Chemoprevention in Barrett’s esophagus is currently applied only in research settings. Identifying pathways that can be targeted by safe, pharmaceutical or natural compounds is key to expanding the scope of chemoprevention. Defining meaningful surrogate markers of cancer progression is critical to test the efficacy of chemopreventive approaches. Combinatorial chemoprevention that targets multiple components of the same pathway or parallel pathways could reduce the risk and improve the efficacy of chemoprevention. Here we discuss the role of chemoprevention as an independent or an adjuvant management option in BE-associated esophageal adenocarcinoma.
Key points
- •
At present, chemoprevention in Barrett’s esophagus is applied only in research settings.
- •
Identifying pathways that can be targeted by safe, pharmaceutical or natural compounds is key to expanding the scope of chemoprevention.
- •
Defining meaningful surrogate markers of cancer progression is critical to test the efficacy of chemopreventive approaches.
- •
Chemopreventive agents that show promise across epidemiologic, preclinical (in vitro and in vivo), and early clinical trials will be more likely to be successful when applied in clinical practice.
- •
Combinatorial chemoprevention that targets multiple components of the same pathway or parallel pathways could reduce the risk and improve the efficacy of chemoprevention.
Introduction
Barrett’s esophagus (BE) is a premalignant condition in which chronic reflux injury leads to replacement of normal squamous mucosa by metaplastic columnar mucosa. Approximately 1 in 10 to 20 adults with reflux in the United States harbor BE, and based on currently available data, the annual incidence of adenocarcinoma in patients with BE is estimated to range from 0.3% to 0.5%. Esophageal adenocarcinoma (EAC) is becoming increasingly common in Western countries over the past several decades. Although the incidences of most malignances are trending downward, the rate of EAC is rapidly increasing. The 5-year survival rate for patients with EAC is approximately 17%, and this dismal prognosis has not changed. Although strategies involving early detection of cancer at a potentially curable stage make sense, these approaches are inherently expensive and are based on limited data. Alternatively, novel interventions, such as ablation or resection of dysplastic Barrett’s mucosa, are being increasingly used in clinical practice; recurrent metaplasia and neoplasia remain problematic. Here we discuss the role of chemoprevention as an independent or an adjuvant management option in BE-associated EAC.
Introduction
Barrett’s esophagus (BE) is a premalignant condition in which chronic reflux injury leads to replacement of normal squamous mucosa by metaplastic columnar mucosa. Approximately 1 in 10 to 20 adults with reflux in the United States harbor BE, and based on currently available data, the annual incidence of adenocarcinoma in patients with BE is estimated to range from 0.3% to 0.5%. Esophageal adenocarcinoma (EAC) is becoming increasingly common in Western countries over the past several decades. Although the incidences of most malignances are trending downward, the rate of EAC is rapidly increasing. The 5-year survival rate for patients with EAC is approximately 17%, and this dismal prognosis has not changed. Although strategies involving early detection of cancer at a potentially curable stage make sense, these approaches are inherently expensive and are based on limited data. Alternatively, novel interventions, such as ablation or resection of dysplastic Barrett’s mucosa, are being increasingly used in clinical practice; recurrent metaplasia and neoplasia remain problematic. Here we discuss the role of chemoprevention as an independent or an adjuvant management option in BE-associated EAC.
Definition and concept of chemoprevention
Chemoprevention refers to a strategy of using pharmaceutical or nutraceutical compounds to prevent the initiation or progression of dysplasia, as well as blocking the invasion of dysplastic epithelial cells across the basement membrane. To clinically apply this strategy, several factors need to be considered. Is there a cohort of patients with a premalignant lesion that can be easily detected or are there known clinical and or pathophysiological conditions that facilitate neoplastic progression? Are there modifiable molecular, biochemical, or cellular derangements that can be targeted using pharmaceutical or nutraceutical compounds, and if so are there intermediate or definitive endpoints that could assess the efficacy of these compounds? In the following discussion, we review the existing scientific evidence to examine the role of chemoprevention in BE.
Risk stratification of cohort
Unlike standard therapeutic paradigms in EAC, chemoprevention has to be undertaken in individuals with reflux or Barrett’s metaplasia in which the risk of lethal EAC is extremely variable depending on dysplasia grade. Therefore, the potential benefit of an intervention has to be carefully weighed against possible side effects or toxicity associated with the intervention. Additionally, existing comorbidities, patient preferences, and local expertise also have to factor into decision-making. For the purpose of chemoprevention, we discuss risk stratification in the following 3 cohorts.
Gastroesophageal Reflux
Gastroesophageal reflux (GERD) is a key risk factor for the development of BE as well as EAC. Patients with long-standing and severe reflux injury are at 6-fold to 8-fold higher risk of developing BE and 7-fold to 43-fold increased risk of EAC. These observations suggest that the GERD population could be a reasonable cohort for chemoprevention. A Gallup survey demonstrated that 44% of adults in the United States experience heartburn at least once a month. A survey of residents in Olmsted County in Minnesota demonstrated that 18% of respondents experienced heartburn at least weekly. Given that nearly one-fifth of the US population has significant reflux symptoms and there is a lack of reflux symptom correlation with the identifiable premalignant lesion of BE, it makes this population a viable chemoprevention cohort only for interventions that are essentially risk free.
Nondysplastic Barrett’s or Barrett’s with Low-Grade Dysplasia
Most the patients diagnosed with BE are nondysplastic or have low-grade dysplasia. Overall, the risk of neoplastic progression in nondysplastic BE is very low. In a Danish population-based, cohort study involving 11,028 patients with Barrett’s esophagus, the incidence rate for adenocarcinoma was 1.2 cases per 1000 person-years in nondysplastic Barrett’s and 5.1 cases per 1000 person-years in Barrett’s with low-grade dysplasia. Another cohort study from Ireland showed a progression rate of 0.17% per year in patients with nondysplastic BE. Similar to these studies, a recent meta-analysis showed the incidence of EAC in patients with nondysplastic Barrett’s epithelium to be 1 case per 300 patient years. Based on these results, if a chemopreventive intervention reduces the risk of cancer by 50%, one would need to treat 2000 nondysplastic or 400 patients with low-grade dysplasia to prevent 1 cancer death. Therefore, if the patient is relatively healthy, with such a low-risk of progression, it is reasonable to use chemopreventive agents that have been deemed extremely safe for use. In this patient population, even a moderate reduction in cancer risk by 50% would translate into a significant public health benefit, as most Barrett’s patients fall under this category.
Patients with Barrett’s with High-Grade Dysplasia and Postablative Barrett’s
Unlike nondysplastic Barrett’s or patients with low-grade dysplasia in whom the risk of cancer progression is low and it takes several years to progress to EAC, up to 20% patients with high-grade dysplastic Barrett’s already have a prevalent EAC. A subset of patients with ablated dysplastic Barrett’s are also at higher risk of recurrent dysplastic BE or EAC. Because of the relatively high risk of neoplastic transformation, it is important to use more effective chemopreventive agents in this cohort, even if these agents carry a considerably higher risk of adverse effects. Alternatively, in this cohort, it will be reasonable to consider chemoprevention as an adjunctive approach along with other therapy, such as endoscopic ablation.
Druggable targets
Similar to therapeutic interventions, target discovery is a critical step in chemoprevention. It is important to understand which pathways are deregulated during progression to EAC and if these pathways could be modifiable with the use of nutraceutical or pharmaceutical compounds. Identification of druggable targets largely depends on data derived from epidemiologic associations and through pathophysiological studies using cell culture and/or animal models to see if these associations could be directly implicated in neoplastic transformation. Because the process of carcinogenesis involves host-environment interactions at local and systemic levels, in the following paragraphs we discuss the target identification along these lines.
Luminal
Acid and bile salts
The esophageal mucosa is constantly exposed to luminal contents. In patients with reflux, gastric and duodenal secretions intermittently cover the esophageal mucosa. Bile salts contained in duodenal secretions in a gastric pH dependent manner play a key role in the chronic injury-inflammation-carcinogenesis cascade. There is significantly more gastric acid and duodenal bile reflux in patients with Barrett’s complicated by dysplasia or cancer compared with Barrett’s without dysplasia. Bile salts that reach the esophagus produce injury over a wide range of pH depending on their p K a. For example, when the reflux is acidic (pH <4), taurine-conjugated bile salts cause injury. When pH is from 4 to 6, glycine-conjugated bile salts cause injury, and unconjugated bile salts cause injury when the pH is neutral or alkaline. Interestingly, the levels of taurine-conjugated and unconjugated bile salts are significantly higher in the esophagus of patients with BE compared with patients with mild or minimal reflux. Even though proton pump inhibitors (PPIs) successfully heal the obvious esophagitis in 80% to 90% of patients, up to one-third of these patients continue to have bile reflux. In addition, chronic use of PPIs deconjugates bile salts by bacterial colonization of the proximal gut. These unconjugated bile salts, in the high-pH environment created by PPIs, are implicated in chronic low-grade inflammation and promote carcinogenesis in BE. Consistent with these epidemiologic associations, chronic reflux of gastric acid and/or duodenal contents in animal models leads to dysplasia and EAC. In canine and rodent models, bile reflux alone induces carcinogenesis in BE.
Human studies that simultaneously target acid production along with bile salt modulation, for example a combination of PPIs and ursodeoxycholic could result in prevention of EAC. Other potential targets can be mucosal protectants or agents that could reduce mucosal permeability. Because exposure of esophageal mucosa to acid and/or bile induces oxidative stress, activates proinflammatory pathways, and initiates aberrant repair process in the mucosa, several mechanisms discussed in the following sections identify additional chemoprevention strategies.
Carcinogens
There is only limited information available on the effects of carcinogen exposure in the development of EAC. However, a weak association exists between cigarette smoking and progression to adenocarcinoma. In a multivariate analysis, the Rotterdam Esophageal Tumor Study Group noted that compared with nonsmokers, the odds of EAC in smokers were 2.3 times higher. Compared with patients with BE, patients with EAC have higher median duration as well as the number of pack-years of smoking (median 29.5 vs 38.5 years, P <.003 and median 15.0 vs 55.25 pack-years, P <.001). Other investigators found no clear association between smoking and neoplastic progression in BE. The differences could simply be due to variable definition of Barrett’s, by including prevalent and incident cases together and inability to control for confounding factors because of small sample size. Preclinical animal studies support a link between carcinogen exposure and EAC. The risk of EAC in mice and rodents increases with 2,6-dimethylnitrosomorpholin or N-methyl-N-benzylnitrosamine carcinogen administration. This is clinically relevant because in the latter half of the past century there was marked increase in the nitrate levels in food and water. Nitrates are concentrated by the salivary glands and secreted into the mouth, where oral microbes reduce these nitrates to nitrites. Moreover, achlorhydria induced by PPI use leads to overgrowth of nitrate-reducing bacteria in the stomach, and these bacteria convert dietary nitrate to nitrite. Contents of gastroduodenal reflux in the lower esophagus change these nitrites to nitrous acid and eventually to nitric oxide. Nitric oxide, through DNA damage, lipid peroxidation, and mutagenesis, promotes tumorigenesis. Targeting carcinogen metabolism or downstream pathways can therefore provide an alternate strategy to prevent carcinogenesis. Similarly, taking steps toward cessation of smoking in high-risk individuals also may prevent neoplastic transformation.
Mucosal
The mucosal compartment of the esophagus allows complex interactions among epithelial, stromal, and immunoregulatory cells. Among the most studied molecular mechanisms that are dysfunctional during neoplastic transformation include prostaglandin biosynthesis, increased proinflammatory cytokine release, increased expression and binding of growth factor ligands to their receptors followed by upregulation of membrane to nuclear signaling, and polyamine synthesis. These derangements provide numerous potential chemoprevention targets.
Arachidonic acid pathway and prostaglandin synthesis
The arachidonic acid pathway is a key mediator of inflammatory response through prostaglandin synthesis, particularly PGE2. This pathway is upregulated during neoplastic progression in BE. PGE2 synthesis is regulated at various steps. The first step is catalyzed by phospholipases to release arachidonic acid from membrane phospholipids. Arachidonic acid is catalyzed by cyclooxygenase (COX) to synthesize PGE2. Among phospholipases, cytosolic PLA2α, because of its high selectivity for a particular type of unsaturated fatty acid, is considered the rate-limiting step for PGE2 synthesis. Transcription factor KLF11, an epigenetic regulator, inhibits PGE2 by silencing cPLA2α and thereby downregulates growth of epithelial cells in BE. In addition, in an in vitro study performed on Barrett’s cell lines by our laboratory, we found that COX-2 inhibition significantly decreases proliferation of BE cells, whereas treatment of these cells with PGE2 (product of COX-2 activity) restored proliferation of cell lines, suggesting a possible role of COX-2 inhibition for chemoprevention in BE. A preclinical study conducted by us established that the use of selective and nonselective COX-2 inhibitors in a rodent model of chronic reflux produced a statistically significant reduction in relative risk of developing EAC by 55% in rats treated with MF-tricyclic and 79% ( P <.01) in those treated with sulindac as compared with the control group. The degree of inflammation was found to be more severe in the control group compared with the study group. In summary, downregulation of PGE2 levels, either by KLF11-mediated inhibition of cPLA2α or through COX2 inhibition, reduces cell growth in vitro and neoplastic transformation in Barrett’s and mucosa of animals with reflux injury. These findings, along with epidemiologic evidence of lower risk of EAC in patients who chronically use nonsteroidal anti-inflammatory drugs (NSAIDs), suggest that intercepting prostaglandin biosynthesis through anti-inflammatory drugs or natural epigenetic modifiers could have chemopreventive potential.
Cytokines
In patients with BE, chronic mucosal injury by acid and bile reflux leads to inflammation and release of cytokines and chemokines. Through ligand-receptor binding, these factors trigger the activation of membrane to nuclear signaling that deregulate cell growth and differentiation and facilitate neoplastic progression. The mechanism involved here can be targeted to prevent neoplastic progression. Several recent preclinical and clinical studies show deregulation of the proinflammatory cytokine cascade that starts with upregulation of interleukin (IL)-1β and tumor necrosis factor (TNF)-α. In vitro and in vivo work support that IL-1β and TNF-α via the IL-6-STAT3 pathway upregulate oncogenic AKT and COX to drive PGE2 biosynthesis. During carcinogenesis in BE, there is upregulation of TNF-α ligand and receptors, as well as polymorphism of the IL-1β gene. Interestingly, transgenic expression of human IL-1β in the esophagus of mice with bile salt injury leads to abnormal differentiation of pluripotent stem cells, metaplasia, and progression to carcinoma via changes in the stromal microenvironment. Members of this cascade are targets of drugs such as NSAIDs, inhibitors of IL-1β, TNF-α, or STAT3. Therefore, interrupting this pathway at various levels could have chemopreventive potential.
Growth factors
Injury-repair is central to the process of carcinogenesis in BE and is associated with upregulation of many growth factors. If the cells express appropriate receptors, growth factors can facilitate cell growth. The ligand secreted by stromal cells or adjacent cells bind to receptors on epithelial cells modulating growth regulatory cytoplasmic and nuclear signaling as well as changing stromal epithelial interactions in a feed-forward manner.
Neoplastic progression in BE is associated with increased expression of epidermal growth factor (EGF), transforming growth factor α, and EGF receptor (EGFR), raising a potential for autocrine or paracrine signaling. Transition from Barrett’s to cancer is associated with overexpression of c-erb B-2 oncogene that translates into constitutively active EGFR that does not require EGF overexpression. Tumors expressing c-erb B-2 have poor outcomes. Because this is a delayed event during carcinogenesis, it therefore does not appear to be a useful chemoprevention target for primary prevention.
The typical growth factor signaling involves activation of downstream membrane to nuclear cascades through kinases that cause posttranslational changes in proteins. In many cancers, ras, src, and myc family of proteins play an important role in transformation. Interestingly, unlike many other cancers, K-ras does not appear to be relevant to neoplastic transformation in Barrett’s but H-ras, src, and myc all show progressive increase during carcinogenesis and could be targeted for chemoprevention.
Differentiation-related extracellular signaling ligands
The WNT glycoprotein family of extracellular signaling ligands is involved in cell growth, motility, and differentiation. WNT inhibitory factor 1 (WIF1), a WNT antagonist, is silenced by promoter hypermethylation, leading to an increase in cellular proliferation, a phenotype that can be rescued by WIF1 restoration. This suggests that agents that activate WIF1, due to their ability to inhibit WNT signaling, could have chemopreventive potential. Restoration of APC and secreted frizzled-related protein-1 (SFRP1) function by reversing promoter methylation could also inhibit WNT signaling and carcinogenesis.
Polyamines
Growth factors, TNF-α, nuclear factor (NF)-kB, and Myc, which are upregulated in neoplastic transformation, activate ornithine decarboxylase (ODC). ODC is the rate-limiting enzyme in polyamine synthesis. Polyamines (putrescine, spermidine, and spermines) have a key role in cell growth. Increased ODC activity and resulting polyamine synthesis is noted during neoplastic progression in BE. Therefore, ODC inhibition by difluoromethylornithine is a reasonable chemopreventive strategy.
Prosurvival and antideath pathways
NF-kB is a proinflammatory, prosurvival transcription factor. There is an upregulation of several members of the NF-kB pathway during cancer development in patients with Barrett’s. As discussed previously, TNF-α, an upstream activator of NF-kB, has been shown to be progressively upregulated in Barrett’s tissues, and IL-8, a downstream transcriptional target of NF-kB, also has been shown to be upregulated in Barrett’s-related adenocarcinomas. In vitro studies using Barrett’s epithelial cells show that physiologic concentrations of bile acid deoxycholic acid (DCA) at neutral pH activate NF-kB. Similar results also are noted in patients undergoing infusion of DCA into the esophagus. DCA induces oxidative stress in Barrett’s mucosa of patients, which causes genotoxic injury, and it also simultaneously induces activation of the NF-κB pathway, which enables cells with DNA damage to resist apoptosis. These molecular mechanisms point to the chemopreventive potential of NF-kB inhibitors, such as BAY 11-7085, AdIκB super repressor, Bortezomib, or curcumin during carcinogenesis in BE.
One of the important concepts during carcinogenesis is uncoupling of cell proliferation from cell death or apoptosis. Several events during carcinogenesis, such as inflammation, generation of reactive oxygen species, lack of antioxidant defenses, and mutagenic pressure exerted by nitroso-compounds, result in hyperproliferation of epithelial cells and mutagenesis. Entry of these mutated cells into further cell division is typically prevented by the tumor suppressor protein p53. Functional p53 induces apoptosis in these abnormal cells if DNA is repaired. Loss of p53 function is a common phenomenon in BE that compromises this protective mechanism. Clonal expansion of these cells with mutated DNA facilitates neoplastic progression in BE. Therefore, pharmacologic measures that upregulate p53 (eg, trans-retinoic acid) may result in chemoprevention by restoration of the apoptotic response. Similar to p53 loss, several other mechanisms help Barrett’s epithelial cells evade apoptosis. Bile acid DCA via Erk-1/2 and p38 MAPK-mediated activation of AP-1 transcription factor induces antiapoptotic protein COX-2. Likewise, an increase in the expression of constitutively active splice variant of the cholecyctokinin-2 receptor in Barrett’s mucosa also leads to activation of antiapoptotic pathways via PKB/Akt. These findings support targeting these kinases either individually or in combination to achieve chemoprevention.
Epigenetic signals
Chromosomal instability influences various genes that modulate the biological and biochemical processes involved in carcinogenesis. To an extent, these genetic changes are modifiable via epigenetic signals, and in certain situations, epigenetic signals actually facilitate the chromosomal instability. Global hypomethylation along with promoter-specific hypermethylation during carcinogenesis in BE leads to inactivation of tumor suppressor gene p16, thereby nullifying its tumor-suppressive action. Further, there is loss of p16 heterozygosity at 9p21 and inactivation of tumor suppressor p53. Loss of function of these key regulatory tumor-suppressive genes results in uncontrolled cell proliferation. Progressive hypermethylation and epigenetic silencing leading to loss of function of the glutathione peroxidase 3 gene also is observed in BE. This epigenetic silencing results in loss of antioxidant defense against repeated oxidative stress from chronic reflux–induced injury. The end result is accumulation of abnormal esophageal cells with a high degree of chromosomal instability. These observations along with introduction of novel epigenetic drugs in clinical practice open a new field of epigenetic chemoprevention.
Immunosurveillance
Normally Fas ligand (Fas-L) binds to Fas and initiates the cell apoptosis cascade. During neoplastic progression in BE, Fas-L is overexpressed and Fas expression is decreased in the epithelial cells. This protects dysplastic Barrett’s epithelial cells from self-destruction and also allows them to evade immune surveillance. Moreover, upregulation of Fas-L expression in dysplastic epithelium could induce apoptosis in Fas-expressing lymphocytes that further compromises tumor surveillance in BE. Strategies to restore immunosurveillance could positively impact the chemopreventive efforts.
Systemic
The mucosal response to luminal insults and epithelial-stromal interaction are modulated by systemic influences, which are outlined as follows.
Obesity-induced disruption of antireflux mechanisms and altered composition of reflux
Visceral obesity increases intra-abdominal pressure and changes the relationship between the gastroesophageal junction (GEJ) and diaphragmatic antireflux mechanisms. These changes increase reflux that eventually causes injury, BE, and EAC. Twenty-four–hour pH monitoring and motility studies reveal that obese patients have increased incidence of asymptomatic reflux and obesity is associated with significant drop in lower esophageal sphincter (LES) pressure. Pharmacologic agents that change LES pressure may therefore have chemopreventive potential in obese patients. Obesity also is associated with dietary and systemic derangements that could alter reflux composition. In a systematic review, McQuaid et al found that the concentration of carcinogenic bile acid in esophageal aspirate was higher in obese patients. Although compliance with dietary modifications is difficult, changes in dietary composition, such as the use of polyunsaturated fatty acids or targeting the obesity-associated derangements in bile salt synthesis with ursodeoxycholate or protecting esophageal mucosa from bile injury (topical protective barriers), could prevent carcinogenesis in BE.
Obesity-induced systemic inflammatory response
There is increasing evidence that obesity is a systemic endocrine derangement that results in a systemic and local proinflammatory state. Adipocytes release high circulating concentration of cytokines like TNF-α, IL-6, IL-1B, IL-10, C-reactive protein, interferon (INF)-ϒ, monocyte chemotactic protein, plasminogen activator inhibitor-1, and fibrinogen. These cytokines create a proinflammatory state, an important connecting link between obesity and various cancers, including EAC. Activated CD8+ T cells in the visceral adipose tissue produce IFN-ϒ resulting in TNF-α production in adipose tissue. Systemic release of this TNF-α acts locally on the subepithelial tissue and switches the stromal macrophage phenotype to modulate inflammation in target tissues such as Barrett’s mucosa. This hypothesis is supported by the findings of increased expression of both the ligand and the receptor of TNF-α during carcinogenesis in BE. This suggests that in overweight patients, adipocytes in endocrine manner, by influencing systemic and local inflammatory response, facilitate neoplastic transformation. This proinflammatory state also upregulates inducible nitric oxide synthase (iNOS) and nitric oxide production, which is typically seen during inflammation and carcinogenesis in BE. As alluded to earlier, this proinflammatory state creates oxidative stress and this altered oxidative tissue state favors mutagenesis to promote neoplasia. Interestingly, there is relative deficiency of antioxidant micronutrients like βcarotene, lycopene, and vitamin C in obese patients and consumption of vegetables and fruits rich in natural antioxidants is associated with a decreased risk of esophageal cancer. Approaches to alter mucosal response to injury through inhibitors of the previously mentioned pathways (anti–TNF-α or IL-1/iNOS inhibitors or dietary antioxidants) could prevent carcinogenesis in BE.
Pathways deranged in the metabolic syndrome
The metabolic syndrome is a cluster of conditions, including increased blood pressure, a high blood sugar level, excess body fat around the waist, and abnormal cholesterol levels, that occur together, increasing the risk of various cancers. One of the characteristic features of metabolic syndrome is insulin resistance. A diet high in energy and animal fat, and low in fiber in combination with physical inactivity contributes to insulin resistance and resulting hyperinsulinemia. Insulin, by activating insulinlike growth factor-1 receptors, stimulates cellular proliferation and inhibits apoptosis via the oncogenic PI3K-AKT-mTOR-S6K1 signaling cascade that facilitates carcinogenesis in BE. Next, leptin, as a part of the metabolic syndrome, increases EAC cell survival through PGE2-mediated activation of EGFR and c-Jun NH2-terminal kinase. Metabolic syndrome is also associated with downregulation of adipokine signaling and the expression of adiponectin receptors that has been correlated with EAC. Thus, obesity, lifestyle changes, and nutritional modification, although difficult to implement, remain important chemopreventive targets in BE.
Hypergastrinemia
Long-term PPI use in patients with BE leads to increased serum gastrin levels. Gastrin is a trophic hormone and promotes cell survival in the gastrointestinal tract. Gastrin binds to the cholecystokinin (CCK2) receptor that in turn induces EGF and the trefoil peptide expression. These gastrin-induced signals lead to COX-2 expression and facilitate carcinogenesis in BE. Interestingly Barrett’s mucosa expresses high levels of CCK2 receptors, and gastrin exposure is known to increase Barrett’s epithelial cell survival. Therefore, strategies to combine PPIs with gastrin inhibitors or the inhibitors of downstream signaling could prevent esophageal cancer.
End point evaluation
Ideally, cancer incidence and associated mortality should be included as primary endpoints in chemoprevention studies, but use of such definitive endpoints is not feasible during neoplastic transformation in Barrett’s. Using cancer risk reduction, although ideal and clinically relevant, is impractical because of the need for a lengthy follow-up and large number of trial participants to gather any meaningful information. To overcome these drawbacks, the markers to assess the chemoprevention response are broadly divided into 2 categories. The first set of markers, such as change in the grade of dysplasia, are clinically relevant. The next set of markers examine the effect of chemopreventive changes on the surrogate markers of neoplastic progression and also assess if the molecular pathways intended to be targeted by the chemopreventive agent are indeed downregulated.
Histopathology
Because grade of dysplasia is clinically used to predict the risk of cancer development and guide therapy in BE, one way to monitor chemoprevention would be to evaluate if the use of a chemopreventive agent downgrades the dysplastic changes or prevents progression to dysplasia in BE. Although histologic grade assessment is the principal predictor of Barrett’s progression, this approach is subjective. Typically, 20% to 30% of patients progress from nondysplastic to low-grade dysplastic Barrett’s and 14% to 28% of patients with high-grade dysplasia progress to cancer within 2 months to 4 years. [Reid, 1988 #1] However, only 85% of expert pathologists can confidently differentiate between EAC and high-grade dysplasia from nondysplastic Barrett’s and low-grade dysplasia. Moreover, agreement among pathologists regarding low-grade dysplasia diagnosis is dismal. In addition, high-grade dysplasia occupies less than 5% of the Barrett’s surface area, and a sampling error may erroneously increase the effectiveness of a chemopreventive agent. Use of degree of dysplasia as the chemoprevention endpoint does require complementary surrogate endpoint biomarkers (SEBs).
Surrogate Endpoint Biomarkers
The surrogate biomarkers that objectively correlate with cancer risk in BE are commonly used in chemoprevention trials. Depending on the phase of chemoprevention trial or the degree of dysplasia in the target population, these surrogate markers are used either independently or as complementary to histopathology. An important concept while selecting SEB is that the marker should not be an irreversible genetic event. Choosing such marker, by virtue of being irreversible, will portray a picture of agent inefficacy or failure regardless of whether or not the preventive agent is showing benefit.
Flow cytometery, image cytometry, and fluorescent in situ hybridization (FISH) show that abnormal p16 and p53 expression due to loss of heterozygosity, promotor silencing, or mutations, along with aneuploidy or hyper-tetraploidy could predict progression of neoplasia. Prasad and colleagues evaluated a biomarker panel by FISH (loss of 9p21 [p16] and 17p13.1 [p53] loci; gains of the 8q24 [c-myc], 17q [HER2-neu], and 20q13 [ZNF217] loci; and multiple gains) on cytology specimens collected before and after photodynamic therapy (PDT) in 31 patients with high-grade BE (HGD)/mucosal cancer. In 6 of 31 patients, abnormal expression was present for at least one marker after PDT in the absence of histologic abnormalities. Two of 6 patients with persistent biomarker abnormalities developed recurrent HGD. The previously mentioned FISH panel was then prospectively applied to patients undergoing PDT. At baseline, 68% of the patients with HGD and 73% with mucosal cancer had FISH abnormalities. On multivariate analysis, p16 loss at baseline predicted the lack of response of HGD/mucosal cancer to PDT at 3-month follow-up (0.32, 95% confidence interval [CI] 0.10–0.96). In a cohort of 181 patients undergoing radiofrequency ablation, the same FISH panel found that the presence of multiple gains at baseline predicted failure of complete remission of dysplasia (hazard ratio [HR] 0.57, 95% CI 0.40–0.82). As pointed out earlier, if these changes are irreversible, use of these markers may spuriously underestimate the chemopreventive potential of an agent. Contrary to this, epigenetic effects, such as global hypomethylation and promoter-specific hypermethylation that are more dynamic in nature, or microRNA will likely be better markers to assess the efficacy of a chemopreventive agent.
Recent studies have demonstrated progressive increase in the levels of GLI-1, vascular endothelial growth factor, transcription factor MCM2, COX-2 enzyme, PGE2, polyamine, and upregulation of prosurvival NF-kB during oncogenesis in BE. These biochemical and molecular markers of neoplastic progression also can assess if the molecular pathways intended to be targeted by chemopreventive agents are actually being downregulated. In addition to these, markers of increased cellular proliferation, along with elevated pro-oncogenic pathways, provide biologically relevant markers. Commonly used biomarkers for increased cellular proliferation include Ki-67, PCNA, and MCM2. These markers, along with markers of apoptosis, such as caspases and BcL2, determine overall cell survival. Another way to assess degree of dysplastic regression in Barrett’s patients would be to quantify the ratio between proliferative and apoptotic markers. Decrease in the ratio with use of chemopreventive agents would suggest dysplastic regression. Quantification of this ratio appears to be a practical approach to monitoring chemoprevention agent efficacy.
Many of the SEBs mentioned previously are currently used in phase 1 and 2 clinical trials either individually or as components of a pathway or to complement histopathology. If molecular, biological, and histologic outcomes change in parallel in response to a chemopreventive agent, there will be higher confidence in accepting the efficacy of a chemopreventive agent.
Clinical studies and randomized trials
The earlier discussion in this review focused predominantly on target identification and monitoring of chemoprevention. These important concepts need complementary clinical studies to predict or validate the efficacy of chemopreventive agents. Although together the information presented here provides a framework to discuss chemoprevention in clinical settings, chemoprevention at the present time remains an experimental approach.
Acid Suppression and Bile Salt Modification
As outlined earlier, acid and bile reflux initiate a cascade of proinflammatory pathways that are known to promote neoplasia. The caveats are that most patients with BE already use PPIs that change the degree of acid and it requires invasive monitoring to examine the effectiveness of interventions to suppress reflux. Therefore, this premise is mainly tested with case-control or observational studies.
Epidemiologic studies show conflicting evidence regarding chemopreventive potential of acid suppression using PPIs. In Veteran Administration setting, follow-up of 236 patients with BE for more than a decade showed approximately 60% risk reduction of dysplastic changes in patients using PPIs compared with those who did not use PPIs. In a larger Dutch study, a more robust risk reduction (up to 75%) of neoplastic progression in BE was noted with acid suppression. Similarly, a prospective review of surveillance data in 350 Australian patients showed that in the patients who delayed the use of PPIs for more than 2 years after Barrett’s diagnosis, there was an increased risk for developing low-grade dysplasia (HR 5.6, 95% CI 2.0–15.7) and high-grade dysplasia or cancer (HR 20.9, 95% CI 2.8–158). However, most of the progression in this study was to low-grade dysplasia, a diagnosis that is commonly contested. Although the clinical relevance of extent of Barrett’s mucosa to the risk of neoplastic progression is not clear, studies show more frequent squamous islands or a partial regression of BE in patients who take PPIs.
Contrary to these results, in a recent Danish study of 9883 patients with BE, a high adherence to PPI use was associated with increased relative risk (RR) of developing high-grade dysplasia and cancer (RR 3.4, 95% CI 1.1–10.5). A meta-analysis conducted by Singh and colleagues demonstrated a 71% risk reduction in development of EAC and high-grade dysplasia in patients who were on PPIs for more than 2 to 3 years, whereas patients who received less than 2 years of therapy did not seem to benefit much from PPI use. Congruent with these observations, small clinical trials addressed the question of whether profound acid suppression can have chemopreventive potential in Barrett’s mucosa. Epithelial cell proliferation, which correlates with dysplasia, decreases during near complete elimination of acid reflux in patients. Villin, which is a marker of differentiation, is more strongly expressed after acid reflux normalization in patients. PPI treatment is also associated with fewer abnormalities in mucosal expression of cell cycle regulatory proteins, such as p16, p21, and cyclins D1 and E, as well as decreased DNA damage in Barrett’s mucosa. In a different patient population of ablated Barrett’s epithelium, when mucosal ablation failed during insufficient acid suppression, a more profound acid suppression resulted in successful ablation. In a short-term study, however, normalization of acid reflux did not appear to be necessary. On the other hand, recurrence of metaplasia was observed after discontinuation of omeprazole maintenance. Because mucosal ablation studies are typically performed with concomitant acid suppression, they do not clearly address the chemopreventive potential and require more definitive studies.
During reflux injury, bile acids are directly implicated in producing esophageal epithelial damage. Bile salt–related injury is pH dependent and the use of PPIs not only alters the pH, but also decreases the overall quantity of reflux. Interestingly, one-third of the patients on PPI treatment continue to have bile reflux. Moreover, different bile acids inflict injury to the epithelial tissue at different pH, as outlined earlier in this review. The composition of bile acids in duodenal bile is modifiable. In healthy individuals, the bile mainly consists of cytotoxic hydrophobic bile acids, such as deoxycholic acid and cholic acid. These cytotoxic bile acids are largely replaced by tertiary hydrophilic bile acids when treated with ursodeoxycholic acid, thus reducing bile acid–induced injury to the esophagus. Although in animals with reflux, ursodeoxycholic acid use led to suppression of inflammation-metaplasia-neoplasia sequence, in patients with BE, urso failed to change markers of differentiation or proliferation. In samples derived from patients with BE and in preclinical investigations, our laboratory found that bile acids trigger activation of GLI-1, an effector of the Hedgehog pathway, which is upregulated during carcinogenesis in BE. GLI-1, through cell cycle regulator CDK2, promotes Barrett’s epithelial cell proliferation. Combination of ursodeoxycholic acid and aspirin via their inhibitory action on GLI-1 downregulate CDK2, decrease cell growth, and prevent cancer in animals. It is therefore possible that if ursodeoxycholic acid is combined with aspirin in patients with BE, it could prevent carcinogenesis.
Nonsteroidal Anti-inflammatory Drugs (Celecoxib)
NSAIDs target COX enzyme, which regulates prostaglandin synthesis from arachidonic acid. COX is implicated in the oncogenic transformation of Barrett’s to EAC. Compared with nonusers, NSAIDs reduce the risk of development of EAC or esophageal junction adenocarcinomas by 40% ( P <.01) in patients who take them frequently and for longer durations of time (>10 years). Based on these epidemiologic data and aforementioned in vivo and in vitro studies, a phase IIb multicenter randomized placebo-controlled trial of celecoxib in patients with BE was conducted (Chemoprevention for Barrett’s Esophagus Trial or CBET). In this study, 100 Barrett’s patients with low-grade or high-grade dysplasia were randomly assigned to treatment (n = 49 to celecoxib 200 mg twice a day) or placebo (n = 51). After 48 weeks of treatment, no difference was observed in the median change in the proportion of biopsy samples with dysplasia or cancer between treatment groups in either the low-grade (median change with celecoxib = −0.09, interquartile range [IQR] = −0.32–0.14 and with placebo = −0.07, IQR = −0.26–0.12; P = .64) or high-grade (median change with celecoxib = 0.12, IQR = −0.31–0.55, and with placebo = 0.02, IQR = −0.24–0.28; P = .88) stratum. No significant differences in total surface area of the BE; in prostaglandin levels; in COX-1/2 mRNA levels; or in methylation of tumor suppressor genes p16, adenomatous polyposis coli, and E-cadherin were found with celecoxib compared with placebo. The study concluded that 200 mg celecoxib twice daily for 48 weeks of treatment did not prevent progression of Barrett’s dysplasia to cancer. The lack of effect could be due to unabated prostaglandin synthesis by COX1 or decreased catabolism of prostaglandin E2 by PGE2 dehydrogenase.
Ornithine Decarboxylase Inhibitors
The polyamines putrescine, spermidine, and spermine, produced during cellular metabolism, are closely associated with inflammation as well as oxidative stress and are signaling molecules for cell growth and differentiation. Ornithine decarboxylase (ODC) is a key enzyme that is involved in the synthesis of polyamines. Patients with BE have markedly increased ODC activity that could potentially promote neoplastic changes in BE. Difluoromethylornithine (DFMO) is an irreversible inhibitor of ODC. Although DFMO prevents growth of Barrett’s epithelial cells, there is no correlation between polyamine levels and ODC activity in patients with Barrett’s. In a cohort of patients with Barrett’s with low-grade dysplasia (n = 10), administration of DFMO (0.5 g/m 2 /d) for 6 months significantly reduced the levels of putrescine, spermidine, and the spermidine/spermine ratio. Furthermore, DFMO downregulated RPL11, which is known to activate tumor suppressor p53 pathway and inhibited KLF5, a transcription factor that promotes cell proliferation. There was also a partial regression of extent of low-grade dysplasia, but clinical relevance of the finding is unclear. In summary, an ODC inhibitor-based chemopreventive approach has mechanistic rationale but only limited success in clinical settings. Combining DFMO with anti-inflammatory drugs, such as NSAIDs or antioxidants, appears to be a reasonable strategy.
Metformin
Metformin lowers serum insulin levels, activates adenosine monophosphate (AMP)- activated protein kinase (AMPK). AMPK activation by metformin increases insulin-dependent glucose uptake and inhibits mTOR, resulting in downregulation of ribosomal protein S6 kinase 1(S6K1) that leads to decreased protein synthesis. This decrease in phosphorylated S6K1 (pS6K1) inhibits cell proliferation. Metformin also has AMPK-independent, indirect antiproliferative effects related to lower systemic levels of insulin. Based on these observations, 74 subjects with BE were recruited to test the chemopreventive potential of metformin (2000 mg/d, n = 38) or placebo (n = 36) for 12 weeks. Biopsy specimens were collected at baseline and at week 12. This was a negative trial, as the percent change in median level of pS6K1 did not differ significantly between groups (1.4% among subjects given metformin vs 14.7% among subjects given placebo; 1-sided P of .80). Metformin was associated with an almost significant reduction in serum insulin levels (median −4.7% among subjects given metformin vs 23.6% increase among those given placebo, P = .80), as well as in homeostatic model assessments of insulin resistance (median −7.2% among subjects given metformin vs 38% among subjects given placebo, P = .06). Metformin had no effects on cell proliferation (on the basis of KI67 assays) or apoptosis (on the basis of caspase 3 assays). This study did raise the possibility that the use of metformin in obese patients with Barrett’s with a high degree of insulin resistance may have a chemopreventive function.
Statins
Preclinical studies show that statins via 3-hydroxy-3-methylglutaryl-coenzyme A reductase-dependent and independent manner exert chemopreventive potential. By inhibiting posttranslational modification of the Ras/Rho superfamily, statins decrease cell growth. Observational and case-control studies also support the chemopreventive potential of statins. Patients with Barrett’s who fill statin prescriptions are at reduced risk of EAC (0.55, 95% CI 0.36–0.86), In the same study a significant trend toward greater risk reduction was noted with longer duration of statin use. In a prospective study of 570 patients with BE in Dutch hospitals during a median follow-up period of 4.5 years, 38 patients (7%) developed high-grade dysplasia or adenocarcinoma. After Barrett’s diagnosis, in 209 (37%) patients who used statins, there was a reduced risk of neoplastic progression (HR 0.47, P = .030, and HR 0.46, P = .048, respectively). In another prospective cohort of 411 patients with Barrett’s, after accounting for variation in use during follow-up and adjusting for age, sex, and smoking, the HR for statin use among patients with high-grade dysplasia was 0.31 (95% CI 0.11–0.86).
In a case-control study comparing statin use between patients with an incident EAC (n = 85) and matched nonprogressive Barrett’s controls (n = 170), there was significantly lower incidence of EAC in statin users (uncorrected odds ratio [OR] 0.45, 95% CI 0.24–0.84). After correction for confounding variables, including aspirin and NSAID use, statin use was still associated with a reduced incidence of EAC (OR 0.57, 95% CI 0.28–0.94). Longer duration of statin use and higher doses were both associated with a significantly greater reduction in EAC. Recently, a case-control study was conducted among patients scheduled for elective endoscopy in primary care settings at a Veterans Affairs center. A total of 303 patients with BE were compared with 2 separate sex-matched control groups: 606 elective endoscopy controls and 303 primary care controls without BE. Statin use was associated with a significantly lower risk of BE (adjusted OR 0.57, 95% CI 0.38–0.87) compared with the combined control groups. The risk of BE was especially lower with statin use among obese patients (OR 0.26, 95% CI 0.09–0.71), as was the risk for BE segments of 3 cm or larger (OR 0.13, 95% CI 0.06–0.30). There was no significant association between BE and nonstatin lipid-lowering medications ( P = .452). Contrary to these studies, a series of nested case-control studies covering 574 UK general practices with 88,125 cases and 362,254 matched controls, showed that the adjusted OR for any statin use and cancer at any site was 1.01 (95% CI 0.99–1.04). Interestingly, statin use had no significant negative association with EAC.
A recent meta-analysis that included 13 studies (including a post hoc analysis of 22 randomized controlled trials) reporting 9285 cases of esophageal cancer among 1,132,969 patients showed a 28% reduction in the risk of esophageal cancer among patients who took statins (adjusted OR 0.72, 95% CI 0.60–0.86). In a subset of patients known to have BE (5 studies, 312 EAC developed in 2125 patients), statins were associated with a 41% decrease in the risk of EAC, after adjusting for potential confounders (adjusted OR 0.59, 95% CI 0.45–0.78). The number needed to treat with statins to prevent 1 case of EAC in patients with BE was 389. Together these studies suggest a chemopreventive potential of statins in the development of BE and progression to EAC. However, randomized prospective studies are warranted.
Polyunsaturated fatty acids
Obesity is rising in epidemic proportions in Western countries. The increased incidence of obesity is attributed to distinct dietary patterns and sedentary life style. Subjects with a body mass index (BMI) in the highest quartile have a fourfold increase in the risk of developing EAC compared with subjects with a BMI in the lowest quartile. It appears that the association between being obese and EAC is mediated via central adiposity. A recent study using computed tomography of the abdomen shows that the excess visceral fat, compared with subcutaneous fat, is more likely to predict EAC risk. In a case-control study of 50 BE cases and 50 controls, Nelsen and colleagues found that both visceral fat and GEJ fat were significantly greater in patients with HGD compared with those without HGD, independent of BMI and GERD. It is proposed that diet rich in saturated fatty acids leads to increased visceral obesity that enhances esophageal inflammation by paracrine mechanisms through proinflammatory phenotype macrophage infiltration in Barrett’s mucosa. Because epidemiologic data also suggest a reduced risk of EAC in populations with a high consumption of fish, and n−3 fatty acids inhibit experimental carcinogenesis, the effects of dietary supplementation with the n−3 fatty acid eicosapentaenoic acid (EPA) on a number of biological endpoints in BE were examined. Fifty-two patients with Barrett’s were randomly assigned to consume EPA capsules (1.5 g/d) or no supplement (controls) for 6 months. There was a significant decline in oncogenic COX-2 protein concentrations in the n−3 group compared with controls ( P <.05). The change in COX-2 protein was inversely related to the change in EPA content ( P <.05). However, cellular proliferation was not different between the 2 groups. Designing an EPA supplementation trial in patients with visceral obesity or high levels of serum saturated fatty acids may show the chemopreventive potential of polyunsaturated fatty acid such as EPA.
Combinatorial Chemoprevention
Most patients with BE are on acid suppressive therapy, which increases postprandial and fasting serum gastrin levels. Gastrin binds to the cholecystokinin (CCK2) receptor, and the downstream signaling increases COX2 expression, which is known to promote carcinogenesis in BE. In fact, gastrin does increase Barrett’s epithelial cell survival by inducing proliferation and inactivating proapoptotic factors to promote carcinogenesis. In a retrospective, patients with the highest quartile of serum gastrin levels were more likely to have high-grade dysplasia or cancer (OR 5.46, 95% CI 1.2–24.8). Because PPIs suppress acid-reflux–related procarcinogenic signaling, combining them with inhibitors of gastrin-dependent signaling appears to be a novel strategy to prevent the development of esophageal cancer. To address this, a recent multicenter double blind, randomized, placebo-controlled, phase 2 clinical trial recruited 114 patients with nondysplastic or low-grade dysplastic BE. Patients were randomized into 3 groups and received 40 mg esomeprazole with either placebo, low-dose aspirin (81 mg) or high-dose aspirin (325 mg) daily for 28 days. Esophageal endoscopic biopsies were performed before and after intervention to measure PGE2 levels and cell growth. It was found that a higher dose of aspirin combined with esomeprazole significantly reduced tissue concentration of PGE2 as well as cell growth (unpublished data) and a low dose of aspirin as well as placebo did not have a significant effect. These findings, along with the ongoing ASPECT trial, could pave the way to use aspirin along with PPI as a chemopreventive strategy in BE in the near future.
Related posts:
Forthcoming Issues
Screening for Barrett’s Esophagus
Cost-Analyses Studies in Barrett’s Esophagus
Epidemiology of Barrett’s Esophagus and Esophageal Adenocarcinoma
Challenges with Endoscopic Therapy for Barrett’s Esophagus
Genetic and Epigenetic Alterations in Barrett’s Esophagus and Esophageal Adenocarcinoma
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
