This Month in Gastroenterology

Article Outline

• Leukocyte Apheresis in Ulcerative Colitis
• Biomarkers in Barrett's Esophagus
• Higher Mutant DNA Fractions in Stool Allows Better Detection Over Plasma in Colorectal Cancer Patients
• Hedgehog Signaling Is Required for Acinar Redifferentiation During Pancreatic Regeneration
• Copyright

Back to Article Outline

Leukocyte Apheresis in Ulcerative Colitis 

Current medical therapy for ulcerative colitis aims at suppression of inflammatory or immune responses through the use of aminosalicylates, steroids, immune suppressants, or infliximab. In severe cases, proctocolectomy can be considered. However, these therapeutic options are not devoid of limitations, such as incomplete responsiveness, potentially serious side effects, or the invasive nature of the intervention. Hence, there is a need for additional therapeutic approaches for ulcerative colitis patients.

Leukocyte apheresis, aimed at improving colitis by removing activated granulocytes and monocytes from the peripheral blood, has been proposed as an alternative treatment approach in ulcerative colitis. Uncontrolled studies reported the induction of clinical remission in substantial proportion of ulcerative colitis patients treated with leukocyte apheresis, without major safety concerns. Based on these data, the apheresis column has been marketed for treatment of ulcerative colitis in Japan and in Europe, although large controlled trials were lacking.

In this issue of Gastroenterology, Sands et al report on the results of a sham-controlled international trial of leukocyte apheresis in ulcerative colitis. Patients with active ulcerative colitis (Mayo score 6–11) despite therapy with aminosalicylates, steroids, azathioprine, or 6-mercaptopurine, were eligible for the trial. The apheresis device (Adacolumn) consists of a monocyte/macrophage adsorption column filled with cellulose acetate beads, and a blood pump that are interposed as an extracorporeal blood circuit between the patient's 2 (donor and recipient) arms. Patients were randomized in a 2:1 ratio to treatment with an active or a sham (with internal bypass) column. Patients underwent 10 treatment sessions over a 9-week period (2 sessions per week the first 2 weeks, followed by 1 session per week for 6 weeks, with a 1-week break), each session consisting of a 60-minute apheresis procedure with a blood flow rate of 30 mL/min. Patients were evaluated for efficacy at week 12.

A total of 168 patients were eligible for randomization, with 112 randomized to active therapy and 56 to sham apheresis. After 12 weeks of therapy, the occurrence of clinical remission (respectively, 17% and 11%) or clinical response (respectively, 44% and 40%) did not differ between the apheresis and sham groups. In addition, the rates of endoscopic remission (17% vs 16%) and changes in Mayo scores (on average 1.5 vs 1.4) were similar in both study arms (Figure 1). Quality of life, assessed with the Short Form-36 and the Inflammatory Bowel Disease Questionnaire, did not differ between both treatment groups. Tolerability and adverse events were similar.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Clinical outcomes in ulcerative colitis patients treated by leukocyte or sham apheresis.

Hence, this study failed to demonstrate efficacy of leukocyte apheresis in the ulcerative colitis patient population that was studied. The results are at variance with high rates of remission reported in previous, mainly uncontrolled, trials with this treatment modality. Besides differences in trial design, it is conceivable that patient selection may have influenced outcomes. Previous uncontrolled studies included mainly steroid-dependent or -refractory patients. A post hoc analysis of the present study suggests potential effectiveness in patients with histologic evidence of severe acute inflammation with ulceration and erosion. However, this possibility awaits confirmation in a separate randomized controlled trial.

See page 400.

Back to Article Outline

Biomarkers in Barrett's Esophagus 

Barrett's esophagus is associated with increased risk for esophageal adenocarcinoma. High-grade dysplasia is considered to be a marker of progression to carcinoma, and surveillance endoscopy with biopsies to detect advanced dysplasia and mucosal neoplasia is a well-accepted recommendation for patients with Barrett's esophagus. The finding of high-grade dysplasia prompted esophagectomy in the past, but over the last decade a number of endoscopic approaches have been developed, including photodynamic therapy and endoscopic mucosal resection.

Although long-term outcomes of photodynamic therapy in high-grade dysplasia are favorable, subgroups of patients do not respond or still progress to carcinoma. There is a clear need for better identification of patients who are less likely to respond to photodynamic therapy, who may require more invasive therapeutic approaches. In this issue of Gastroenterology, Prasad et al report on a prospective study evaluating the ability of genetic alterations to predict in patients with high-grade dysplasia or mucosal neoplasia in a Barrett's esophagus. In patients undergoing photodynamic therapy and in patients who chose to remain under surveillance, fluorescence in situ hybridization to detect 5 genetic alterations that previously have been associated with high-grade dysplasia was performed on esophageal brush cytology specimens.

A total of 129 patients (71 who underwent photodynamic therapy and 55 patients who were under surveillance only) were recruited for the study. Clinical characteristics and prevalence of genetic alterations were comparable for both groups. At follow-up surveillance biopsies after 3 months, 40% of the patients did not have evidence of any dysplasia or carcinoma. On univariate analysis, increasing length of the Barrett's segment, P16 allelic loss, and P53 allelic loss predicted lack of response. In multivaraiate analysis, P16 allelic loss was a significant predictor of lack of response (Figure 2). The authors propose that loss of P16 may lead to loss of apoptosis induction by photodynamic therapy.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2. 

  Receiver operator characteristic curve for a cross-validated model with P16 loss as a predictor of lack of response to photodynamic therapy (adjusted for clinical variables).

Thus, this study suggests identified P16 allellic loss as a predictor of poor response to photodynamic therapy. This observation suggests that the use of well-selected biomarkers, such as P16, may help in the selection of the most favorable therapy in individual patients, for instance with esophageal high-grade dysplasia.

See page 370.

Back to Article Outline

Higher Mutant DNA Fractions in Stool Allows Better Detection Over Plasma in Colorectal Cancer Patients 

In addition to colonoscopy, computed tomographic colonography, flexible sigmoidoscopy, and fecal occult blood and fecal hemoglobin immunohistochemical tests, stool DNA tests are a suggested method to screen for colorectal cancer. Current stool DNA tests, like other fecal testing methods, have poor sensitivity, and newer technologies could improve detection rates. Plasma-based tests for detection of colorectal cancer have not been commercially developed; however, mutations central to colorectal cancer pathogenesis have been detected in plasma in some patients with colorectal cancer.

In the study by Diehl et al, 25 patients representing all 4 stages of colorectal cancer were assessed for mutations in APC, TP53, KRAS, PIK3CA, and CTNNB1 in the primary tumor, and compared mutation detection in stool and plasma (stool and plasma samples were collected 6–12 days postcolonoscopy and before bowel preparation for surgery). Because stool contains bacterial and human DNA, stool was enriched for human DNA by a reversible electrophorectic capture affinity protocol; subsequently, stool and plasma DNA were examined by beads, emulsions, amplification, and magnetics technology, or BEAMing, a way to convert single DNA template molecules to singe beads containing tens of thousands of exact copies of the template. BEAMing afforded quantitative assessment of mutant DNA, at an optimal amplicon size of ∼100 bp (over larger sizes) in stool. With a median background of mutations in human lymphocytes (from polymerase chain reaction errors during amplification) at 0.0009%, median mutant DNA fragments from stool was 1,000-fold higher than background but did not vary across stage, with stages I, II, III, and IV having 0.83%, 0.31%, 0.29%, and 0.62%, respectively (Figure 3). Twenty-three of 25 patients (92%) have mutant DNA detected in their stool by BEAMing, an improvement over the current single-base extension approach for fecal DNA detection, which found mutant DNA in only 60% of patients, which may be due to the abundancy of mutant DNA. From plasma DNA, only 8 of 16 patients (50%) had detectable mutant DNA, and in the 8 positive, the median fraction of mutant DNA was 0.42%, similar to that of stool (0.37%).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 3. 

  Mutations in fecal DNA and TNM stage. The horizontal bars shows the median fraction of mutant DNA. The whiskers represent the minimum and maximum values that were found for each indicated stage.

The study indicates that BEAMing technology, with a sensitivity of detecting one mutant template among 10,000 normal templates, can identify 92% of colorectal cancer patients when stool is examined, and 50% of colorectal cancer patients when plasma is examined. Detection does not seem to be dependent on stage. Colorectal tumors shed about 0.40% of total human DNA in stool and plasma. The total amount of DNA available could be one influence on lower plasma detection rates. Additionally, detection of large adenomas, which could be more difficult to detect in plasma DNA, will need to be evaluated.

See page 489.

Back to Article Outline

Hedgehog Signaling Is Required for Acinar Redifferentiation During Pancreatic Regeneration 

The ability of the pancreas to regenerate after acute injury is at least partially dependent on upon surviving acinar cells to dedifferentiate, induce a pancreatic progenitor phenotype, then regenerate new cells to repopulate acini. Although sonic hedgehog signaling has been demonstrated in embryonic tissue patterning and implicated in the maintenance of stem cells in a number of adult tissues, the role of hedgehog signaling in pancreatic regeneration is not clear.

In the study by Fendrich et al, cerulein-induced pancreatic injury was performed in wild-type C57BL/6J mice, wild-type mice treated with the steroid alkaloid cyclopamine that inhibits hedgehog signaling through its direct interaction with Smoothened, a key signal transducer for hedgehog, and in 2 genetic mouse models in which hedgehog signaling is abrogated for the Smoothened allele. Cerulein produces acute pancreatic injury that is highest 2 days after treatment, with full histologic recovery and restoration of acinar cell mass by day 7 after treatment. During injury, transient metaplastic epithelium derived from acinar cells formed, which were highly proliferative that up-regulated sonic hedgehog signaling by the first day post-cerulein injection that abated significantly by day 5 and completely absent by day 7 posttreatment. Cyclopamine, which blocks the hedgehog signaling up-regulation, failed to block the degree of cerulein-induced pancreatic injury; however, significant impairment of pancreas regeneration was evident with a marked reduction in differentiated acinar cells and its corresponding amylase expression, but with the persistence of transient metaplastic epithelium through day 7 post-cerulein treatment (Figure 4), highly suggestive of the need for hedgehog signaling to allow redifferentiation of the transient metaplastic epithelium. Cerulein injury in Pdx-Cre;smo^fl/fl and tamoxifen-induced Ela-CreERT2;smo^fl/fl, mouse models null for Smothened either in the entire pancreas or specifically in adult acinar cells, respectively, showed a failed regeneration response identical to that seen with cyclopamine.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 4. 

  Quantification of initial injury and subsequent regeneration using 2-dimensional fluorescent intensity profiles of cerulein-mediated injury in mouse pancreata. The upper panel demonstrated similar reductions in amylase staining in each group on day 2, indicating similar severity of initial cerulein-mediated (Cae) injury in each group. In the lower panel, failure to generate amylase-expressing cells on day 7 is seen in the setting of either pharmacologic or genetic blockade of hedgehog signaling. Either systemic blockade of hedgehog signaling by cyclopamine (Cae+Cyc), or deletion of Smoothened either throughout the pancreatic epithelium (Pdx-Cre;smo^fl/fl Cae) or specifically in acinar cells (Ela-CreERT2;smo^fl/fl Cae) results in failure to regenerate exocrine cell mass. Wt, wild-type mice; PBS, phosphate-buffered saline, a negative control.

This study indicates that hedgehog signaling after cerulein-mediated pancreatic injury is necessary for dedifferentiated acinar cells to contribute to tissue renewal. This finding has implications for patients with acute pancreatic injury, as well as a possible link with the formation of pancreatic neoplasia.

See page 621.