‘Apoptotic colopathy’ is an umbrella term signifying a pattern of injury where the gastrointestinal biopsy shows a colitic picture with apoptosis as the predominant histological feature. Although the entities within apoptotic colopathy share a common histological feature— ‘apoptosis’, there is a list of varied clinical differential diagnoses that produce this similar histological pattern of injury. These include graft-versus-host disease, drug-induced injury due to multiple drugs (in particular, mycophenolate mofetil, check point inhibitor therapy and some others), infections (particularly cytomegalovirus, adenovirus and some others), immune disorders and other miscellaneous causes. However, the management of these varied differentials is strikingly different, thus necessitating an algorithmic approach for accurate diagnosis and optimal patient management. A definitive diagnosis requires interpretation of varied histological findings in the appropriate clinical context including clinical history, drug history and laboratory findings. This review will focus on the histopathological findings of varied entities that can manifest as ‘apoptotic colopathy’ on assessment of colonic biopsies.
- surgical pathology
Statistics from Altmetric.com
‘Apoptotic colopathy’ is an umbrella term describing a pattern of injury featuring apoptosis as the predominant histological feature. Although an increase in apoptosis is the common histological thread, there is a list of varied differential diagnoses which can produce this pattern of injury. Given the different management protocols in these varied entities, a pragmatic approach is required for a definitive diagnosis and optimal patient management.
‘Apoptotic colopathy’ on a mucosal biopsy may be easy to miss because apoptotic bodies often require high power for diagnosis as well as a thorough assessment of multiple sections and may be confused for lymphocytes, especially by relatively less experienced eyes. Lymphocytes are usually the size of a red blood cell, are recognised as intact cells and show a uniform size with well-defined nuclei. On the other hand, apoptotic bodies, appear as variably sized bits and pieces of degenerating dust or cellular debris.
This current review will focus on the histological findings of various clinical entities which show ‘apoptosis’ as a common pattern of injury.
Graft-versus-host disease (GVHD)
GVHD involving the gastrointestinal (GI) tract has been reported to complicate haematopoietic stem cell transplant (HSCT) in about 30%–70% of allogeneic bone marrow transplant recipients and about 5%–10% autologous stem cell transplantation and very rarely can be seen after solid organ transplantation or blood transfusion as well.1–3 However, the histopathological features of GI GVHD in all these scenarios are similar. GVHD occurs when the donor T cells recognise recipient antigens and incite an inflammatory response, manifested by target organ injury—particularly in the skin, liver, GI tract and lung.2 4 5 Although traditionally being characterised into an acute and chronic phase, the clinical and pathological distinction between acute and chronic GI-GVHD is not entirely clear and will not be discussed in this review.2 3 6 This review will focus predominantly on histopathological features of colonic GVHD as a potential cause of apoptotic colopathy. Notably, endoscopic findings do not correlate well with histopathological findings and range from apparently normal mucosa to severe ulceration.2
‘Apoptotic colopathy’ or the presence of epithelial cell apoptosis in colonic biopsies is the hallmark histological feature of GI GVHD.2 3 5–8 The epithelial cell apoptosis in GVHD is typically coupled with relatively sparse mononuclear inflammation (figure 1A), although scattered eosinophils and neutrophils may be seen. The apoptotic bodies are found first in deep crypts of the colon (the proliferative compartment), can be pretty impressive and have been referred to as ‘exploding crypt cells’, a term describing intracytoplasmic vacuoles filled with nuclear dust and karyorrhectic nuclear debris (figure 1B).2 3 9 In mild cases, apoptotic bodies may be the only histological clue, while in more severe cases, apoptotic microabscesses (defined as cluster of 5 or more apoptotic bodies), hypereosinophilic crypts (crypts lined by atrophic, thin or hypereosinophilic epithelium), crypt abscesses, loss of crypts, frank epithelial destruction and ulcers are seen (figures 1C–D and 2A–D).8
A grading scheme has been recommended for acute GVHD, as described below.9 10 However, correlation with clinical symptoms and patient outcome is weak and as such not many pathologists actually grade biopsies.2
Grade I—isolated crypt apoptosis, without crypt loss.
Grade 2—crypt apoptosis with loss of isolated crypts.
Grade 3—crypt apoptosis with loss of two or more contiguous crypts.
Grade 4—extensive crypt loss with mucosal ulceration/denudation.
How many apoptotic bodies in a colonic biopsy are considered significant?
In a NIH-sponsored consensus project, the minimum criteria for apoptotic bodies varied from one apoptotic body per biopsy fragment, to total number of apoptotic bodies at least equal to the number of biopsy pieces and to scattered apoptotic bodies in more than one crypt.2 5 In routine clinical practice, most pathologists require the presence of at least 1–2 apoptotic bodies per biopsy fragment for a histological diagnosis of ‘consistent with GVHD’ in the absence of other confounding factors (such as concurrent infections or clinical suspicion of drug reaction).2 5
Since apoptosis due to bowel preparatory regimen may mimic GVHD, isolated colonic surface epithelial apoptotic bodies are typically not regarded as a diagnostic evidence of GVHD.
GVHD may have a patchy distribution, hence, studies have recommended that at least 8 and even up to 20 serial sections should be reviewed to avoid false-negatives.2 5 Ancillary studies, such as the use of immunostains for caspase 3 are not recommended in routine clinical practice.2 5
Naked neuroendocrine cell clusters in the lamina propria is another feature of GVHD. Endocrine cells are typically more resistant than other epithelial cells to the cytotoxic injury in GVHD and become more conspicuous due to loss of proliferating cells in the base of crypts (figure 2C).3 8 11 Studies have also shown that endocrine cell aggregates correlate with the severity of GVHD.8
Endothelial cell injury, manifesting as pericapillary haemorrhage within the lamina propria, has also been reported in GI-GVHD.3 Eosinophils can be seen in GVHD; although when prominent (>15/10 high power field (HPF)) would favour a diagnosis of mycophenolate mofetil (MMF)-induced colonic injury over GVHD (discussed in detail later in this paper).
A NIH Consensus Project has proposed recommendations for biopsy reporting of the final diagnosis categories in patients with provisional diagnosis of GVHD. These include
Not GVHD (biopsies with no evidence of GVHD).
Possible GVHD (evidence of GVHD but other aetiologies are also likely such as cytomegalovirus (CMV) colitis with apoptotic bodies found in the vicinity of CMV inclusions, MMF-associated colitis or clinical features that suggest or favour a drug reaction).
Consistent with GVHD (clear histological evidence of GVHD but with mitigating factors, such as coexisting CMV colitis, yet the presence of abundant apoptotic cells not associated with CMV-infected cells; limited sample; rare apoptotic epithelial cells with no alternative explanation or recent chemotherapy or radiotherapy).
GVHD (unequivocal evidence of GVHD).
A pitfall to keep in mind is that the conditioning regimens given to bone marrow transplantation patients can lead to epithelial cell apoptosis, nuclear atypia and crypt cell degeneration, features that resemble GVHD. However, these features usually resolve by day 20 after transplantation. Thus, a diagnosis of GVHD should be made with caution within the first 20 days after transplantation.12 Also, it is worth remembering that acute GVHD is generally observed during or after stem cell engraftment, which usually occurs after the second to third week post-transplantation.6 13
To summarise, a definitive diagnosis of GI GVHD depends on constellation of clinical, histological and laboratory findings. GVHD is much more common with HSCT than solid organ transplant. If the biopsy shows apoptotic colopathy and the patient has concomitant cutaneous and or liver GVHD, then they would be at considerable risk for developing GI-GVHD as well. Knowledge of the clinical history, including drug history, suspected infections, timing of biopsy after transplantation and biopsy/clinical findings in other organs such as skin and liver is crucial for an accurate diagnosis and avoiding misinterpretation of subtle histopathological findings. Importantly, dual pathology may exist and patient with GI-GVHD are extremely susceptible to CMV infection and hence CMV immunostains are recommended in any case of apoptotic prominence, even in the absence of obvious viral inclusions on routine H&E sections.
Mycophenolate mofetil-induced colitis
‘Apoptotic colopathy’ is one of the patterns of injury that can be seen with ‘Mycophenolate mofetil (MMF/CellCept)-induced colitis’ and can exhibit many overlapping histological features with GVHD.8 14–21
MMF is an immunosuppressive medication and is used to prevent rejection in both solid organ transplants (such as liver, kidney, heart) and bone marrow/peripheral stem cell transplant. It is also used to treat autoimmune and inflammatory diseases, such as psoriasis, lupus, rheumatoid arthritis, autoimmune dermatoses, among others.
MMF-induced colitis usually occurs at a dose ranging between 1000 to 2000 mg per day.17 Bone marrow transplant patients who are receiving MMF for GVHD prophylaxis are at risk for developing both GVHD and MMF colitis, and the need to differentiate between the above two aetiologies is crucial given the differing management protocols. While MMF colitis would require cessation or decreased dose of the drug, GVHD would require increased immunosuppression.
MMF-colitis has been reported to cause prominent crypt apoptosis at the crypt bases, along with lamina propria oedema, mixed lamina propria infiltrate, normal or mild architectural distortion, isolated crypt damage, active colitis with focal ulcers, mild regenerative epithelial changes and possible proximal accentuation of pathological changes; many of these features overlap with GI-GVHD (figure 3A–C).8 14–21
Studies have shown isolated crypt damage (scattered, dilated damaged/degenerated crypts variably lined by atrophic epithelium with eosinophilic luminal debris) to be a feature of MMF colitis (figure 3C), and it remains rather unclear as to how these are different from the ‘hypereosinophilic crypts’ which are reported to be more common in GVHD than MMF colitis.3 However, still in the latter, the lamina propria shows significantly more eosinophils and neutrophils as well as appears to lack apoptotic microabscesses and the aggregates of endocrine cells within lamina propria, when compared with GVHD. On the other hand, features favouring GVHD include endocrine cell aggregates, apoptotic microabscesses, architectural distortion and a relative lack of eosinophilia.8 15 Also, normal biopsy specimens from sites distant from focal lesions (such as focal ulcers seen in MMF colitis) should raise the question of MMF associated colitis rather than GVHD.5 On the contrary, apoptotic bodies in sites other than the colon are suggestive of GVHD.
A recent study has shown that a combination of an eosinophil count of >15 per 10 HPF, in conjunction with a lack of lamina propria endocrine cell aggregates and lack of apoptotic microabscesses can be used by pathologists to help separate MMF from GVHD-induced colitis in routine clinical practice. These three features showed sensitivity, specificity, positive predictive value and a negative predictive value of 76%, 93%, 81% and 90%, respectively by receiver operating characteristic analysis.8
Checkpoint inhibitor (CPI) therapy
Immune CPIs are novel drugs that stimulate an antitumour response by targeting immune cell checkpoints and have been used in patients with advanced malignant melanoma, non-small cell lung carcinoma, renal cell carcinoma, microsatellite unstable colorectal carcinoma and urothelial carcinoma. Some of the approved immune CPIs include CTLA-4 inhibitors like ipilimumab and anti-PD-1 drugs like nivolumab, pembrolizumab and avelumab.22 23
CTLA-4 inhibitors have been reported to cause an ‘autoimmune-like colopathy’ consisting of neutrophilic inflammation (including neutrophilic cryptitis and crypt microabscesses) as well as increased crypt epithelial apoptosis. Dense, predominantly lymphocytic infiltrate in the lamina propria along with frequent plasma cells and eosinophils are also seen, along with a variable increase in intraepithelial lymphocytes (IELs).22 23
Anti-PD-1-induced colitis is reported to exhibit an active colitis with increased apoptosis and crypt atrophy/dropout, somewhat similar to the above-described features of anti-CTLA-4 colitis. The biopsy shows neutrophilic crypt microabscesses/cryptitis, increased crypt epithelial cell apoptosis and presence of crypt atrophy and dropout (figure 3D). Atrophic crypts can show features suggestive of ‘apoptotic microabscesses’, including prominent attenuation of crypt epithelium along with luminal apoptotic debris mixed with inflammatory cells.22 23
These histopathological features are not entirely specific and are commonly seen in colitides of some other aetiologies (mentioned in this review) as well and hence correlation with clinical history (history of malignancy and particularly drug history), knowledge of the above varied CPI therapy-induced histological patterns and clear communication with the clinical team are critical for pathologists to differentiate CPI therapy-induced injury from its histopathological mimics in the GI tract.
Idelalisib, a phosphatidylinositol-3 kinase δ inhibitor, approved for the treatment of patients with relapsed chronic lymphocytic leukaemia/small cell lymphoma, follicular lymphoma and indolent non-Hodgkin's lymphoma, has also been associated with GI side effects.
The colonic mucosal biopsies have been reported to show a ‘triad’ of epithelial cell apoptosis, intraepithelial lymphocytosis and neutrophilic cryptitis. Awareness of the histological features of idelalisib-associated enterocolitis, along with the clinical history of the drug use is essential to distinguish Idelalisib-associated apoptotic colopathy from its potential mimics.22 24 25
Conditioning regimens can be a cause of ‘apoptotic colopathy’ in the early post-transplant period (particularly in the first 20 days after HSCT) and can cause histological features indistinguishable from mild GVHD (such as increased crypt apoptotic bodies, mitotic activity, regenerative changes along with relatively sparse lamina propria inflammatory cells).12 Timing of the biopsy after the transplant is required to make this distinction.
Other miscellaneous drugs/therapy-induced changes
There are other drugs associated with apoptotic colopathy and GVHD-like features (increased crypt apoptosis, cystically dilated crypts lined by flattened and degenerated epithelium and luminal apoptotic debris) and these include antimetabolites (methotrexate, capecitabine) and tumour necrosis alpha inhibitors (etanercept and infliximab). The patients clinically present with transitory watery diarrhoea, which improves on cessation of these drugs.15 26 Radiation injury can also cause prominent apoptotic injury to the epithelium.17
CMV infection is a major cause of morbidity in immunosuppressed patients and also can rarely cause disease in immunocompetent hosts. One of the common causes that may manifest as ‘apoptotic colopathy’ on colonic biopsies, CMV can present alone or as a dual pathology in immunosuppressed patients who may be overtly susceptible to a superimposed CMV infection such as those with acute GVHD or on mycophenolate therapy or idiopathic inflammatory bowel disease (IIBD) patients. The histopathological picture varies—while some immunosuppressed patients may show scattered crypt apoptotic bodies, in association with the viral inclusions and mount only minimal inflammatory response characterised by mild acute neutrophilic inflammation; some others may show severe inflammatory response characterised by frank ulcers (figure 4A and B). Crypt atrophy, characterised by regenerative crypts lined by attenuated epithelium and/or crypt drop out is seen. Mucosal ulcers, if present, are associated with activated endothelial cells, prominent acute inflammation (particularly around blood vessels) and intravascular fibrin thrombi. Viral inclusions are usually identified in the endothelial cells, but may also be identified in glandular epithelial cells or stromal cells and can be sparse. Both intranuclear (usually large, amphophilic surrounded by clearing leading to an owl’s eye appearance) and/or intracytoplasmic (usually small, brightly eosinophilic granules) may be seen (figure 4B). H&E slides usually show typical viral inclusions; however, CMV immunohistochemical stain is considered the ‘gold standard’ for the identification of CMV in tissue sections.3 27 28 Most pathologists, as a standard practice, obtain CMV immunohistochemistry on any endoscopic biopsy in HSCT patient containing gland apoptosis, regardless of the patient’s CMV blood copy number and even if no viral nuclear inclusions are identified on routine H&E stain. Many studies also support that immunohistochemistry facilitates CMV detection in suspicious cases that lack classic viral inclusions or when exuberant inflammation or reactive cellular changes obscure detection of viral inclusions.29 30
Although CMV inclusions can be recognised on histological assessment of H&E slides in many cases, studies have confirmed that detection of CMV DNA via PCR in biopsies is a sensitive method for detection of the organism, although is not routinely used in clinical practice.28 31
CMV infection, as a cause of apoptotic colopathy, may mimic acute GVHD in patients with a history of stem cell transplant. Of course, the presence of viral inclusions confirms CMV infection. But, GVHD and CMV infection can coexist and it may not be easy to separate the effects of each in colonic biopsies. Severe crypt injury and marked apoptosis in a biopsy even away from the areas of scattered rare viral inclusions would favour GVHD as the major cause of mucosal injury and CMV more likely being a superimposed infection.
Only few studies have described adenovirus as a cause of ‘apoptotic colopathy’. Histological findings are variable and include crypt apoptosis, lymphoplasmacytic infiltrate of the lamina propria and acute neutrophilic inflammation within the crypts and lamina propria. Biopsies may also show small tufts of surface epithelial cells showing regenerative nuclear atypia and abundant cellular degeneration. Amphophilic, smudgy appearing intranuclear inclusions may be identified on the H&E stain. Immunostain for adenovirus can highlight scattered adenovirus immunopositive cells typically within or just below the surface epithelium and are occasionally shed into the lumen.3 4
Other miscellaneous infections
Other infections such as cryptosporidiosis and patients with AIDS often show non-specific crypt cell degenerative changes and apoptosis, which can mimic GVHD. The cryptosporidial spores can be recognised lying just above surface or crypt epithelium (figure 4C).
Common variable immunodeficiency (CVID)
CVID results in hypogammaglobulinaemia, reduced ability to produce antibody in response to infections or vaccines and a predilection for chronic/recurrent infections. It can cause wide variety of histological patterns in the GI tract, one of which is apoptotic colopathy. Studies have reported prominent apoptosis in colonic biopsies in almost half of the patients with CVID. The other features reported in the colonic biopsy are a paucity of plasma cells in combination with intraepithelial lymphocytosis that is usually more pronounced in the deep mucosa and the presence of lymphoid aggregates, along with variable foci of architectural distortion. Notably, the presence of plasma cells does not preclude a diagnosis of CVID, as approximately one-third of the biopsy samples may show plasma cells. Neutrophils when present in high numbers would suggest the presence of superimposed infections. When such findings are encountered in biopsies, along with a lack of history of transplantation or drugs, this should alert the pathologist to think about immunodeficiency states.4 32 33
Traditionally considered a disease of the small intestine, autoimmune enteropathy (AIE) can affect other sites in the tubular gut-like colon. AIE is an uncommon disorder exhibiting variable clinical and pathological features and is typically characterised by diarrhoea, villous atrophy in the small intestine, lack of response to dietary restrictions and evidence of autoimmunity.
The histological findings in the colonic mucosa are variable and range from increased crypt epithelial cell apoptosis (with or without inflammation) and is seen in approximately one third of the biopsies. The other histological findings reported are variable foci of chronic active inflammation, characterised by expansion of the lamina propria by a lymphoplasmacytic inflammatory cell infiltrate and Paneth cell metaplasia in the distal colon, neutrophilic cryptitis and crypt microabscesses and increased IELs. Atrophic crypts with luminal debris have also been reported as is crypt architectural distortion with crypt dropout.34
Apoptotic colopathy has also been reported in colonic biopsies from patients with severe T cell deficiency.35
Other miscellaneous and rather uncommon cause includes reports of apoptotic colopathy in patients with malignant thymoma.13 The authors have encountered a case of prominent apoptosis with paucity of goblet cells and increased IELs in a colonic biopsy taken from patient with Stevens-Johnson syndrome (figure 4D). Additionally, many types of bowel preparation formulas may induce crypt cell apoptosis and which can be particularly marked in some individuals without any crypt destruction and is often more prominent in the right colon than in the left colon and rectum.
To summarise, apoptotic colopathy is a histological pattern of injury which can be seen in varied clinical scenarios (table 1). The most important aetiologies seen in clinical practice include GVHD; Drug-induced injury (particularly MMF, CPI therapy and some others), infections (CMV and some others), immune disorders along with some other miscellaneous causes. Although there is an overlap of histological features in these above described entities, still there are some other histomorphological features described in this review that may favour one aetiology over the other (table 1). However, most importantly, an accurate diagnosis would require correlation with clinical history (such as that of stem cell or solid organ transplant or susceptibility to infections or timing of the biopsy), drug history and clinical/endoscopic findings. An open communication between the pathologist and the clinical team is a key to accurate diagnosis and optimal patient management.
Some important aetiologies of apoptotic colopathy include Graft-versus-host disease (GVHD), drug-induced injury (mycophenolate mofetil and some others), infections (CMV and some others), immune disorders along with some other miscellaneous causes.
Histologic features favoring GVHD over Mycophenolate induced colitis include endocrine cell aggregates, apoptotic microabscesses,architectural distortion and a relative lack of eosinophilia (<15 per 10HPF).
Dual pathology may exist and patients with apoptotic colopathy of some other causes(such as GVHD) are susceptible to CMV infection and hence CMV immunostains may be considered in any case of apoptotic prominence.
An accurate diagnosis would require correlation with clinical history (such as that of stem cell or solid organ transplant or susceptibility to infections or timing of the biopsy), drug history, clinical/endoscopic findings, along with an open communication with the clinical team.
Handling editor Cheok Soon Lee.
Contributors The authors participated actively in this manuscript and can take public responsibility for the content of this paper. All authors critically revised the paper for intellectual content and have approved the final version of the manuscript.
Disclaimer The manuscript has not been and will not be submitted elsewhere for publication. The authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Competing interests None declared.
Patient consent Not required.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.