].

Two main pathways to GC exist worldwide. The predominant pathway involves gallstones and resultant cholecystitis and affects women to a greater extent than men. The risk of developing gallstones in response to environmental factors is genetically determined, as shown by the marked tendency of gallstones to cluster in families [34]. The other pathway involves an anomalous pancreatobiliary duct junction (APBDJ), a congenital malformation of the biliary tract that is more frequent in Japan, Korea, and possibly China, than in Western countries [28]. In APBDJ, the premature junction of common bile and pancreatic ducts results in regurgitation of pancreatic juice into the gallbladder, leading to bile stasis and inflammation, though generally less severe than that resulting from gallstones [28].

Currently the prevention of gallbladder cancer in high risk populations depends upon the diagnosis of gallstones and removal of the gallbladder. Indeed, a strong inverse association between number of cholecystectomies and GC incidence and mortality rates can be found in many countries. The increase of GC mortality reported in Chile in the 1980s was related to decreased rates of cholecystectomies [35]. Increased rates led to the removal of gallbladders at risk, and a reduction of GC incidence and mortality in Europe and the United States [36].

Unfortunately, information about the genetic changes involved in gallbladder carcinogenesis is limited. Most studies have focused on gene abnormalities and deletions ("loss of heterozygosity") at chromosomal regions harboring known or putative tumor suppressor genes [28]. It appears, however, that TP53 inactivation has an important and early role in gallbladder carcinoma associated with gallstones and chronic inflammation. This inactivation would abrogate the tumor suppressor function of the p53 protein resulting in impairments in cell cycle control, cellular repair and apoptosis.

In contrast, KRAS mutations are frequent and early events in tumors associated with APBDJ [28] but detected less often in gallbladder carcinomas associated with gallstones. KRAS is an oncogene that encodes a protein that is a member of the small GTPase family. A mutation in this gene results in an abnormal protein implicated in several malignancies, including lung adenocarcinoma, ductal carcinoma of the pancreas and colorectal carcinoma among others.

Chlamydophila pneumoniae and lung cancer

Lung cancer is the leading cause of cancer death in the United States and many countries in the Western world. In 2002, the most recent year for which statistics are available, 90,121 males and 67,509 females died from lung cancer [37]. About 6 out of 10 people with lung cancer die within 1 year of finding out they have lung cancer. Between 7 and 8 will die within 2 years [38]. Although patients may experience a partial or complete response to treatment, most patients relapse and die. Increased dosage of chemotherapy or length of treatment has not been beneficial [39].

Chlamydophila (formerly Chlamydia) pneumoniae infection has been implicated in several chronic lung diseases by serology and direct antigen detection. Acute lower respiratory tract infection caused by C. pneumoniae seems often to precede attacks of asthma in both children and adults but is also involved in some exacerbations of chronic bronchitis. More importantly it seems to be strongly associated with chronic obstructive lung disease irrespective of exacerbation status. Moreover, persistently elevated C. pneumoniae antibody titers have been observed in sarcoidosis and lung cancer [40].

C. pneumoniae is a Gram-negative bacillus and an intracellular parasite that causes respiratory infection in more than 50% of adults. The route of transmission is usually by aerosol and in most cases these infections are mild. The bacterium is, however, an important cause of pneumonia, bronchitis, sinusitis, rhinitis and chronic obstructive pulmonary disease [41]. Respiratory infections from C. pneumoniae vary in different countries and populations, being endemic in the United States and epidemic in Scandinavian countries [19].

After acute infection the C. pneumoniae intracellular life cycle is characterized by the development of metabolically inert (and thus antibiotic resistant) atypical "persistent" inclusions. These inclusions contain increased quantities of chlamydial heat shock protein 60, a highly immunogenic protein implicated in the pathogenesis of chronic chlamydial infections. The resulting clinical course is acute symptomatic illness followed by chronic respiratory symptoms. Research also suggests that persistent C. pneumoniae inflammation correlates with increased risk of lung cancer [16, 17, 19]. Prospective and retrospective studies both report that individuals with elevated IgA antibody titers to this organism have 50% to 100% increased lung cancer risk [15].

In a study by Kocazeybek et al. [19] the relationship between chronic C. pneumoniae infection and lung carcinoma was examined. A total of 123 patients who were smokers and diagnosed with lung carcinoma based on clinical and laboratory (radiological, cytological) findings were examined. 101 (82.1%) of the cases were male. 70 had small-cell, 28 squamous-cell and 7 large-cell carcinomas, while 18 had adenocarcinoma. 123 healthy controls were matched to the cancer patients by age, gender, duration of smoking and locality.

Blood samples (5 ml) were withdrawn at the time of diagnosis (or enrollment for controls) and 1 month later. Values between IgG ≥512 and IgA ≥40 were set as the criteria for chronic C. pneumoniae infections. In male patients with lung carcinoma, IgG antibody titers of ≥512 and IgA antibody titers of ≥40 were found at a higher rate than in the control group, however, this ratio was not significant for female patients. These elevations in antibody titers were found in a total of 62 (50.4 %) cases, 54% of the male patients and 36% of the female patients. Chronic C. pneumoniae infections were seen statistically more often in male patients with carcinoma who were aged 55 years or younger than in controls (P < 0.001). No difference was reported between male patients with lung carcinoma over age 55 and controls or in blood titers between female patients and controls.

The relationship between C. pneumoniae infection and lung carcinoma was studied by Littman et al. [42] in a large prospective case-control study to investigate whether IgA antibody titers to C. pneumoniae were associated with lung cancer risk. A total of 508 pairs were enrolled and included both current and former smokers. Serum was collected at baseline and annually thereafter. Antibody determinations of each lung cancer subject and matched control were tested simultaneously in the same titration series in a blinded fashion. C. pneumoniae titers (IgA or IgG) ≥16 were considered seropositive, which was consistent with the cutoff used in other studies. Subjects were matched by age, gender, and smoking status at baseline. The median age of cases and controls was 59 years and about half were women. All subjects were also examined for demographic, lifestyle, dietary, and racial and ethnic factors. Lung cancer subjects had a heavier smoking history than controls.

After adjusting for a history of chronic bronchitis or emphysema, lung cancer subjects were more likely to have IgA titers ≥16 (55.4% vs. 51.3%) and ≥256 (5.1% vs. 2.5%) to C. pneumoniae than controls. Individuals with antibody tiers IgA ≥16 had 1.2 times the risk of lung cancer (95% confidence interval, 0.9–1.6) compared to those with lower titers. Investigators reported a significant trend (P = 0.007) of increasing odds ratios with increasing IgA titers primarily due to an odds ratio of 2.8 (95% confidence interval, 1.1–6.7) associated with titers ≥256. Elevated IgA was reported with squamous cell carcinomas and to a lesser extent, for small cell carcinomas and adenocarcinomas. There was no evidence of a stronger association with elevated IgG titers however. Subjects with race not classified as White or Black were more likely to have IgA titers ≥16. No significant differences in seropositivity were found, however, based on smoking behaviors.

Streptococcus bovis and colorectal cancer

Colorectal cancer (CRC) is a common malignancy in developed countries and is the 3rd most common cancer in the United States [38]. Greater than 80% occur sporadically [43]. The American Cancer Society estimates that there will be about 104,950 new cases of colon cancer and 40,340 new cases of rectal cancer in 2005 in the United States. Combined, they will cause about 56,290 deaths. The risk of colon cancer increases after the age of 40 and rises exponentially from the ages of 50 to 55. In fact, more than 9 out of 10 people found to have colorectal cancer are older than 50 [38].

Survival of CRC is related to the stage of disease at the time of the initial diagnosis. Between 1985 and 1997, death rates of colon cancer in the United States declined slightly due to earlier detection of primary tumors, via stool blood tests, sigmoidoscopy, colonoscopy, and screening tests for serum carcinoembryonic antigen concentration (CEA) [44]. The 5-year survival rate for CRC patients is greater than 90% when tumors are detected at a localized early stage. After the cancer has spread regionally and involves adjacent organs or lymph nodes, the rate drops to 40–65%; survival is less than 10% for patients with distant metastases. Therefore, there is an urgent need to develop effective treatment strategies to reduce morbidity and mortality. Surgery is currently the primary treatment modality for this disease. By the time the patient presents with recurrent symptoms, however, the disease is rarely curable by surgery even when combined with other therapies [45].

Several species of bacteria have been linked to chronic infections of the colon and increased risk of colon cancer including Escherichia coli [46] and several streptococci [47, 48]. Recent studies, however, have validated earlier findings of an association between colon cancer and Streptococcus bovis [11, 12]. As early as 1951, McCoy and Mason [49] suggested a relationship between colonic carcinoma and the presence of infectious endocarditis. It was not until 1974 [50] that the association of Streptococcus bovis and colonic neoplasia was recognized, as 25–80% of patients who presented with a S. bovis bacteremia had a colorectal tumor. The incidence of S. bovis associated colon cancer has been determined as 18% to 62% [14].

S. bovis is a normal inhabitant of the human gastrointestinal tract that can cause bacteremia, endocarditis, and urinary infection [51]. Although S. bovis is the 2^nd greatest cause of infectious endocarditis from streptococci [50], it is frequently associated with gastrointestinal lesions, especially carcinoma of the colon [12, 51–53]. Notably, the colonic neoplasia may arise years after the presentation of the condition of bacteremia or infectious endocarditis [12].

A retrospective review of forty-five documented cases of S bovis bacteremia was conducted by Gold et al. [12]. Subjects were identified by a search of computerized bacteriology records from one tertiary referral hospital and 1 community hospital located in the same city. Patient records were reviewed to identify the presence of colonic neoplasia, the use of gastrointestinal endoscopy, and the presence of gastrointestinal or extraintestinal malignancies. Seventeen patients (41% of adult patients) underwent colonoscopy. Colonic neoplasia was present in 16 patients (39% of adults). Invasive cancer was present in 13 patients (32% of adults), 8 of these had malignant lesions arising within the gastrointestinal tract, 3 affecting the colon and 5 patients had extraintestinal malignancies. The authors concluded that S. bovis bacteremia was associated with both colonic neoplasia and extracolonic malignancy.

It has been demonstrated that S. bovis or its wall extracted antigens (WEA) were able to promote carcinogenesis in rats [12]. In one of these investigations a total of 10 adult rats received i.p. injections of the carcinogen azoxymethane (AOM) (15 mg/kg body weight) once per week for 2 weeks. Fifteen days after the last injection of AOM (week 4) the rats were randomly divided into three groups. Twice per week during 5 weeks, the rats received, by gavage either S. bovis (10¹⁰ bacteria Group I), WEA (100 μg Group II) and controls (Group III).

One week after the last gavage (week 10), they found that administration of either S. bovis or its antigens promoted the progression of preneoplastic lesions. There were increased formations of hyperproliferative aberrant colonic crypts, enhanced expression of proliferation markers and increased production of IL-8 in the colonic mucosa. Normal rats treated with the bacteria did not develop hyperplastic colonic crypts, however. The authors concluded that S. bovis exerts its pathological activity in the colonic mucosa only when preneoplastic lesions are established.

Under identical experimental conditions Streptococcus gordonii was substituted for S. bovis. The number of preneoplastic lesions in the colon of S. gordonii-treated rats was similar to rats treated with AOM alone (22 ± 2). The authors suggested that S. bovis and its wall extracted antigens, unlike S. gordonii, act as promoters of carcinogenesis in a chemically-induced animal model.

In another investigation Biarc et al. [11] isolated 12 S. bovis cell-associated proteins (S300) and WEA. Cells of the human colonic epithelial cell line Caco-2 originally derived from an adenocarcinoma were grown to confluence and allowed to differentiate. These cells were stimulated with 200 ul of either S. bovis WEA (50 μg/ml) or cell-associated proteins S300 (100 μl).

The purified S300 fraction was able to trigger the human cell line and rat colonic mucosa to release chemokines (human IL-8 or rat CINC/GRO) and prostaglandin E₂ (PgE₂). The 12 S. bovis proteins were highly effective in the promotion of pre-neoplastic lesions in azoxymethane treated rats. In fact the S300 proteins were able to induce a 5-fold increase in PGE₂ secretion from Caco-2 cells, as compared with cells stimulated with WEA. The study found that PGE₂ release in the human cells correlated with an over-expression of cyclooxygease-2 (COX-2).

Evidence has shown that over-expression of COX-2 has a major role in mucosal inflammation [47] and is associated with inhibition of apoptosis [54] and enhancement of angiogenesis [55], which favor cancer initiation and development. It was reported by Biarc et al. [11] that S. bovis proteins also promoted cell proliferation by triggering mitogen-activated protein kinases (MAPKs), which can increase the incidence of cell transformation, the rate of genetic mutations and up-regulate COX-2. The investigators concluded that colonic bacteria such as S. bovis can contribute to cancer development particularly in chronic infection/inflammation diseases where bacterial components may interfere with cell function [11].

E. coli, crohn's disease and colon cancer

Inflammatory bowel disease (IBD) includes both ulcerative colitis (UC) and Crohn's disease (CD). Both of these disorders have an increased risk of colorectal cancer (CRC) [38, 46, 56]. Although colorectal cancer (CRC) in individuals with IBD only accounts for 1–2% of all cases of CRC in the general population, it is considered a serious complication of the disease and accounts for approximately 15% of all deaths in patients with IBD. The magnitude of the risk has been found to differ, however, in population-based studies [56–58]. Recent figures suggest that the risk of colon cancer for people with IBD increases by 0.5–1.0% yearly, 8–10 years after diagnosis. The magnitude of CRC risk also increases with early age at IBD diagnosis, longer duration of symptoms, and extent of disease, with pancolitis having more severe inflammation and a higher risk of dysplasia-carcinoma progression [56].

E. coli are found at higher levels in inflammatory bowel disease (IBD), therefore, studies have examined the mechanisms that may explain this phenomenon. A cell culture study by Martin et al [46] attempted to quantify and characterize mucosa-associated and intramucosal bacteria, particularly E. coli, in these inflammatory conditions. Their hypothesis was that the disease-associated alterations in mucosal glycosylation found in inflammatory bowel disease and colon cancer might predispose to altered recruitment of bacteria to the mucosa.

Mucosa-associated bacteria were isolated from biopsy samples of Crohn's disease, (n = 14); ulcerative colitis, (n = 21); noninflamed controls, (n = 24) and at surgical resection of colon cancer, (n = 21). Results found that mucosa-associated and intramucosal bacteria were cultured more commonly in Crohn's disease (79%, P = 0.03; and 71%, P < 0.01, respectively), and colon cancers (71% and 57%) than in noninflamed controls (42% and 29%) but not ulcerative colitis (38% and 48%). Mucosa-associated E. coli, which accounted for 53% of isolates, were more common in Crohn's disease (6/14; 43%) than in noninflamed controls (4/24, 17%), and intramucosal E. coli more common in Crohn's disease (29%; controls, 9%).

E. coli expressed hemagglutinins in 39% of Crohn's cases and 38% of cancers but only 4% of controls, and this correlated (P = 0.01) with adherence to embryonic intestinal cells (I407) and colon adenocarcinoma cells (HT29). Although close apposition of E. coli resulted in release of pro-inflammatory cytokines, cellular invasion by bacteria was not essential to this process [46].

Aspinell [59] suggested that the bacterial adherence found by Martin et al. [46] might result from activation of virulence genes following contact of the organisms with the inflamed mucosal cells. Martin et al. [46] found, however, that the mucosal isolates expressed of none of the known virulence genes, other than adherence genes. Martin and co-workers concluded that their findings supported a central role for mucosally adherent bacteria in the pathogenesis of Crohn's disease. They postulated that similar, lower grade, inflammatory changes could contribute to the risk of sporadic cancer development [46].

The authors stated however that it was certainly possible that the presence of the bacteria in the sub-mucus niche in human Crohn's disease and colon cancer could have been encouraged by disease-associated changes [46] in the mucosa. If true, their findings would result from colonization coincidental to the disease-associated alterations in mucosal glycosylation found in inflammatory bowel disease and colon cancer.

A study conducted by Masseret et al.[60] examined the E. coli strains isolated from patients with Crohn's disease (CD) with chronic ileal lesions (n = 14), early endoscopic recurrent lesions (n = 20), without endoscopic recurrence (n = 7), and controls (n = 21). Genetically linked E coli strains were isolated significantly more frequently from patients with chronic and recurrent CD (24/33 patients) than from controls (9/21) (p < 0.05). Most patients operated on for chronic ileal lesions (78.5%) harbored E coli strains belonging to the same cluster (p < 0.002 v controls). The prevalence of patients with early recurrent lesions harboring E coli strains belonging to this cluster was high but not significant. 21 of 26 strains isolated from patients with active CD demonstrated adherent ability to differentiated Caco-2 cells, indicating that most of the genetically related strains shared a common virulence trait. Comparison of E coli strains recovered from ulcerated and healthy mucosa of patients operated on for CD demonstrated in each patient that a single strain colonized the intestinal mucosa. The authors suggested that although a single E coli isolate was not found in Crohn's ileal mucosa, some genotypes were more likely than others to be associated with chronic or early recurrent ileal lesions.

S. typhi and susceptible populations

As previously stated, certain populations have an increased risk of gallbladder cancer (GC), however certain individuals may be predisposed to S. typhi infection which appears to increase the risk of GC. In an investigation by deJong et al. [61], three unrelated individuals with severe, idiopathic mycobacterial and Salmonella infections were found to lack IL-12Rβ1 chain expression. Interleukin-12 (IL-12) is a cytokine that promotes cell-mediated immunity to intracellular pathogens, such as S. typhi, by inducing type 1 helper T cell (T_H1) responses and interferon-γ (IFN-γ) production. IL-12 binds to the high-affinity β₁/β₂ heterodimeric IL-12 receptor (IL-12R) complexes on T cell and natural killer cells. The cells of these patients were deficient in IL-12R signaling and IFN-γ production and their remaining T cell responses were independent of endogenous IL-12. IL-12Rβ₁ sequence analysis revealed genetic mutations that resulted in premature stop codons in the extracellular domain. The genetic absence of IL12-Rβ₁ expression represented an immune deficiency in these 3 patients. Interestingly, these patients did not develop any abnormal infections with other viral, bacterial, or fungal pathogens. The defect in IFN-production and extreme susceptibility to mycobacterial and Salmonella infections in these patients appeared to be a direct result of their lack of IL-12R expression and signaling. The authors concluded that selective susceptibility to mycobacterial and Salmonella infections, however, suggested that the type-1 cytokine pathway was essential for controlling resistance to the intracellular pathogens and that no redundant protective immune mechanism could compensate for this deficiency.

Respiratory conditions and increased susceptibility to lung cancer

75–90% of people who develop lung cancer are smokers, however, only a small proportion of smokers develop lung cancer [42]. Hence, epidemiological studies such as that of Littman et al [42] and Kocazeybek et al. [19] have been conducted to more closely identify risk factors. Identifying genetic factors that increase a smoker's risk of developing lung cancer may help scientists to better understand the etiology of lung cancer and more effectively target high-risk groups for screening. Additionally, genetic factors have been identified that appear to predict the prognosis of certain lung cancer patients [62]. For example, mutations affecting the epidermal growth factor receptor (EGFR) were significantly associated with specific genetic alterations. Supervised clustering analysis based on EGFR gene mutations elucidated a subgroup including all EGFR gene mutated tumors, which showed significantly shorter disease-free survival

To analyze the genetic alterations of primary lung adenocarcinoma in a high-throughput way, Shibata et al. [62] used laser-capture micro-dissection of cancer cells and array comparative genomic hybridization focusing on 800 chromosomal loci containing cancer-related genes. They identified a large number of chromosomal numerical alterations, including frequent amplifications. Three subgroups of lung adenocarcinoma were characterized by distinct genetic alterations and were associated with smoking history and gender. The authors concluded that multiple carcinogenic pathways exist; certain abnormalities appear related to gender and smoking while others may impact survival [62].

Tumors and coley's toxins

Spontaneous tumor regression has followed severe bacterial, fungal, viral and protozoal infections. For hundreds of years this phenomenon inspired the development of the earliest cancer therapies. Reports of spontaneous remissions of advanced cancers infections can be found in the late nineteenth and early twentieth centuries. Many of these unexplained cures followed bacterial infections accompanied by high fevers.

An American surgeon, Dr. William Coley began the first well-documented use of bacteria and their toxins to treat end stage cancers. Coley first used live Streptococcus pyogenes cultures. Problems with the predictability of patient responses caused him to develop a safer vaccine in the late 1800's composed of two killed bacterial species, S. pyogenes and Serratia marcescens. In this way he could simulate an infection with the accompanying fever without the risk of an actual infection [95, 96].

Coley's vaccine was widely used to successfully treat sarcomas, carcinomas, lymphomas, melanomas and myelomas. Complete, prolonged regression of advanced malignancy was documented in many cases. The combined reports of Coley and others estimated the 5-year survival rate at 80% in malignancies for which no treatment existed. Even in patients considered in the terminal stages of cancer some remarkable recoveries were reported with the patient often outliving the cancer [97].

Coley considered 4 points critical to success: (1) initiation of a naturally occurring infection with fever, (2) avoidance of immune tolerance by gradually increasing the dosage, (3) directly injecting the vaccine into the tumor when accessible, and (4) a minimum of 6 months of injections to avoid recurrences. Today little credence is given to the febrile response in fighting cancer [96, 97].

A retrospective study was conducted in 1999 to compare the 10 year survival rate of patients treated by Coley's vaccine with modern conventional therapies. Richardson et al. [95] tried to match 128 of Coley's cases with 1,675 controls from the Surveillance Epidemiology End Result (SEER) cancer registry. The 2 populations were matched by age, gender, ethnicity, stage and radiation treatment status. Limitations included sample size and staging of patients receiving Coley's vaccine. The authors concluded that "Given the tremendous advances in surgical techniques and medicine in general, any cohort of modern patients should be expected to fare better than patients treated 50 or more years ago. Yet no such statistical advantage for the modern group was observed in this study." These findings were supported by case reports of spontaneous remissions or significant benefits when accidental infections occurred [98–100].

What role may a febrile response play in the remission of a tumor? Hobohm [101] offers the following hypothesis. Fever causes a cascade of events of inflammatory factors which activate resting dendritic cells (DC) that lead to the activation of T-cells. Cancer-cell specific T-cells usually remain in a state of anergy, most likely due to the absence of danger signals that usually accompany tissue destruction and inflammation upon acute infection [102]. A feverish bacterial infection may have a 3-fold beneficial effect. First, many infectious agents release endotoxins, like LPS, induce inflammatory cytokines and stimulate DC. Second, both thymocyte proliferation and generation of allo-specific CTL are increased with temperature in vitro [103]. Third, the vasculature of a tumor is more fragile than that of normal tissues and therefore more prone to destruction by the immune response. An infection causing hemorrhagic necrosis could trigger febrile collapse of the tumor vasculature [104, 105]. Interestingly, the affinity of certain streptococci for binding to fibrinogen and fibrin may account for the 'homing' of bacterial enzymes to tumors as these cells are abundant in such proteins [106].

The mechanism by which infection cures cancer has been investigated. It has been suggested by Zacharski and Sukhatme [96] that the tumor regression observed by Coley and others is due to the activation of plasminogen. For example when the streptococcal spreading factor known as streptokinase (SK) combines with host plasminogen, plasmin is released. Plasmin triggers protease cascades that degrade plasma and extracellular matrix proteins. These mechanisms of degradation are toxic to tumor cells, disrupt the tumor extracellular matrix, alter tumor growth and inhibit metastasis [96]. The notion that plasminogen activators like SK might result in the remissions reported by Coley is appealing as they appear to spare healthy cells while attacking tumors. Zacharski et al. [107] hypothesized that although the potent enzymes produced by plasminogen activation may have a direct effect on cancer cells it was more likely they disrupted the cell-extracellular matrix of the tumor.

Investigators report that the traditional best treatment options for some candidate tumor types, such as advanced soft tissue sarcomas, breast cancer and melanoma, have not improved patient outcome substantially since Coley's day [96, 108, 109]. Currently, biologic response modifier therapies have moved beyond the nonspecific immunotherapy of Coley's era and laid the foundation for today's approaches. Zarcharski and Sukhatme [96] suggest that the early success of Coley's toxins are leading to therapies that engage the host's immune system against an individual's tumor offering new hope for cancer patients.

Autologous tumor cell vaccine therapy is an example of this new approach to cancer treatment. These vaccines differ markedly from conventional cytotoxic drug therapy that affect both normal and tumor cells. Tumor vaccines stimulate an individual's cell-mediated immune response by targeting the patient's tumor antigens. While efficacy of standard chemotherapy relates to the dose of the drug, the efficacy of a tumor vaccine is more complex, involving host-vaccine interactions [110]. These include: (1) immunogenicity of the vaccine regarding tumor-associated antigens as opposed to self; (2) the host's immune response in terms of immune recognition and effector mechanisms; and (3) the development of host systemic cell-mediated immunity, including long-term immunologic memory, (3–5 years). Therefore, the potency of the vaccine is not determined by immunogenicity alone but by its ability to induce the host anti-tumor response [110].

Bacillus calmette-guérin and autologous tumor cell vaccine vs. colon cancer

Certain tumor antigens are, however, normally weak immunogens. Therefore the use of adjuvants and the intradermal route of injection have, in some cases, produced an optimum antigenic vaccine. These adjuvant vaccines have induced effective host recognition of tumor-associated antigens and improved patient survival. For example, preliminary evidence by Hoover et al [111] suggested that active specific immunotherapy (ASI) of colon cancer using autologous tumor cell vaccines had potential in improving recurrence-free interval and survival. ASI assumes there are distinct tumor antigens on an individual's cancer cells that are either absent or in lower concentration on normal cells. The vaccine attempted to stimulate host's immune defenses against tumor-associated antigens by enhancing the immunogenicity of the patient's own tumor cells with an immunomodulating adjuvant, such as Bacillus Calmette-Guérin (BCG).

In a study by Hoover et al. [111] 80 eligible subjects with colon (47) or rectal (33) cancer were enrolled into a prospectively randomized, controlled clinical trial of active specific immunotherapy (ASI). An autologous tumor cell-Bacillus Calmette- Guerin (BCG) vaccine was used to determine whether ASI could improve disease- free status and survival. Eligible subjects had colon or rectal cancers extending through the bowel wall or had positive lymph nodes providing adequate cells from the primary tumor. Wide surgical removal of all tumors was performed with histologically proven clear margins and removal of involved lymph nodes. Prior to randomization individuals were screened for metastatic disease. Colon cancer and rectal cancer subjects were in separate but parallel studies and randomized into groups treated by resection alone or resection plus ASI. 3–4 weeks following surgery, both controls and treatment subjects were skin tested for immune competence and sensitivity to tuberculin purified protein derivative (PPD). Vaccines were begun in the ASI treatment group 4–5 weeks following surgery to allow for adequate immune recovery from surgery and anesthesia. A total of 24 colon and 17 rectal subjects composed the treatment group. With a median follow-up of 93 months, there was a significant improvement in survival (two-sided P = .02; hazards ratio, 3.97) and disease-free survival (two-sided P = .039; hazards ratio, 2.67) in all eligible colon cancer patients who received ASI. With a median follow-up of 58 months, no benefits were seen in patients with rectal cancer who received ASI. The authors concluded that the study suggested that ASI may be beneficial to patients with colon cancer.

In 2005, Uyl-de Groot et al. [110] conducted a multicenter, randomized controlled phase III clinical trial with Stage II and III colon cancer patients using ASI. Autologous tumor cells were used with the immunomodulating adjuvant Bacillus Calmette-Guérin (BCG) in a vaccine (OncoVAX^®). Patients were randomized to receive either OncoVAX^® or no therapy after surgical resection of the primary tumor. The vaccine was processed within 48 h after surgery in order to have viable, metabolically active, autologous tumor cells.

Analysis of prognostic benefit with a 5.8 year median follow-up, showed that the beneficial effects of OncoVAX^® were statistically significant at all endpoints including recurrence-free interval, overall survival, and recurrence-free survival in Stage II colon cancer patients. Surgery alone cures 65% of Stage II colon cancer patients. For the remaining patients, OncoVAX^® in an adjuvant setting significantly prolongs recurrence-free interval and significantly improves 5-year overall survival and recurrence-free survival. Unfortunately, no statistically significant prognostic benefits were achieved in Stage III patients [110].

Immunization with bacillus calmette-guérin vs. lung cancer

Grant et al [39] hypothesized that optimal chemotherapy with or without radiation followed by active immunization could eliminate microscopic residual disease and prolong survival. Immunization with GD3, a ganglioside expressed on the surface of most small cell lung cancers (SCLC) had not evoked a strong immune response. Therefore BEC2, a large xenogenic protein which mimics GD3, was judged to be a good immunogenic candidate. This approach had proven successful in extending the lives of melanoma patients [112].

Chapman et al. [113] conducted a phase II trial comparing 5 dose levels of BEC2. The study population consisted of 15 patients with small cell lung cancer (SCLC). All subjects had completed standard therapy and had achieved a partial or complete response. Patients received a series of five intradermal immunizations consisting of 2.5 mg of BEC2 plus BCG over a 10-week period. Blood was collected for serological analysis, and outcome was monitored. All patients developed anti-BEC2 antibodies, despite having received chemotherapy with or without thoracic radiation. Anti-GD3 antibodies were detected in five patients, including those with the longest relapse-free survival. The median relapse-free survival for patients with extensive stage disease was 10.6 months. In patients with limited stage disease a median relapse-free survival had not been reached with a follow-up of >47 months and only one of the 7 patients in this group relapsed. The authors reported that immunization of SCLC patients using BEC2 plus BCG after standard therapy could induce anti-GD3 antibodies and was safe. The survival and relapse-free survival in this group of patients was substantially better than those observed in similar patients receiving standard therapy.

A Phase III trial was conducted to evaluate BEC2 plus BCG as adjuvant therapy for limited small-cell carcinoma after chemotherapy and irradiation [114]. A total of 515 subjects were randomly assigned. Unfortunately, in this trial there was no improvement in survival, progression-free survival, or quality of life in subjects that were vaccinated. A trend toward prolonged survival was observed in the one third of subjects who developed a humoral response (p = 0.085), however.

The effectiveness of vaccines for several cancers was examined in a series of investigations. Xiang et al. [115] tested vaccines for human melanoma using the mutant S. typhi strain SL7207 as a DNA carrier. Tolerance against self-antigens was broken by genetically fusing ubiquitin with MHC I derivatives. Another approach coupled tumor-specific antibodies to functional IL-2. This combination in addition to oral vaccination with plasmid-encoded tumor antigens significantly enhanced protection against carcinoma of the colon [116], carcinoma of the lung [117] and melanoma [118]. Unstable cancer cells provided a challenge, however. Interestingly, this was overcome by targeting stable, proliferating endothelial cells of the tumor vasculature. This novel approach effectively inhibited angiogenesis [119].

Helicobacter pylori and esophageal adenocarcinoma

In industrialized countries the incidence of H pylori has been steadily decreasing [120]. The incidence of esophageal cancer (EA), however, is increasing [121]. Surprisingly, there is evidence that these two trends may be related. Several studies have determined that virulent strains of H pylori are found less commonly among patients with Barrett's esophagus and EA when compared with controls [122–124]. This led to studies that found positive associations among the increased incidence of obesity, GORD, Barrett's esophagus and EA [125].

Recently, a nested case-control study was conducted by de Martel et al [126] to assess the association between H. pylori infection and the risk of development of EA. Of a total of 128,992 members of an integrated health care system who had participated in a multiphasic health checkup (MHC) during 1964–1969, 52 patients developed EA during follow-up. Three randomly chosen control subjects from the MHC cohort were matched to each cancer subject, on the basis of age, gender, race, date and site of the MHC. Data on cigarette smoking, alcohol consumption, body mass index (BMI), and education level were obtained. Serum samples collected at the MHC were tested for IgG antibodies to H. pylori and to the H. pylori CagA antigen associated with H. pylori virulence.

Subjects with H. pylori infections were less likely than uninfected subjects to develop EA odds ratio (OR, 0.37) 95% confidence interval (CI, 0.16–0.88). This significant association was restricted to cancer subjects and control subjects <50 years old (OR, 0.20) (95% CI, 0.06–0.68). Interestingly, in patients with H. pylori infections, the OR for EA in those who tested positive for IgG antibodies to the CagA protein was similar to that for those who tested negative for it. BMI ≥25 and cigarette smoking, however, were strong independent risk factors for EA. The authors found, however, that the absence of H. pylori infection, independent of cigarette smoking and BMI, was associated with an increase in the risk of development of EA [126].

An epidemiological study in Sweden sought to determine whether BMI was associated with esophageal malignancies compared to gastric adenocarcinoma and controls. In a nationwide, population-based case-control study by Lagergren et al [127], between 1995 through 1997, a total of 189 patients with adenocarcinoma of the esophagus and 262 patients with adenocarcinoma of the gastric cardia were enrolled. These patients were compared with 167 patients with incident esophageal squamous cell carcinoma and 820 healthy controls.

Odds ratios were determined from BMI and cancer case-control status and ratios estimated the relative risk for the two adenocarcinomas studied. Calculations used multivariate logistic regression with adjustment for potential confounding factors. The adjusted odds ratio was 7.6 (95% CI, 3.8 to 15.2) among persons in the highest BMI quartile compared with persons in the lowest. Obese persons (persons with a BMI>30 kg/m²) had an odds ratio of 16.2 (CI, 6.3 to 41.4) compared with the leanest persons (persons with a BMI<22 kg/m²). The odds ratio for patients with cardia adenocarcinoma was 2.3 (CI, 1.5 to 3.6) in those in the highest BMI quartile compared with those in the lowest BMI quartile and 4.3 (CI, 2.1 to 8.7) among obese persons. Although a strong dose-dependent relation existed between BMI and esophageal adenocarcinoma, esophageal squamous-cell carcinoma was not associated with BMI. A modest but significant increase in intragastric acidity was also observed following the cure of H pylori infection which the authors postulated could contribute to gastroesophageal reflux disease (GORD).

The incidence of EA has increased rapidly over the last 30 years. During this period, the prevalence of Helicobacter pylori has decreased. Trends of increasing esophageal adenocarcinoma can be linked causally to increasing GORD which can be linked to an increasingly obese population. There appeared to be no plausible biological mechanism of association between H pylori, obesity, and GORD until studies of ghrelin, however.

Ghrelin was the first circulating hormone demonstrated to stimulate food intake in man. This peptide is produced in the stomach and regulates appetite, food intake, and body composition. The effects of ghrelin were examined in H pylori positive asymptomatic subjects by several investigators [128–130]. In a randomized double-blind cross-over study, by Wren et al. [129], ghrelin was shown to acutely enhance appetite and increase food intake in 9 healthy human subjects. There was a clear-cut increase in calories consumed by every individual from a free-choice buffet (mean increase 28 +/- 3.9%, p < 0.001) during ghrelin versus saline infusions. Furthermore, visual analogue scores for appetite were greater during ghrelin compared to saline infusion. Ghrelin had no effect on gastric emptying, however. The authors concluded that endogenous ghrelin was a potentially important new regulator of the complex systems controlling food intake and body weight.

Evidence is accumulating that ghrelin may explain the relative rarity of H. pylori among patients with Barrett's esophagus and EA. Findings from these studies and others support the notion that H. pylori may have a "protective" effect against EA [122, 124]. Studies have found that curing H pylori infection increased plasma ghrelin in healthy asymptomatic subjects which may lead to increased appetite, weight gain and contribute to the increasing obesity seen in Western populations where the prevalence of H pylori is low. This evidence supports the notion that decreasing incidence of H pylori infection may lead to increased levels of plasma ghrelin and that this hormone appears to be a factor in increasing obesity which elevates the risk of GORD which is positively associated with Barrett's esophagus and increased risk of esophageal adenocarcinoma. It appears that the absence of H. pylori infection may be one of several factors that leads to the increased incidence in EA effect observed in Western populations.

The implications for treatment of individuals with H. pylori infection were addressed by Nakajima and Hattori [131]. They systematically reviewed the literature and estimated the expected annual incidence of esophageal or gastric cancer with and without eradication of H. pylori infection in patients with chronic atrophic gastritis. The expected annual incidence of gastric cancer in patients with corpus atrophy with persistent infection was at least 5.8-fold higher than that for esophageal adenocarcinoma after the eradication of infection at all ages. Even for patients with accompanying reflux esophagitis or Barrett's esophagus, the incidence of gastric adenocarcinoma with persistent infection was higher than that of esophageal adenocarcinoma after eradication of infection. The authors concluded, therefore, that if eradication of infection lowers the incidence of gastric cancer, it should be recommended for patients with corpus atrophy at all ages irrespective of the presence of reflux esophagitis or Barrett's esophagus, especially in populations having a high prevalence of gastric cancer [131].

In summary, increased BMI has been linked with the elimination of H. pylori infection. As the sphincter mechanism at the esophagogastric junction is weakened by weight it is not surprising that obese individuals have a higher incidence of gastric reflux or GORD [127]. GORD may lead to the development of Barrett's esophagus, which increases the risk of EA by 40-fold [132, 133]. The study by Lagergren [127] provides evidence that these associations may be related as increasing body mass was associated with a stepwise increase in the risk of EA. If eradication of H. pylori infection lowers the incidence of gastric cancer, however, it should be recommended for patients with corpus atrophy at all ages irrespective of the presence of reflux esophagitis or Barrett's esophagus, especially in populations having a high prevalence of gastric cancer [131].

Attenuated bacteria: Promising carriers of DNA vaccines

Attenuated bacteria will enhance stimulation of the innate immune system yet increase the safety of a vaccine, [134] therefore they may be ideal for the delivery of vaccines. Animal studies have shown that attenuated S. typhimurium strains can successfully deliver a variety of genetically engineered DNA vaccine plasmids for therapeutic vaccination of mice against model tumors [117, 118, 135].

The identification of bacterial "carriers" for DNA vaccines that target cancer cells by site-specific colonization may allow the selective delivery of vaccine plasmids into tumor cells [136]. Colonization of these species may be considered coincidental to favorable conditions provided by the tumor yet prove clinically useful. Ultimately, however, the safety and efficacy of recombinant therapeutic agents expressed by plasmids must be conducted in appropriate animal models.