Fungal Infections Associated With the Use of Novel Immunotherapeutic Agents

Abstract

Purpose of the Review

Recent concerns have emerged regarding the potential of immunotherapy to cause infection. In this review, we summarize the current literature on invasive fungal infections that occur during treatment with immune checkpoint inhibitors and chimeric antigen receptor T cell therapy.

Recent Findings

Fungal infections are uncommon with the use of checkpoint inhibitors. Most cases are caused by invasive aspergillosis and pneumocystis pneumonia and occur in patients requiring high dose corticosteroids for the management of immune-related adverse events. Conversely, fungal infections are commonly reported during therapy with CAR T cells. Most cases are caused by invasive aspergillosis and candidiasis and are likely the result of prolonged neutropenia following the conditioning regimen or immunosuppressant use for the management of cytokine release syndrome and neurotoxicity.

Summary

Treatment-related toxicities that require prolonged immunosuppressive agents appear to play a key role in the development of fungal infections during immunotherapy. Ongoing surveillance is needed to fully address the risks of fungal infections with these novel agents.

Introduction

Immunotherapy has revolutionized the treatment of cancer over the past decade. Rather than exerting a direct effect on malignant cells, immunotherapeutic agents target and modulate the immune system to fight cancer, primarily by enhancing T cell-mediated killing. The primary classes of immunotherapeutic agents include immune checkpoint inhibitors (ICIs) and chimeric antigen receptor (CAR) T cells [1, 2].

Despite their established clinical benefit, immunotherapy is often accompanied by significant side effects, including immune-related adverse events (irAEs) and cytokine release syndrome (CRS), which are important limiting factors for their use. The treatment of these side effects often involves the use of immunosuppressants, including high dose corticosteroids [3, 4], thereby increasing the risk for infection. Since the introduction of immunotherapeutic agents into clinical practice, concerns have emerged regarding their potential to cause infection. Conversely, ICIs have also been reported to mitigate infections and are now being studied for use in the treatment of sepsis and chronic infections [5, 6]. Few studies have focused on addressing the risk of fungal infections with the use of ICIs and CAR T cell therapy. In this article, we aim to review the available literature and discuss the burden of fungal infections with the use of these novel agents to treat cancer.

Fungal Infections in Patients with Cancer

The incidence of invasive fungal infections (IFIs) varies widely among different populations and types of cancer and according to the definitions utilized by different retrospective cohorts. Fungal infections occur most commonly in patients with hematologic malignancies and in hematopoietic stem cell transplantation (HSCT), with an overall reported incidence ranging from 1.3–19% for proven or probable IFI [7–11]. On the other hand, the risk appears to be much lower in patients with multiple myeloma or lymphoma, ranging from 0.4 to 3% [12]. The incidence of IFIs is also affected by the use and choice of antifungal prophylaxis during intensive chemotherapy or after HSCT [13, 14]. A recent meta-analysis found that the risk of invasive aspergillosis (IA) in patients with hematologic malignancies was 11% during remission-induction chemotherapy without the use of prophylaxis and 4% when prophylactic agents were used [15]. In contrast to hematologic malignancies, IFIs are relatively uncommon in patients with solid organ malignancies. A retrospective evaluation of postmortem data obtained between 1993 and 2002 in one large cancer center reported an incidence of IA of 0.69% in patients with solid malignancies compared with 17.3% in patients with hematologic malignancies [16]. IFIs are associated with unacceptably high mortality rates. As an example, the mortality attributable to IA can be as high as 18% in patients with solid malignancy and 30% in patients with hematologic malignancies, with some studies reporting even higher values [17].

Aside from the disease-related immunosuppression inherent to certain hematologic malignancies, multiple factors may contribute to increase the susceptibility to fungal infections. These include chemotherapy-associated neutropenia, the use of T cell depleting agents, receipt of hematopoietic stem cell transplantation (HSCT), development of graft-versus-host disease (GvHD), and the use of corticosteroids. Given the high overall and attributable mortality associated with fungal infections, special consideration should be given whenever employing a novel therapeutic modality. While an increased risk may be clearly expected with the use of certain agents (e.g., the anti-CD52 monoclonal antibody alemtuzumab) [18], others (such as the Bruton’s tyrosine kinase inhibitor ibrutinib) may exhibit unexpected effects on fungal immune surveillance [19–21].

Fungal Infections Associated With the Use of Immune Checkpoint Inhibitor Therapy

Immune checkpoints are inhibitory pathways that regulate the adaptive immune response and prevent host tissue destruction by T cells. While critical for maintaining self-tolerance, the downregulation of T cells has the undesired effect of diminishing the immune system’s ability to fight cancer. ICIs act by releasing T cells to exert cytotoxic effect, thereby leading to destruction of malignant cells. Unfortunately, given its lack of specificity, normal host cells often suffer collateral damage during the process. This translates clinically into the well-known irAEs that are often encountered during treatment and most commonly affect the skin, gastrointestinal, endocrine, hepatic, and pulmonary systems. As a result, many patients require the use of high doses of corticosteroids or infliximab in order to control the irAEs [4].

Ipilimumab, a cytotoxic T lymphocyte antigen 4 (CTLA-4) inhibitor, was the first of its class to receive FDA approval in 2011 for the treatment of late-stage melanoma [22]. Shortly thereafter, the programmed death-1 (PD-1) and programmed death-ligand (PD-L1) inhibitors were added to this list, which has now expanded to include seven different agents approved for use in the treatment of different types of cancer, either as monotherapy or combination therapy (i.e., nivolumab and ipilimumab for advanced melanoma and renal cell carcinoma). Multiple clinical trials are ongoing with a goal of expanding the use of ICIs over a wide variety of indications.

The burden of fungal infections on patients receiving ICIs is not defined clearly but appears to be low based on case series conducted to date. Reports of fungal infections emerging during treatment have raised concerns for a possible direct or indirect association to this drug class. Upon review of the available literature, we encountered only 10 case reports of such infections (Table 1) [23–31]. The most common infection was pulmonary aspergillosis with a total of 6 cases, followed by 3 cases of Pneumocystis jirovecii pneumonia (PJP) and one case of Fusarium species lung infection. In most of these cases, infections occurred in the setting of concomitant use of high dose immunosuppressive agents for treatment of irAEs (6) and/or in the presence of lymphopenia (2), both well-known risk factors for fungal infection. As an exception to these observations, Inthasot et al. reported a case of presumed pulmonary invasive aspergillosis (IPA) in a 57-year-old woman treated with nivolumab for non-small cell lung cancer (NSCLC), in which there were no reported irAEs or immunosuppressant use [27]. The authors did not mention whether additional risk factors were present in this case, such as neutropenia or cannabis use. Recently, Si et al. also reported a case of PJP in an 18-year-old woman treated with pembrolizumab for primary mediastinal B cell lymphoma (PMBCL), in which there were no significant exposure to corticosteroids or presence of absolute lymphopenia. The patient was 10 months post autologous stem cell transplantation and had discontinued PJP prophylaxis 4 months prior to development of symptoms [31].

Table 1.

Case reports of fungal infections associated with ICIs

Agent	Malignancy	Infection	Age	Sex	Comorbidities	Prior chemo	IrAE	Steroid use	Cytopenia	Outcome	Ref.
Durvalumab	NSCLC	Pleural aspergillosis	63	M	COPD, DM, HCV	Yes	No	No	Lymphopenia	Survived	[23]
Durvalumab	NSCLC	Pulmonary aspergillosis	68	M		Yes	Possible	Yes		Survived	[24]
Nivolumab	NSCLC	Pulmonary fusariosis	56	M	HTN	Yes	Yes	Yes	Pancytopenia	Death†	[25]
Nivolumab	NSCLC	Pulmonary aspergillosis*	65	M	CPPA, smoking, tuberculosis	Yes	No			Survived	[26]
Nivolumab	NSCLC	Pulmonary aspergillosis	57	F		Yes	No	No		Survived	[27]
Ipilimumab	Melanoma	Pulmonary aspergillosis	68	M		No	Yes	Yes		Death†	[28]
Ipilimumab	Melanoma	PJP	69	F	CLL, low IgG	Yes	Yes	Yes‡			[29]
Ipilimumab	Melanoma	PJP	63	F		No	Yes	Yes			[29]
Pembrolizumab	NSCLC	Pulmonary aspergillosis	62	F		No	No	Yes		Death	[30]
Pembrolizumab	PMBCL	PJP	18	F		Yes	No	No	No	Survived	[31]

^*Exacerbation of CPPA with aspergilloma

^†Death due to underlying malignancy

^‡Concomitant infliximab use

In one of the ten reported cases we analyzed, the patient had a prior diagnosis of chronic progressive pulmonary aspergillosis but developed an acute progression of the fungus ball with associated infiltration while on nivolumab [26]. The authors speculated that the patient developed a nivolumab-induced hyperreaction to the organism, resembling an immune reconstitution syndrome. Similar reactions have been reported during checkpoint inhibitor therapy in the setting of Mycobacterium tuberculosis infection [32–34]. In fact, there is an increasing body of evidence suggesting that reactivation of latent tuberculosis infection may represent a direct complication of therapy [33–35]. PD-1 inhibition promotes an increase in Th1 and Th17 responses while suppressing the Th2 response [36]. This pattern of T helper cell differentiation would be expected to enhance immunity against fungal infections and support the concept that ICIs can paradoxically worsen infection by leading to an exaggerated immune response against fungal organisms, at least in a subset of patients. Extending this concept, PD-1 inhibition has also been reported to potentially trigger bronchopulmonary aspergillosis (ABPA) [37, 38], which is usually associated with an exaggerated Th2 response. However, some studies suggest that Th17-mediated inflammation may play an important role in ABPA [39] and treatment-related enhanced Th17 responses could offer a potential explanation for this scenario.

Interestingly, given its ability to upregulate cell-mediated immunity, ICIs are now being studied for use during sepsis, but available data are limited. Murine [40], in vitro [41] and ex vivo [42] studies of bacterial sepsis, appears to suggest a potential benefit for utilizing PD-1 inhibition during bacterial sepsis. ICIs have also been reported to improve survival in murine models of fungal sepsis [43] and IA [44]. In a recent case report, nivolumab was used in combination with interferon-γ to successfully treat a case of abdominal invasive mucormycosis [45]. This was the first case report of an ICI being used to treat a human IFI, but given the concomitant administration of interferon-γ, treatment success cannot be attributed solely to the use of nivolumab.

The risk for fungal infections with the use of ICIs has been evaluated in two retrospective studies (Table 2). Del Castillo et al. studied 740 patients who received ICIs for the treatment of melanoma. Six fungal infections were reported to occur during the 4-year study period, represented by three cases of PJP, two cases of IPA, and one case of Candida bloodstream infection. The overall risk for infection was low but significantly increased when concomitant use of corticosteroids was required. Among the different ICI treatment regimens utilized, the combination of ipilimumab/nivolumab gave rise to the highest risk of overall infections, which is likely explained by the higher incidence of IrAEs in the combination group [47•]. A more recent retrospective study evaluated the frequency of infections in 167 patients with lung cancer that were treated with nivolumab. Fungal infections developed in 1.2% of patients, represented by one case of probable IPA and one case of Candida esophagitis [46]. The incidence of IA in this study was 0.6%, which is similar to the previously reported postmortem studies of patients with solid organ malignancies [16].

Table 2.

Retrospective studies of fungal infections associated with ICIs

Agent	Malignancy	Infection (s)	n	Patients, No. (%)	Study timing	Ref.
Nivolumab	NSCLC	Total	167	2 (1.2)	Dec 2015-Jun 2017	[46]
		Pulmonary aspergillosis		1 (0.6)
		Candida esophagitis		1 (0.6)
Checkpoint inhibitors*	Melanoma	Total	740	6	Dec 2010-Oct 2014	[47•]
		Invasive aspergillosis		2
		PJP		3
		Candida BSI		1

^*Checkpoint inhibitors included ipilimumab, nivolumab and pembrolizumab monotherapies and the combination of ipilimumab with nivolumab

In conclusion, there appear to be multiple direct or indirect mechanisms by which immune checkpoint inhibitors could lead to the development of IFIs. First, immune-related adverse events often complicate treatment with checkpoint inhibitors and result in the administration of high doses of corticosteroids, which is a well-known risk factor for the development of IFIs. Second, ICI therapy can lead to immune-mediated neutropenia, although this process appears to be rare [25, 48]. Third, it is possible that by activating T cell functions, ICIs can unmask previously unrecognized infections, by restoring or augmenting microbe-elicited inflammation and tissue damage. As clinicians gain further experience with ICI therapy and its applications, it will be essential to conduct ongoing surveillance to fully address the risks of fungal infections with these novel agents.

Fungal Infections Associated With the Use of CAR T Cell Therapy

The development of adoptive cell therapy and CAR T cell therapy for the treatment of cancer has exerted a tremendous impact in the field of immunotherapy. In this treatment modality, T cells are genetically modified to express specific CARs that target different molecules in malignant cells. Initial clinical studies focused mainly on treatment of refractory/relapsed B cell hematologic malignancies by targeting CD19 but have now advanced to include new targets such as CD22 for the treatment of B cell malignancies, BCMA for multiple myeloma, and CD33 and CD123 for myeloid malignancies. Furthermore, clinical trials are now evaluating the role of CAR T cell treatment for solid tumor malignancies [1].

One of the major limitations of CAR T cell treatment is the potential for life-threatening toxicities, which include neurotoxicity and cytokine release syndrome (CRS). The treatment of these toxicities involves the administration of high doses of immunosuppressants, including corticosteroids and tocilizumab, which poses a significant risk for infection [3]. Aside from that, candidates for CAR T cell treatment often have impaired immunity due to their underlying malignancy and are often heavily pre-treated with chemotherapeutic agents. Moreover, many of these candidates have a history of prior HSCT and GvHD, significantly increasing the risk for fungal infection. Although the true impact of CAR T cell treatment on the development of fungal infections is difficult to determine, the use of antifungal prophylactic agents is common in this population, and studies tend to underestimate the risk of infection.

Only a few studies have been designed to examine the risk of infection with the use of CAR T cells, with most involving cohorts from early clinical trials utilizing CD19 CAR T cells (Tables 3 and 4). Park et al. evaluated 53 patients who had received CD19 CAR T cells for the treatment of B-ALL during a 6-year period. Five fungal infections occurred in four patients, corresponding to a 7.5% incidence in the cohort [49•]. Most infections (80%) occurred during the first 30 days after infusion, and most (60%) were caused by probable pulmonary aspergillosis (in accordance with the EORTC/MSG published criteria). Except for one case of proven pulmonary mucormycosis, all other fungal infections occurred while patients were neutropenic.

Table 3.

Early fungal infections associated with CAR T cells in cohort studies

Target	Malignancy	Prophylaxis	Infection (s)	Total no. of patients	IFI, patient no. (%)	Study timing	Ref.
CD-19 19–28z	ALL	Micafungin	Total	53	4 (7.5)	May 2010-Aug 2016	[49•]
			Saccharomyces cerevisiae BSI		1 (1.9)
			Pulmonary aspergillosis*		2 (3.8)
			Pulmonary mucormycosis		1 (1.9)
CD-19	ALL,CLL,NHL	Fluconazole	Total†	133	4 (3)‡	May 2013-Sep 2016	[50]
			Pulmonary aspergillosis*		1 (0.8)
			Sinus mold infection		1 (0.8)
			Candida glabrata BSI		2 (1.5)
			Candida glabrata pneumonia		1 (0.8)
			Candida bracarensis pneumonia		1 (0.8)
Pooled		Voriconazole	Total	109	2 (1.8)	Oct 2017-Jul 2018	[55]
CD19/22	B-cell malignancy		Fungal infection NOS§	84	1 (1.2)
BCMA	MM		–	16	0
CD19/22 post HSCT	B-cell lymphoma		Fungal infection NOS§	9	1 (11.1)
CD19 (CAT)	ALL		Fungal chest infection	14	2 (14.3)	Apr 2016-Dec 2018	[51]
CD19	ALL¶		Total	83	1 (1.2)	2014–2017	[52]
			Cunninghamella pulmonary infection#		1 (1.2)

^*Classified as probable invasive aspergillosis

^†All had severe CRS or neurotoxicity requiring tocilizumab and/or corticosteroids. Three patients were HSCT recipients (autologous or allogeneic)

^‡Six infections occurred in 4 patients

^§NOS not otherwise specified

^¶The majority of patients had ALL with the exception of 1 patient with mixed leukemia and 1 with B-cell lymphoma

^#Unclear if this represented worsening of a previous infection or a new infection

Table 4.

Late fungal infections associated with CAR T cells in cohort studies

Target	Malignancy	Infection (s)	Total no. of patients	Patients with IFI, no. (%)	Study timing	Ref.
CD19	ALL, CLL, NHL	Total* (> 28 days)	119	2 (1.7)	May 2013-Sep 2016	[50]
		Aspergillus fumigatus sinusitis		1 (0.8)
		PJP		1 (0.8)
CD19	ALL, CLL, NHL	Total (≥ 90 days)	54†	4 (7.4)	Jul 2013-Feb 2017	[53]
		Aspergillus		2 (3.7)
		Candida		1 (1.9)
		Coccidioides		1 (1.9)
CD-19 19–28z	ALL	Pulmonary aspergillosis (> 30 days)	32	1 (3.1)	May 2010-Aug 2016	[49]
CD19	ALL	Death from systemic mycosis (> 30 days)	75	1 (1.3)‡	Apr 2015-Apr 2017	[54]
CD19		None	48	0 (0)	2014–2017	[52]

^*Both patients were allogeneic HSCT recipients

^†Number of patients for available data with infection

^‡At least one case of fungal infection reported. Not all infections had their etiology specified

In a separate cohort of 133 patients, IFIs were reported in 5% of patients. Early (≤ 28 days) and late (> 28 days) fungal infections occurred in 3% and 1.7% of patients, respectively, and corresponded to 14% and 8.6% of all infection events [50••]. All patients either had severe CRS or neurotoxicity requiring immunosuppressants or were HSCT recipients. There was only one case of PJP which appeared to have been a breakthrough infection [50••]. Additional studies have reported incidences ranging from 1.2 to 14.3% with the use of CD19 CAR T [51–54]. When utilizing CD19/22-targeted CAR T cells, Luo et al. reported an incidence of 1.2% with CAR T treatment alone and 11.1% when treatment was preceded by HSCT [55]. Overall, it appears that invasive fungal infections were more frequently encountered when micafungin was utilized as antifungal prophylaxis and less frequently with the use of voriconazole. However, the lack of head-to-head comparison between prophylactic regimens does not permit definitive conclusions to be drawn.

Cordeiro et al. analyzed infections that occurred after day 90 post CD19 CAR T cell infusion in 54 patients. In this study, fungal infections represented 2.6% off all infections and included a case of coccidioidomycosis [53]. Another case of disseminated coccidioidomycosis after treatment with CD19 CAR T cells has also been described by Zahid et al. Although confounded by multiple factors, including prior allogeneic HSCT, chronic GvHD of skin and mouth, and use of corticosteroids and rituximab, the authors proposed that the use of anti-CD19 directed therapy could increase the susceptibility to disseminated infection, by depleting B cells and leading to an IgG subclass deficiency [56].

Recently, Haidar et al. [57] reported a 3% incidence of invasive mold infection among CAR T cell recipients at a large academic center, corresponding to a total of 2 cases. The first patient had prolonged neutropenia of more than 3 months duration and developed invasive Fusarium solani infection 5 days post CAR T cell infusion. The second patient developed refractory neutropenia after CAR T cell infusion complicated by invasive fungal sinusitis due to Mucorales, which occurred more than 4 months after CAR T cell infusion.

Data on other types of CAR T cells is still very limited. No fungal infections were reported in early clinical trials with CD30 CAR T cells [58]. In a phase I study of 25 patients who received BCMA CAR T cells for the treatment of relapsed/refractory MM, there was one reported death from candidemia following receipt of steroids for the treatment of CRS [59]. Additional studies have either not reported or not addressed fungal infections with the use of BCMA CAR T cells. A single patient died of pulmonary aspergillosis during a trial with CD22 CAR T cells for B-ALL in China. However, in that case, the infection was present prior to enrollment, and death was considered unrelated to CAR T cell toxicity [60].

By depleting B cells, CD19 CAR T cell therapy often lead to hypogammaglobulinemia and could potentially affect antibody-dependent cell-mediated immunity. The importance of B cells in host defense against fungal infections is primarily linked to susceptibility to Pneumocystis. In murine models of PJP, B cells have been shown to facilitate CD4 T cell activation and play an important role in clearing infection [61]. In contrast, humoral immunity does not play a major role in host defense against candidiasis or aspergillosis. Furthermore, the use of the B cell-depleting agent rituximab has been associated with an increased risk of PJP in retrospective studies [62, 63]. However, PJP has not been frequently reported after CAR T cell treatment, perhaps due to widespread use of prophylaxis. Since B cell immunity does not appear to play a major role against invasive mold infections, it is likely that T cell and myeloid cell depletion during CAR T cell conditioning is the main risk factor for invasive candidiasis or invasive mold infections, rather than receipt of the adoptive cell therapy per se. Moreover, late hematologic toxicity has been observed with CAR T cell therapy, and prolonged neutropenia is not uncommon [53]. The neurologic toxicity or cytokine release syndrome related to CAR T therapy often requires the use of high dose corticosteroids. Finally, prior receipt of myelosuppressive or myeloablative chemotherapy and additional immunosuppressive agents, such as alemtuzumab, often lead to impaired cell-mediated immunity. These factors, either alone or in combination, could explain the mechanisms underlying infection during CAR T cell treatment.

Conclusion

Although fungal infections are more commonly reported after CAR T cell treatment, they are rare with the use of immune checkpoint inhibitor therapy. However, the true risk may be underestimated due to concurrent use of antifungal prophylaxis, in particular with CAR T cell therapy. Additional confounding factors in assessing infectious risk of these therapies include the prior receipt of myelosuppressive or myeloablative chemotherapy, T cell-depleting agents, and corticosteroids. After the initiation of checkpoint inhibitor and CAR T cell therapy, the emergence of significant toxicities that require prolonged immunosuppressive treatment appears to play a key role in the development of fungal infections.