Cytomegalovirus in Steroid-Refractory Immune Checkpoint Inhibition–Related Colitis

Monoclonal antibodies blocking cytotoxic T-lymphocyte associated antigen 4 (CTLA4) and programmed death 1 (PD-1) and its ligand programmed death ligand 1, so-called immune checkpoint inhibitors (ICPs), have become important treatment strategies not only for several metastatic cancers (including NSCLC) but also for melanoma and renal cell cancer.1Haanen J.B.A.G. Carbonnel F. Robert C. et al.Management of toxicities from immunotherapy: ESMO clinical practice guidelines for diagnosis, treatment and follow-up.Ann Oncol. 2017; 28: iv119-iv142Abstract Full Text Full Text PDF PubMed Scopus (1302) Google Scholar However, ICPs can lead to immune-related adverse events such as diarrhea (27%–54% with anti-CTLA4 treatment) and colitis (8%–22% with anti-CTLA4 treatment and 1%–2% with anti–PD-1 treatment).1Haanen J.B.A.G. Carbonnel F. Robert C. et al.Management of toxicities from immunotherapy: ESMO clinical practice guidelines for diagnosis, treatment and follow-up.Ann Oncol. 2017; 28: iv119-iv142Abstract Full Text Full Text PDF PubMed Scopus (1302) Google Scholar, 2Larkin J. Hodi F.S. Wolchok J.D. Combined nivolumab and ipilimumab or monotherapy in untreated melanoma.N Engl J Med. 2015; 373 (1270–1201)Crossref Scopus (5110) Google Scholar, 3Robert C. Ribas A. Wolchok J.D. Anti-programmed-death-receptor-1 treatment with pembrolizumab in ipilimumab-refractory advanced melanoma: a randomised dose-comparison cohort of a phase 1 trial.Lancet. 2014; 384: 1109-1117Abstract Full Text Full Text PDF PubMed Scopus (1432) Google Scholar Such adverse events are most frequently induced by combination therapy.1Haanen J.B.A.G. Carbonnel F. Robert C. et al.Management of toxicities from immunotherapy: ESMO clinical practice guidelines for diagnosis, treatment and follow-up.Ann Oncol. 2017; 28: iv119-iv142Abstract Full Text Full Text PDF PubMed Scopus (1302) Google Scholar Specifically, in advanced NSCLC, a meta-analysis on 2722 patients treated with PD-1/programmed death ligand 1 inhibitors showed immune checkpoint inhibition–related colitis (IRC) in 0.4%.4Luo W. Wang Z. Tian P. Li W. Safety and tolerability of PD-1/PD-L1 inhibitors in the treatment of non-small cell lung cancer: a meta-analysis of randomized controlled trials.J Cancer Res Clin Oncol. 2018; 144: 1851-1859Crossref PubMed Scopus (37) Google Scholar The rate of radiological signs of colitis may be as high as 11%.5Alessandrino F. Sahu S. Nishino M. et al.Frequency and imaging features of abdominal immune-related adverse events in metastatic lung cancer patients treated with PD-1 inhibitor.Abdom Radiol. 2019; 44: 1917-1927Crossref PubMed Scopus (25) Google Scholar When IRC (Common Terminology Criteria for Adverse Events grade ≥2) occurs, ICPs are discontinued and administration of immunosuppressants, especially prednisolone or tumor necrosis factor-α antibodies (e.g., infliximab), is started. In therapy-resistant colitis tacrolimus, mycophenolate mofetil or vedolizumab have also been used.1Haanen J.B.A.G. Carbonnel F. Robert C. et al.Management of toxicities from immunotherapy: ESMO clinical practice guidelines for diagnosis, treatment and follow-up.Ann Oncol. 2017; 28: iv119-iv142Abstract Full Text Full Text PDF PubMed Scopus (1302) Google Scholar, 6Bergqvist V. Hertervig E. Gedeon P. et al.Vedolizumab treatment for immune checkpoint inhibitor-induced enterocolitis.Cancer Immunol Immunother. 2017; 66: 581-592Crossref PubMed Scopus (149) Google Scholar Cytomegalovirus (CMV) can be involved in steroid-refractory colitis in patients with inflammatory bowel disease (IBD),7Kim Y.S. Kim Y.H. Kim S.J. et al.The prevalence and efficacy of ganciclovir on steroid-refractory ulcerative colitis with cytomegalovirus infection: a prospective multicenter study.J Clin Gastroenterol. 2012; 46: 51-56Crossref PubMed Scopus (71) Google Scholar as well as in steroid-refractory cases of IRC.8Lankes K. Hundorfeanb G. Harrerc T. et al.Anti-TNF-refractory colitis after checkpoint inhibitor therapy: possible role of CMV mediated immunopathogenesis.Oncoimmunology. 2016; 5e1128611Crossref PubMed Scopus (50) Google Scholar, 9Franklin C. Rooms I. Fiedler M. et al.Cytomegalovirus reactivation in patients with refractory checkpoint inhibitor-induced colitis.Eur J Cancer. 2017; 86: 248e256Abstract Full Text Full Text PDF Scopus (44) Google Scholar However, the currently available literature on the latter is scarce. We treated two patients with CMV infection who had steroid-refractory IRC and discuss the diagnostics, clinical relevance, and treatment of CMV in IRC. Patient A is a 73-year-old female with stage IV melanoma treated with ipilimumab (3 mg/kg) and nivolumab (1 mg/kg) three times weekly. After two cycles she experienced development of diarrhea with hematochezia. Stool cultures were negative. Colonoscopy showed a mild, nonulcerative, nonspecific colitis of the entire colon, including the rectum. Histological examination showed inflammatory changes compatible with IRC (Fig. 1A). The result of immunohistochemistry (IHC) staining for CMV, performed retrospectively, was negative. Administration of prednisolone (1 mg/kg body weight/d) was started and stool frequency decreased. When prednisolone was tapered, diarrhea increased (10–12 times daily). Infliximab was administered (first dose 5 mg/kg, second dose 10 mg/kg) along with prednisolone, 1 mg/kg, without improvement of diarrhea. Sigmoidoscopy showed a continuous severe nonspecific ulcerative colitis (Fig. 2A). Histological examination revealed more severe inflammation with many CMV-positive cells on IHC (Fig. 1B–D). A test for fecal CMV DNA was positive. Treatment with ganciclovir (5 mg/kg body weight intravenously every 12 hours) was started and the dose of prednisolone was decreased to 0.5 mg/kg. One week later there was no significant improvement in clinical and endoscopic severity of colitis (Fig. 2B). Fecal cultures remained negative for CMV. The serum CMV viral load was 42,000 copies per milliliter. The established treatment was continued, and semi-elemental tube feeding was added. No significant improvement occurred, and a new endoscopy was performed 10 days after the last. Here, the colitis appeared worse on histological examination, but fewer CMV-positive cells were seen. The serum CMV load decreased to fewer than 500 copies per milliliter. Treatment with ganciclovir was continued and the dose of prednisolone was increased to 2 mg/kg. After 4 weeks of treatment, ganciclovir was changed to prophylactic valganciclovir (900 mg/d). A new blood sample showed undetectable CMV levels. Sigmoidoscopy and histological examination showed less severe colitis, including fewer CMV-positive cells. Stool frequency finally improved to fewer than four times daily about 6 weeks after the start of administration of prednisolone at a dose of 2 mg/kg. Prednisolone was slowly tapered and stopped without redevelopment of diarrhea. The patient achieved a durable partial tumor response.Figure 2Endoscopic images of patients A and B with immune checkpoint inhibition–related colitis. (A) Distal aspect of the colon in patient A with immune checkpoint inhibition–related colitis after treatment with prednisolone and infliximab. The colon mucosa is severely inflamed, with erythema, edema, vulnerable mucosa, absent vascularity, and ulcerations. (B) Distal colon in patient A after 1 week of treatment with ganciclovir for cytomegalovirus infection. No improvement is seen; in fact, ulcerations seems a bit worse. (C) Proctum in patient B, who experienced development of diarrhea after two cycles of nivolumab. No normal vasculature is seen; the mucosa is edematous and lightly erythematous, with some erosions. Endoscopy was performed 6 days after treatment with prednisolone. (D) Colon of patient B after treatment with prednisolone and infliximab; severe colitis (erythema, edema, and no vasculature with superficial white mucosal coating) without signs of recovery.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 3Images from histological examination of patient B with immune checkpoint inhibition–related colitis. (A) Biopsy from the first endoscopy of patient B (hematoxylin and eosin; original magnification, ×20) shows inflammation typical of immune checkpoint inhibition–related colitis (i.e., increased cellularity of the lamina propria combined with active inflammation (cryptitis). There was also a minimal patchy increase in intraepithelial lymphocytes. Immunohistochemistry performed retrospectively on the biopsy samples from the first endoscopy were negative for CMV (data not shown). (B) Biopsy sample from the second endoscopy (hematoxylin and eosin; original magnification, ×20) shows similar to slightly more pronounced inflammation. (C) Cells with viral/cytopathic changes (hematoxylin and eosin; original magnification, ×60) are inconspicuous (arrow) and mainly found within the lamina muscularis mucosae and submucosa, whereas only a few are present in the lamina propria of the mucosa (not shown). (D) Immunohistochemistry for CMV (hematoxylin and eosin; original magnification, ×20) demonstrated the area with the most staining cells. Of 10 biopsy samples, one had 38 cells positive for CMV, four had one or two cells positive for CMV, and five had no cells positive for CMV. The highest count in a single high-power field (one HPF = 0.21 mm2) was 24 CMV-positive cells.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Patient B is a 54-year-old female with unresectable adenocarcinoma of the lung treated with nivolumab (flat dose of 240 mg twice weekly). Five days after the second dose she experienced development of diarrhea 10 to 14 times daily. Stool cultures were negative for CMV. Initially, the patient's complaints improved as a result of administration of prednisolone at a dose of 1 mg/kg. Colonoscopy performed 6 days after she started taking prednisolone showed a mild proctitis with some miliary erosions up to 20 cm from the anal verge (Fig. 2C). The results of histological examination were consistent with IRC (see Fig. 3A). The dose of prednisolone was decreased to 30 mg after 1 week of treatment, after which the stool frequency increased to 12 times per day. Infliximab, 5 mg/kg, was added to prednisolone, 60 mg. Initially, the stool frequency decreased, but it started to increase again 5 days later. Sigmoidoscopy showed a diffuse nonulcerative colitis with a white cover up to 20 cm (proximal border not seen [Fig. 2D]). Histological examination then showed more severe colitis without CMV (see Fig. 3B–D). Stool cultures remained negative for CMV. Infliximab, 10 mg/kg, provided no persistent improvement and prednisolone was increased to 2 mg/kg. Again, diarrhea improved for a few days and then increased. A new sigmoidoscopy (about 1 month after the previous one) showed an erosive and ulcerative colitis of the left colon (proximal border not seen). Again, histological examination showed severe active inflammation, but now with signs of chronicity and some CMV-positive cells on IHC. The CMV viral load in the serum was 356 copies per milliliter. The low level of serum CMV DNA and low number of CMV-positive epithelial cells were not considered significant, and antiviral therapy was not administered. Because of steroid- and infliximab-refractory colitis, administration of tacrolimus was initiated (4 mg twice a day) and diarrhea improved over the next days. When tacrolimus was tapered and stopped, diarrhea returned. After the patient started taking tacrolimus and prednisolone again, her stool became normal once again. Next, prednisolone was tapered first. After tapering of the tacrolimus, her stool remained normal. In summary, we have presented two cases of steroid- and infliximab-refractory IRC in which CMV was detected but was considered to play different roles. Colon biopsy samples from patient A demonstrated a large number of CMV-positive cells and the serum contained a high viral load, which was considered relevant CMV disease. After initiation of antiviral therapy and tapering of the steroids there was no clinical improvement. Improvement did not occur until the patient was treated with both antiviral therapy and a high dose of steroids. This was thought to imply that the IRC predominated over the CMV colitis, although both were considered important. In contrast, histologically, patient B had a much lower number of CMV-positive cells, which coincided with low serum viral load. In this case, CMV was not considered to be clinically relevant. Here, clinical symptoms improved without antiviral therapy while the immunosuppressant therapy was continued and was even intensified to obtain clinical remission. The pathogenic role of CMV in colitis remains debatable.10Magro F. Gionchetti P. Eliakim R. et al.Third European evidence-based consensus on diagnosis and management of ulcerative colitis. Part 1: definitions, diagnosis, extra-intestinal manifestations, pregnancy, cancer surveillance, surgery, and ileo-anal pouch disorders.J Crohns Colitis. 2017; : 649-670Crossref PubMed Scopus (995) Google Scholar, 11Shukla T. Singh S. Loftus E.V. Bruining D.H. McCurdy J.D. Antiviral therapy in steroid-refractory ulcerative colitis with cytomegalovirus: systematic review and meta-analysis.Inflamm Bowel Dis. 2015; 21: 2718-2725Crossref PubMed Scopus (62) Google Scholar In ulcerative colitis CMV may be an innocent bystander and self-limiting without antiviral therapy.12Criscuolia V. Cas A. Orlandoa A. et al.Severe acute colitis associated with CMV: a prevalence study.Dig Liv Dis. 2004; 36: 818-820Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar, 13Rahier J.F. Magrob F. Abreue C. et al.Second European evidence-based consensus on the prevention, diagnosis and management of opportunistic infections in inflammatory bowel disease.J Crohns Colitis. 2014; 8: 443-468Abstract Full Text Full Text PDF PubMed Scopus (707) Google Scholar In refractory ulcerative colitis, however, CMV is associated with more frequent disease relapse and even colectomy.14Kim Y.S. Kim Y.H. Kim J.S. et al.Long-term outcomes of cytomegalovirus reactivation in patients with moderate to severe ulcerative colitis: a multicenter study.Gut Liver. 2014; 8: 643-647Crossref PubMed Scopus (31) Google Scholar Another difficulty is determining when CMV infection should be regarded as clinically relevant. IHC performed on colon biopsy samples is considered highly sensitive and specific in IBD.13Rahier J.F. Magrob F. Abreue C. et al.Second European evidence-based consensus on the prevention, diagnosis and management of opportunistic infections in inflammatory bowel disease.J Crohns Colitis. 2014; 8: 443-468Abstract Full Text Full Text PDF PubMed Scopus (707) Google Scholar However, the presence of CMV in tissue biopsy samples detected only by polymerase chain reaction is probably less relevant, as patients without histological signs of CMV colitis (hematoxylin and eosin [HE] staining and/or IHC) can recover while receiving steroids without antiviral therapy.7Kim Y.S. Kim Y.H. Kim S.J. et al.The prevalence and efficacy of ganciclovir on steroid-refractory ulcerative colitis with cytomegalovirus infection: a prospective multicenter study.J Clin Gastroenterol. 2012; 46: 51-56Crossref PubMed Scopus (71) Google Scholar Wethkamp et al. suggested that 600 CMV copies per 105 cells (i.e., from biopsy samples) could discriminate clinically relevant (i.e., response to antiviral therapy) from latent infection in patients with IBD.15Wethkamp N. Nordlohne E. Meister V. Helwig U. Respondek M. Identification of clinically relevant cytomegalovirus infections in patients with inflammatory bowel disease.Mod Pathol. 2018; 31: 527-538Crossref PubMed Scopus (9) Google Scholar Another study showed that more than 250 copies per milligram of biopsy sample was associated with steroid-refractory colitis.16Roblin X. Pillet S. Oussalah A. et al.Cytomegalovirus load in inflamed intestinal tissue is predictive of resistance to immunosuppressive therapy in ulcerative colitis.Am J Gastroenterol. 2011; 106: 2001-2008Crossref PubMed Scopus (146) Google Scholar Antiviral therapy might be reserved for patients with detectable CMV in biopsy samples, as in a previous study CMV disappeared from the serum without antiviral intervention in almost all cases with HE-negative biopsy results.17Matsuoka K. Iwao Y. Mori T. et al.Cytomegalovirus is frequently reactivated and disappears without antiviral agents in ulcerative colitis patients.Am J Gastroenterol. 2007; 102: 331-337Crossref PubMed Scopus (167) Google Scholar Serum CMV could be used for monitoring treatment, as was done in the case of patient A. The clinical course is another factor to take into consideration, as 79% of patients with steroid-refractory IBD with CMV (11 of 14) improved with ganciclovir, whereas 55% of patients responsive to steroids (13 of 31) did not require antiviral therapy.7Kim Y.S. Kim Y.H. Kim S.J. et al.The prevalence and efficacy of ganciclovir on steroid-refractory ulcerative colitis with cytomegalovirus infection: a prospective multicenter study.J Clin Gastroenterol. 2012; 46: 51-56Crossref PubMed Scopus (71) Google Scholar In IBD, it has been suggested that a combination of clinical response to steroids, presence of five or more inclusions per colon biopsy sample (HE staining or IHC), and serum viral load be used to guide decision making in treatment of CMV colitis.18Beswick L. Ye B. van Langenberg D.R. Toward an algorithm for the diagnosis and management of CMV in patients with colitis.Inflamm Bowel Dis. 2016; 22: 2966-2976Crossref PubMed Scopus (31) Google Scholar The next question is whether the immunosuppressant medication should be tapered when antiviral therapy is started. The European Crohn's and Colitis Organization advises that discontinuation of immunomodulators should be considered in cases of a severe steroid-resistant colitis with CMV and states that immunomodulators should be discontinued when CMV has turned into systemic disease.13Rahier J.F. Magrob F. Abreue C. et al.Second European evidence-based consensus on the prevention, diagnosis and management of opportunistic infections in inflammatory bowel disease.J Crohns Colitis. 2014; 8: 443-468Abstract Full Text Full Text PDF PubMed Scopus (707) Google Scholar Data on the role of CMV in steroid-refractory IRC are very scarce. One reported case concerns a steroid- and infliximab-refractory colitis in which CMV was detected by IHC and polymerase chain reaction on colon biopsy samples. Serum was positive for CMV as well. This patient required ganciclovir for a prolonged period while steroids were tapered.8Lankes K. Hundorfeanb G. Harrerc T. et al.Anti-TNF-refractory colitis after checkpoint inhibitor therapy: possible role of CMV mediated immunopathogenesis.Oncoimmunology. 2016; 5e1128611Crossref PubMed Scopus (50) Google Scholar Franklin et al.9Franklin C. Rooms I. Fiedler M. et al.Cytomegalovirus reactivation in patients with refractory checkpoint inhibitor-induced colitis.Eur J Cancer. 2017; 86: 248e256Abstract Full Text Full Text PDF Scopus (44) Google Scholar studied 370 patients treated with an ICP; 41 experienced development of colitis. Five patients were refractory to steroids and infliximab, and CMV was detectable in all five. In refractory cases, CMV DNA in colon biopsy samples was more often detectable in refractory cases (four of five [80%] versus one of 15 [7%]), as was CMV DNA in serum (four of five [80%] versus none of 12 [0%]). The four cases with CMV-positive DNA in biopsy samples were successfully treated with ganciclovir.9Franklin C. Rooms I. Fiedler M. et al.Cytomegalovirus reactivation in patients with refractory checkpoint inhibitor-induced colitis.Eur J Cancer. 2017; 86: 248e256Abstract Full Text Full Text PDF Scopus (44) Google Scholar This seems to be in line with our cases; patient A, who had many CMV-infected cells, required antiviral therapy, whereas patient B, who had a much lower number of CMV-infected cells, did not. In conclusion, CMV can play a role in steroid-refractory IRC. We suggest testing for CMV in steroid-refractory IRC by taking multiple colonic biopsy samples, as the load of CMV-positive cells (based on IHC) is likely to indicate the necessity for antiviral treatment. Serum viral load may be useful in determining effectiveness of antiviral therapy, especially when clinical improvement is lacking. We suggest starting antiviral treatment in steroid-refractory cases with a high number of CMV-positive cells on IHC (the data are insufficient to generate a cutoff point, but it is probably >50 CMV-positive cells in a single biopsy sample) or in patients with a relatively low number of CMV-positive cells on IHC accompanied by a high viral load in the serum. Tapering of immunosuppressants should be considered during active CMV colitis. However, in the case of lack of improvement, persisting IRC should be considered as a cause. Dr. van Turenhout, Dr. Berghuis, and Dr. van Dieren drafted the article and developed the concept of the case report. Dr. Snaebjornsson performed histological analysis of the cases and provided the histological images. Dr. Snaebjornsson, Dr. Wilgenhof, Dr. Burgers, and Dr. Haanen performed critical revision of the article for important intellectual content.