Evidence from Case Series
An overview of all extracted case series is depicted in Table 1. In total, 215 UC patients, 52 CD patients, 5 patients with indetermined colitis, and 21 controls were studied. Assessment of CMV was primarily done using histology (ie, H&E staining), identifying typical large cells containing basophilic intranuclear inclusion and frequently surrounded by a clear halo (“owl's eye”). Also, immunoperoxidase staining for CMV antigens was often performed to confirm these findings. PCR amplification techniques were only used in 3 studies. CMV prevalence varied greatly between studies, however, true prevalence rates in IBD patients cannot be abstracted from the type of studies that are under consideration. All reported cases received corticosteroids. In 5 of 9 studies, patients received some form of immunosuppression (thiopurines or cyclosporine). Interestingly, antiviral treatment did not induce a clinical response in all patients and did not prevent surgery in some. 15–17 Whether this was a result of a lack of antiviral therapy or the severity of UC is not clear. The wide variety of conclusions drawn from these studies can be appreciated from Table 1.
Table 1. Overview of Case Reports Addressing the Presence of CMV in IBD
|Case||Series||Primary Objective||Patients||CMV Assessment||CMV Prevalence||Co-medication at at Baseline||Outcome||Conclusion||Ref.|
|2001||Papadakis||Outcome of CMV||UC5||Histology PCR||UC + CD:||CSA, steroids.||10/10 gancyclovir,||CMV screening||16|
| || ||infection||CD 3||tissue Serology||10/1895||azathioprine,||5/10 incomplete||necessary in|| |
| ||complicating||IC2||6-MP||remission,||refractory IBD||IBD||2/10 no change|| || |
|2001||Cottone||CMV prevalence||UC 55||Histology PCR||UC 5/19 (SR)||2/7 azathioprine||5 clinical||CMV causes||15|
| ||in refractory||CD 7||buffy coat||CD 2/19 (SR)||4/7 steroids||remission after||refractory||antivirals||colitis|
|2000||Alcala||CMV prevalence||UC 39||Histology||UC 7/39 (SR)||39/39 steroids||39/39 surgery||CMV complicates||42|
| || ||in resected UC||CD 0||5/39 CSA||(No antivirals)||refractory UC||patients|| || |
|1999||Kaufman||CMV prevalence||UC 8||Histology||UC 8/172||10/11 steroids||11/11 surgery||CMV more||43|
| || ||in reseted IBD||CD 3|| ||CD 3/376||I/11 6-MP||(No antivirals)||prevalent in|| |
| || ||patients||UC than in CD|| || || || || || |
|1999||Vega||Report on own||UC 7||Histology||UC 7/7||9/9 steroids||2/9 surgery||CMV may play a||17|
| || ||experience||CD 2|| ||CD 2/2||4/9 CSA|| ||role in IBD|| |
| || || || || || || || ||pathogenesis|| |
|1992||Wakefield||CMV prevalence||UC21||PCR tissue||UC 17/21||UC 16/21 steroids||Unspecified||CMV (+coexist-||27|
| || ||in IBD||CD 29||PCR PBMC||CD 19/29||CD 12/29 steroids|| ||ing HHV-6)|| |
| || || ||Ctrl 21|| ||Ctrl 6/21||Ctrl: no steroids|| ||may play a role|| |
| || || || || ||PBMC: none|| || ||in IBD pathogenesis|| |
|1986||Eyre-Brooks||CMV prevalence||UC 20||Histology||UC 2/20||UC 2/2 steroids||3/3 surgery||C M V does no||23|
| || ||in resected IBD||CD 3|| ||CD none||CD none|| ||alter course of|| |
| || ||patients||IC3|| ||IC 1/3||IC steroids|| ||disease|| |
|1979||Swarbrick||Relationship of||UC 24||CMV serology||UC 19/24||29/29 steroids||No change in titre||Unlikely that||44|
| || ||CMV,||CD 5|| ||CD 4/5|| ||during active||CMV affects|| |
| || ||Chlamydia and|| || || || ||disease||clinical IBD|| |
| || ||Yersinia in|| || || || || ||course|| |
| || ||IBD|| || || || || || || |
|1977||Cooper||CMV prevalence||UC46||Histology||UC 6/46||46/46 steroids||5/6 CMV+||CMV alters||45|
| || ||in resected UC||CD 0|| || || ||patients: toxic||clinical course|| |
| || ||patients|| || || || ||megacolon||of the disease|| |
Evidence for CMV Latency and Reactivation
CMV infection is common, occurring in 40%–100% of persons, 4 and CMV persists indefinitely within the host. The current view is that peripheral blood monocytes constitute a major site of viral latency and that differentiation into tissue macrophages as a result of proinflammatory cytokines and chemokines leads to a productive CMV reactivation and infection. 18–21 Because Papadakis et al 16 demonstrated that all patients with severe refractory IBD tested positive for CMV IgG antibodies, it seems plausible that the observed CMV complications are the result of CMV reactivation rather than de novo infection. Indeed, initial data on CMV seroprevalence in UC patients showed higher rates in patients versus controls. 22 However, true prevalence rates of CMV in the IBD population cannot be abstracted from the type of studies that are described in this manuscript.
Eyre-Brook and Dundas 23 made an important observation: Patients, positive for CMV, and not receiving antiviral therapy but undergoing colectomy were clinically in remission after surgery, and no CMV was detected in the remaining bowel. Moreover, because of low platelets, splenectomy was performed in one patient undergoing colectomy for therapy-resistant UC, and no evidence of CMV as found in the spleen. This strongly suggests that CMV reactivation in active colitis is confined to the diseased region of the colon, and confirms that de novo CMV infection in this patient population is highly unlikely.
Evidence for Cellular Targets Within the GI Tract
Localization and characterization of CMV-infected cells within the diseased mucosa of IBD patients could provide insight into the cellular targets of the replicating virus during inflammation. In biopsy specimens containing inclusion bodies on H&E staining, CMV was detected primarily in endothelial cells and the perivascular region. 24 In colonic tissue specimens of both CD and UC patients, CMV inclusions were primarily identified in the lamina propria and submucosa at sites of mucosal ulceration, as well as in endothelial sites. 23 Using immunohistochemistry, it was shown that CMV was present primarily in endothelial cells in the gastrointestinal tract. 24,25
These findings are in keeping with earlier studies demonstrating that endothelial cells are a common target for CMV infection in vivo regardless of the organ involved. 26 Endothelial cells can serve as fully permissive hosts for CMV in vitro provided the natural endothelial cytopathogenicity of the virus is preserved. The vascular endothelium represents the interface between circulating immunocompetent cells and the lamina propria of the gut. The endothelium is thought to be an active participant in immunomodulation during IBD: endothelial cells are capable of regulating leukocyte migration, expressing HLA molecules and presenting antigen. Furthermore they respond to a wide variety of cytokines, and inducibly express immunoreactive cellular adhesion molecules.
Evidence of Other Herpesviruses in IBD
Using nested polymerase chain reaction (nPCR), Wake-field et al investigated whether other herpes family members were present in IBD tissue specimens. 27 Herpes simplex virus-1 (HSV1) and varicella zoster virus (VZV) could not be detected, whereas viral DNA of both human herpes virus-6 (HHV6) and Epstein-Barr virus (EBV) were present in samples from Crohn's patients as well as ulcerative colitis. A remarkable observation in this study was the low incidence of viral DNA in peripheral blood mononuclear cells in comparison with the incidence of viral DNA present in colonic tissue samples, suggesting that viral replication is restricted to the diseased area of the gut. Using in situ hybridization, EBV was detected in resected colon specimens of 7 of 11 Crohn's patients and 3 of 5 patients suffering from ulcerative colitis. 28 These patients did receive corticosteroids, but none was treated with immunosuppressives. Interestingly, no EBV-positive cells were found in the noninflamed parts of the resected IBD specimens; again, this strengthens the hypothesis that viral multiplication is a localized phenomena strictly occurring within the inflamed mucosa.
CMV is a ubiquitous herpesvirus that latently infects the majority of adults, both the acute infection and subsequent reactivation of the virus being relatively benign and not associated with considerable pathology in an immunocompetent individual. In contrast, primary infection or reactivation in immunocompromised patients is associated with significant morbidity and mortality. The exact role of CMV in the IBD population remains unclear, and our aim was to review systematically all available evidence from the literature and to generate a number of assumptions to develop rational recommendations for future research. In short, critical appraisal of the available studies does not allow any firm conclusions on CMV seroprevalence in IBD, nor on CMV pathogenicity or rational therapeutic approaches using antivirals. Detection of the virus in IBD patients does not necessarily indicate that CMV is causing disease. Indeed, CMV shedding and viremia can occur even when disease due to CMV is not present.
Key questions that remain unanswered are as follows: (i) Is CMV reactivation caused by concomitant use of immuno-suppressives? (ii) Is CMV reactivation secondary to the inflammatory response seen in IBD? (iii) What is the optimal therapeutic strategy for active IBD with concomitant CMV infection? (iv) What is the best possible diagnostic tool for CMV screening and monitoring during disease activity?
When considering the first question, one has to consider the possibility that all latently infected IBD patients receiving immunosuppressives frequently produce infectious virus, and reactivation is due primarily to the failure of the host immune response to remove infectious virus. This is not supported by the published literature. Although reactivation can be rapidly induced in animal models by several types of immunosuppression, these types included whole body irradiation, 29 treatment with cytotoxic drugs such as cyclophosphamide 30,31 or antibody-mediated depletion of lymphoid cells, 32 regimens not used in IBD. Moreover, these strategies resulted in high levels of cell death and cytokine release, which may themselves play a role in viral reactivation. The role of azathioprine or cyclosporine during CMV replication has, to our knowledge, not been reported.
One of the intriguing outstanding questions is the role of mucosal inflammation in inducing and sustaining CMV reactivation. During active IBD, local expression of a wide range of proinflammatory cytokines including TNF-α, IFN-β, and IL-2 is induced, responsible for activation of transcription factors (NF-κB, AP-1) which in turn induce expression of chemokines (IL-8, MIP-1α, RANTES, MCP-1) and adhesion molecules (E-selectin, ICAM-1, VCAM-1) crucial for recruiting immunocompetent cells. Stimulated by this cytokine/chemokine release from the affected gut, it is likely that latently infected circulating monocytes and dendritic cells migrate to these sites of inflammation and differentiate into permissive cells supporting active replication. 21,33–36 As mentioned earlier, it has been proposed that endothelial cells can serve as permissive hosts for CMV during inflammatory responses and thus serves as a key mediator for CMV reactivation. CMV-infected endothelial cells powerfully stimulate T cells to produce IL-2 and to proliferate. 37 Consequently, activated T cells produce both IFN-γ and TNF-α, further driving the proinflammatory responses and potentially aggravating the course of the disease.
What is the optimal therapeutic strategy for active IBD complicated by CMV infection? In the reviewed literature it is suggested that some patients benefit from CMV eradication during episodes of active IBD. However, not all patients showed clinical improvement, and some needed a colectomy even after successful antiviral treatment. These observations underscore the need to discriminate between clinical conditions of CMV shedding from those where CMV actually is causing aggravation of inflammatory bowel disease. To bring all published observations within a logical framework, we propose a schematic representation of sequelae during bowel inflammation in CMV seropositive patients (Fig. 1). During the initiation phase, mucosal inflammatory responses cause increases in local cytokine/chemokines production, triggering the hosts' latently infected cells. Patients' serology will reveal IgG positivity, but detection methods like PCR and histology are negative. Subsequently, stimulated monocytes differentiate into plasma cells and dendritic cells migrating to the inflamed site. Evidence suggests that this reactivation phase is restricted to the area of inflammation. Infected cells can now be detected in biopsy specimens on H&E-staining. Finally, during the consolidation phase, active replication takes place predominantly in endothelial cells, and possibly is responsible for further aggravating the inflammatory response. Viral particles are shed both into the circulation as well as in the gut lumen as can be detected by PCR analysis of blood and feces, respectively. The impact of active viral replication on disease severity remains to be elucidated.
CMV infection is a significant cause of morbidity and mortality in immunocompromised individuals, such as transplant recipients and AIDS patients, therefore robust quantitative detection methods have been developed in recent years. 38 The quantitative detection of CMV DNA in the blood compartment by PCR is increasingly used to identify patients at risk for CMV disease and for monitoring antiviral treatment. In addition, a sensitive PCR assay was recently described for the detection and quantitation of CMV DNA in human fecal specimens, which could well prove a particular promising tool for identifying CMV replication in IBD patients. 39 In severely ill IBD patients, endoscopy with biopsies is a high risk and poorly tolerated invasive procedure. Moreover, it has been suggested that CMV involvement occurs more frequent in the right colon, 40 making endoscopy an unattractive candidate for CMV screening. This makes a strong case for introducing CMV PCR in fecal samples as detection method, and indeed, the first results in a cohort of steroid-refractory ulcerative colitis patients confirmed its clinical use. 41 However, if CMV PCR is not available, an alternative would be to biopsy the distal colon to evaluate the presence of CMV.
In conclusion, although CMV causes significant clinical morbidity in IBD patients, no significant progress was made toward unraveling its pathogenic role. This is mainly due to the observational nature of research that has been produced during the last decades. Based on the available literature, we have proposed a model for the sequence of events during bowel inflammation in seropositive patients. Prospective studies in large cohorts of IBD patients are needed to identify patients at risk for CMV disease. Controlled studies using antiviral therapy in active ulcerative colitis with signs of CMV infection are currently underway.