Monoclonal antibodies targeting the immune regulatory “checkpoint” receptors programmed death receptor 1 (PD-1), its ligand, PD-L1, and cytotoxic T cell associated antigen 4 (CTLA-4) have demonstrated remarkable efficacy against advanced cancers [1]. The natural roles of these immune checkpoints include preventing T cell over-activation, inducing anergy, maintaining peripheral immune tolerance, and contributing to T cell exhaustion in local inflammatory environments [2, 3]. By inhibiting these regulatory receptors, immune checkpoint inhibitors (ICIs) boost the anti-tumor effector functions of T cells [2]. Simultaneously, ICI-induced loss of tolerance leads to off-target immune related adverse events (irAEs) [4]. The frequency and severity of these events are an important limitation of immunotherapy, leading to treatment interruption and even discontinuation [5].

Immune related (entero) colitis (irEC) is among the most common severe irAEs leading to the discontinuation of immunotherapy [6–9]. High dose glucocorticoids are the first line management of immune mediated enterocolitis [10–13]. Current guidelines recommend continuation for at least 4 to 6 weeks after resolution of irEC [10–13]. Not infrequently, patients are intolerant to steroid tapers or require prolonged courses of glucocorticoids for symptom control. Prolonged use of glucocorticoids is associated with multiple complications that include serious infections, hyperglycemia, osteoporosis, and altered mental status. In patients with irEC who fail steroid tapers or are steroid unresponsive, inhibitors of tumor necrosis factor alpha (anti-TNFα) have been used to suppress mucosal inflammation [14, 15]. Retrospective studies of patients who received anti-TNFα agents for irEC have shown that compared to steroids, these agents lead to more rapid symptomatic improvement and a shorter duration of steroids without affecting time to ICI treatment failure or overall survival [6, 14–16]. For patients who require anti-TNFα agents to control irEC, ICIs are generally discontinued out of concern for recurrent, potentially treatment refractory, colitis. Recent work in mouse models of cancer has shown that co-administration of ICI and anti-TNFα upfront led to improved tumor responses and decreased colitis severity, an approach that has not yet be studied in prospective trials in humans [17, 18]. In this report, we present our institutional experience on patients with different malignancies who were treated concurrently with anti-TNFα and single or combination ICIs.

Included in this series are patients evaluated and treated at the Massachusetts General Hospital who were referred to the gastroenterology service for new gastrointestinal complaints (abdominal pain, protracted nausea and vomiting, or diarrhea) that arose during treatment with one or more ICIs. ICIs were administered as standard-of-care or as a part of a clinical trial. Details of the patients’ medical history, malignancy, and prior cancer therapeutics were reviewed in the chart. Data pertaining to ICI use and irEC development and management include doses of ICI, presenting grades of diarrhea and colitis, glucocorticoid dose and number of steroid taper attempts, infliximab dose and frequency, doses of infliximab to clinical remission, and doses of ICI concurrently administered with infliximab. Diarrhea and colitis were graded using the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 at disease onset [19]. Upper endoscopy and/or colonoscopy/flexible sigmoidoscopy were performed as clinically indicated at the Massachusetts General Hospital Endoscopy Unit. Consent was obtained from all patients. Pathology was reviewed by board certified pathologists. Radiologic imaging data were obtained as indicated by the treatment protocol. This retrospective study was approved by the institutional review board of the Massachusetts General Hospital.

The biological rationale for combining anti-TNFα therapies with ICIs comes from recent insights into the role of TNFα in tumor immunology. TNFα produced in the setting of anti-PD-1 blockage leads to impaired CD8+ tumor infiltrating T lymphocyte responses [17]. Additionally, TNFα increases activation-induced cell death in T cells, limiting their viability in the tumor microenvironment [18, 20]. In a mouse model of melanoma, concurrent treatment with anti-PD-1 and anti-TNFα led to improved anti-tumor responses [17]. More recently, in a murine colon cancer model, concurrent treatment with anti-TNFα and combined anti-CTLA-4 and anti-PD-1 improved survival when compared with double checkpoint inhibition treatment alone [18]. When colitis was concomitantly induced in the tumor bearing mice through dextran sodium sulfate (DSS), the mice that received anti-TNFα and double checkpoint inhibition had better colitis amelioration and improved overall survival [18]. By blocking TNFα both studies showed an increase in CD8+ T cells numbers and viability in the tumor microenvironment and draining lymph nodes [17, 18]. These findings add to a growing body of literature that implicates innate inflammation in tumor promotion [21–23].

TNFα plays as essential role in the pathogenesis of irEC. Patients with irEC have up-regulated mucosal TNFα and a local activation of the TNFα gene signature [18]. Additionally, mucosal TNFα levels predict irEC steroid responsiveness with higher mucosal TNFα levels predicting lower steroid responsiveness [24]. Multiple groups have previously reported on treatment of severe irEC with TNFα blokecrs [8, 14, 15, 25]. However, after receiving anti-TNFα, patients’ ICI therapy was usually discontinued. In this case series, we report our institutional experience with patients who received immunotherapy and anti-TNFα concurrently.

Patients 2 and 4 were both treated with a combination of ipilimumab and nivolumab while Patients 1, 3, and 5 received monotherapy with pembrolizumab, ipilimumab, or cemiplimab, respectively. Patients 2, 3, and 4 all received a different ICI months to years prior to the regimen immediately associated with colitis. Given that the irAE to some ICI may manifest up to 2 years after therapy [26], the irEC they experienced may have a mixed component due to previous exposure to multiple ICIs. The onset of symptoms was sooner and the severity was worse in the patients who received a combination of ipilimumab and nivolumab compared to those receiving a single agent ICI, consistent with findings reported in the literature previously [6, 26].

Patient 1 had a history of segmental colitis associated with diverticulitis (SCAD) prior to initiation of ICI. Overlapping features of ICI and SCAD were seen on the colonic biopsy. The pathophysiology of SCAD is incompletely understood but the syndrome is thought to overlap with IBD [27]. Some retrospective studies reported increased risk of irEC in patients with baseline active inflammatory and autoimmune diseases, placing the patient at a higher risk of irEC [7, 28].

On presentation with symptoms, all patients were initiated on glucocorticoid therapy for irEC and achieved good control. Patients 1, 3, 4, and 5 were unable to maintain symptom control with glucocorticoid tapering. The decision to resume ICI after irEC carries a serious risk of relapse and is often done on an individual basis [29]. Some studies have shown that 50–60% of these patients have relapse of irEC [14, 29]. In one study, after irEC resolution, ICIs were restarted concomitantly with vedolizumab, an α4β7 integrin inhibitor that blocks T cell trafficking to the gut, only one out of eight patients had recurrence of irEC [14]. Although vedolizumab is a reasonable approach for treatment of glucocorticoid refractory irEC, inhibition of T cell trafficking into the gut may be risky in patients with gastrointestinal malignancies (e.g. patient 2) where antitumor T cells would also require access to the gastrointestinal mucosa. Similarly, vedolizumab could inhibit responses to gastrointestinal metastases, which are found in approximately 5% of patients with melanoma and are often not seen on surveillance imaging [30, 31]. We favor infliximab as initial biologic therapy for irEC for these reasons, as well as the potential antitumor benefit associated with TNFα blockade.

The decision to initiate concurrent ICI and anti-TNFα therapy in our cohort was driven by an inability to taper steroids and concern for irEC recurrence. Infliximab was administered at a dose of 5 mg/kg. The frequency of infliximab infusions varied depending on ICI regimen and irEC severity. In general, patients were loaded on infliximab using standard dosing at weeks 0, 2, and 6. Maintenance therapy frequency was selected based on the assumption that patients receiving ongoing immunotherapy in the setting of irEC would behave like patients with severe IBD, and that they may require more frequent infliximab administration than the standard 8 week interval. In addition, for patient convenience, infliximab was infused on the same schedule as the immunotherapy, with infusions occuring on the same day, though the drugs were not given simultaneously.

After initiation of concurrent anti-TNFα and ICI therapy, all patients continued to receive ICI without recurrence of symptoms. Follow up endoscopies in Patients 1, 2, and 4 showed resolution of acute inflammatory features. Staging scans in Patients 1, 2, 3, and 5 after concurrent therapy demonstrated overall disease stability. Patient 4 had progression of his disease despite targeted therapy, immunotherapy, and radiation. He is currently being considered for surgical management for symptom control. After receiving anti-TNFα and 12 further doses of ICI with no disease progression, Patient 1 developed a Clostridium difficile infection after which he had recurrence of irEC. After treatment with immunotherapies, gastrointestinal disruption (e.g., due to infection) may cause alteration in the gut microbiota and the local immune compartment, resulting in a breach in tolerance leading to irEC [4, 7].

Our experience adds to the growing animal literature showing that concurrent anti-TNFα and ICI therapy is safe, does not negatively impact tumor control and is associated with a better side effect profile. Importantly, patients were able to continue on immunotherapy. We suggest that concurrent anti-TNFα be considered in patients who develop severe irECs early in their immunotherapy course where additional immunotherapy is likely to provide a benefit, and where other treatment alternatives are either unavailable or have a low likelihood of providing benefit. Prospective data will be necessary, however, to clearly define populations where anti-TNFα concurrent with immunotherapy is both safe, and leads to improved tumor outcomes.

An ongoing phase I clinical trial (NCT03293784) is evaluating the safety and tolerability of treating metastatic melanoma with ICIs combined with either infliximab or certolizumab, a similar anti-TNFα agent. Our experience helps provide evidence of the safety of combination treatment with anti-TNFα and ICI, which we propose should accelerate the initiation of a phase II clinical trial to examine the impact of TNFα blockade on both irAEs and antitumor immunity.