How Could Antibiotics, Probiotics, and Corticoids Modify Microbiota and Its Influence in Cancer Immune Checkpoint Inhibitors: A Review

ABSTRACT

Immunotherapy has become a new paradigm in oncology, improving outcomes for several types of cancer. However, there are some aspects about its management that remain uncertain. One of the key points that needs better understanding is the interaction between immunotherapy and gut microbiome and how modulation of the microbiome might modify the efficacy of immunotherapy. Consequently, the negative impact of systemic antibiotics and corticosteroids on the efficacy of immunotherapy needs to be clarified.

INTRODUCTION

Immune checkpoints (ICP) are modulators of the immune response and have an essential role in self-tolerance. ICP could inhibit T-cell function through several mechanisms, such as inhibition of dendritic cell-mediated T-cell activation (CTLA-4) or inducing T-cell exhaustion at sites where there is ongoing inflammation (PD-1/PD-L1) (1). Some neoplasms use ICP in favor of reducing the antineoplastic effect of the host immune system.

Immune checkpoint inhibitors (ICI) act against specific molecules of the immune system, which have a suppressor function. Furthermore, ICI increase the antineoplastic effect of the host immune system. Treatments with ICI have been shown to improve cancer prognosis in patients with several neoplasms, having become the standard of care in some instances (2–4). Nevertheless, the benefit of ICP is restricted to a limited number of patients, and only 25 to 30% of individuals who receive ICI achieve clinical benefit. Several tumor cells or oncogenic pathway-related factors are associated with favorable outcomes with ICI (2–6). Other host-related or extrinsic factors, such as those that modify the microbiome, might influence overall efficacy of ICI (7).

To establish the influence of the microbiome on the response to ICI, we will briefly review the relationship between the microbiome and ICI efficacy and how altering the microbiome might modulate the efficacy of immunotherapy in cancer patients.

MICROBIOME AND CANCER

The microbiome encompasses all host-associated microorganisms present on epithelial barriers, constituting a dynamic ecosystem (8, 9). It is formed by approximately 10¹⁴ microorganisms, such as bacteria, fungi, viruses, and archaea. It is located in different body surfaces and compartments, with the gastrointestinal (GI) tract being one of its main locations. Its composition is influenced by several factors, such as diet, drug therapies, age, and genetics (8, 9). The immune system at the GI tract works to achieve a mutually beneficial equilibrium between these microbes and the host, involving both the innate and adaptive immune systems. The mechanisms aiming to protect the host from bacterial damage or invasion are counterbalanced by processes that restrain the antibacterial response, preventing collateral damage from inflammation (10). Several studies on animal models found that immunological responses vary depending on the bacterial species present in the gut microbiota (10). In this way, the microbiome contributes to shape and modulate the immune system, exerting an overall profound impact on the peripheral immune system. The microbiome might have a direct role in local and distant carcinogenesis, providing toxic or oncogenic products, or an indirect role by promoting inflammation and/or immunosuppression. Conversely, some bacteria have been found to have distant antitumor activity by potentiating host immune responses (8).

THE MICROBIOME AND IMMUNOTHERAPY IN CANCER PATIENTS

The microbiome apparently influences the efficacy of several antineoplastic therapies, including immunomodulatory agents such as ICI. Recently, a review by Gong et al. reported that several interdependence mechanisms and pathways have been proposed to explain the effect of the microbiome on the efficacy of ICI (11).

Some experiments have found that in mice with altered gut microbiomes, whether by administrating antibiotics or in germfree mice, ICI treatment outcomes are poorer (12, 13). Routy et al. provided an example of the impact of the human microbiome on the effect of ICI. In their experiment, fecal microbiota transplantation (FMT) from cancer patients who responded to ICI, into germfree or antibiotic-treated mice, was found to improve the antitumor effects of PD-1 blockade, as opposed to FMT from nonresponding patients (14).

A relationship between Bacteroides fragilis and the efficacy of anti-CTLA-4 has also been postulated, as B. fragilis usually is overexpressed in ileum, stimulating Th1 cells, after CTLA-4 blockade (15). Similarly, other Bacteroides spp. might induce a Th1 response that contributes to an optimal effect therapeutic from anti-CTLA-4 (16, 17). In a recent review, Ma et al. identified that other species, such as Bifidobacterium longum, Bifidobacterium breve species, and Faecalibacterium spp., have been associated with increased anti-PD1 efficacy (15, 18). Bifidobacterium seems to increase gamma interferon (IFN-γ) and CD8⁺ T cells in tumors, stimulate the maturation of dendritic cells, and increase the amount of tumor-specific cytotoxic T lymphocytes (15, 16). Moreover, several other bacterial species and genera, such as Akkermansia muciniphila, Clostridiales, and Ruminococcaceae, if present in the microbiome, prevent leaky colon and systemic immunosuppression (14) and are related to better outcomes (15, 18).

IMMUNE-RELATED ADVERSE EVENTS, RESPONSE, AND THE MICROBIOME IN PATIENTS RECEIVING ICI

ICI-induced T-cell activation is not tumor specific. ICI-induced immune response may damage host tissues, causing immune-related adverse events (IrAE). These can present as colitis, endocrine dysfunction, skin disorders, pneumonitis, or myocarditis (among others). Most IrAEs are mild to moderate and tend to appear in the first 12 weeks after initiating treatment (although they may occur even 6 months after ICI discontinuation) (1). Some factors, such as therapeutic dosage, the use of combination checkpoint blockade, or microbiome-related factors, suggest an association with an increased risk of IrAE (19).

A correlation between IrAE and ICI efficacy has been postulated. In a secondary analysis of the CA209-003 study, a multicenter, phase 1 trial that included 270 patients with non-small-cell lung cancer (NSCLC), melanoma, or renal cancer, this association was assessed (20). Patients with treatment-related adverse effects had a significantly increased long-term overall survival (OS) compared with patients who did not experience these kind of adverse effects (Δ, 19.8 months; 95% confidence intervals [CI], 13.8 to 26.9 months). In some cases with early discontinuation due to grade 3 adverse events, tumor regression was persistent.

The relationship between IrAE and the microbiome has also been previously explored. Several changes in colonic mucosa after administration of anti-CTLA-4 have been observed (17). Bacteroides spp. are associated with favorable tumor responses to anti-CTLA-4 without significant intercurrent intestinal toxicity (17). Likewise, Dubin et al. proposed the Bacteroidetes phylum as a biomarker correlated with protection against colitis secondary to anti-CTLA-4. The Bacteroidetes phylum was found in colitis-resistant patients and could stimulate T-regulatory cell differentiation and limit inflammation (21). Similar results were proposed by Chaput et al., relating Bacteroidetes with colitis-free patients and low anticancer response (22). Conversely, microbiota with high levels of Firmicutes seems to predispose to a better response but with more risk of colitis (22).

MECHANISMS OF INFLUENCE ON MICROBIOME AND IMMUNOTHERAPY ANTIBIOTICS

Antibiotics.

Antibiotics modify the gastrointestinal microbiome by diminishing bacterial diversity (23). It may take longer than 6 weeks to restore the microbiome depending on the duration, use, and type of antibiotic (4, 22, 23). Antibiotics, as microbiome-modifying agents, are expected to influence the overall efficacy of this type of immunotherapy (7).

Several observational studies and recent reviews suggest that antibiotic use decreases the efficacy of ICI (4, 7, 14, 23–27). For instance, Routy et al. addressed the impact of antibiotics within 2 months before, or 1 month after, the initiation of PD1-PD-L1 monoclonal antibodies. This was based on a cohort of 249 patients with non-small-cell lung cancer (NSCLC) (140 patients), renal cancer (67 patients), or urothelial cancer (42 patients) (14). They found that progression-free survival (PFS) (4.1 versus 3.5 months; P = 0.017) and OS (11.5 versus 20.8 months; P < 0.001) were significantly poorer in patients who had received antibiotics (14). The impact of antibiotic use and ICI efficacy was also explored in a retrospective cohort study including 360 patients with NSCLC and renal cancer from two U.S. centers (23). They reported a significant and independent PFS (1.9 versus 7.4 months; hazard ratio [HR], 3.1; 95% confidence interval [CI], 1.4 to 6.9; P < 0.01) and OS (median 17.3 versus 30.6 months; HR, 3.5; 95% CI, 1.1 to 10.8; P = 0.03) reduction in patients who had received systemic antibiotics. The analysis of antibiotic timing on ICI efficacy concluded that the reduction in ICI efficacy was most marked in those patients who had received antibiotics within 30 days before ICI compared with those receiving antibiotics within 60 days before ICI. The process of microbiome recovery in the latter could explain these differences.

In a recent retrospective observational study, which included 291 patients with advanced NSCLC, melanoma, or renal cancer, it was found that patients who received at least 1 course of systemic antibiotics 2 weeks before or during ICI therapy had lower PFS (3.1 versus 6.3 months; P = 0.003) and OS (10.4 versus 21.7 months; P = 0.002) (4). In the multivariate analysis including other prognostic factors, the authors concluded that the antibiotic received was an independent predictor of shorter PFS and OS. Interestingly, the authors categorized antibiotic use into two groups, a single antibiotic group (up to 7 days of a single antibiotic) or a cumulative antibiotic group (more than one antibiotic or a single antibiotic for more than 7 days), with poorest outcomes in the latter (4).

Recently, Hakozaki et al. observed shortened median PFS (1.2 versus 4.1 months) and OS (7.9 months versus not reached) among 13/90 patients with NSCLC treated with antibiotics prior to nivolumab but without statistically significant differences after multivariate analysis (7). In a retrospective single-center study by Galli et al., overall antibiotic use was not associated with PFS or OS. Nevertheless, when antibiotic exposure was considered by the means of the antibiotic-immunotherapy exposure ratio (AIER), defined as days of antibiotic/days of immunotherapy, high antibiotic exposure was independently associated with decreased PFS and OS (3). A pooled analysis, including 2,740 cancer patients from 19 eligible studies, found a significant decrease in PFS (HR, 1.84; 95% CI, 1.49 to 2.26; P < 0.001) and OS (HR, 2.37; 95% CI, 2.05 to 2.75; P < 0.001) (28). Recently, a pooled analysis including 1,512 patients from a phase II POPLAR and a phase III OAK group (both evaluated atezolizumab versus docetaxel in NSCLC) has been published (29). In the analysis, the effect of antibiotics taken between 30 days before and 30 days after the start of the study was evaluated. They observed unfavorable prognosis (shorter OS but unchanged PFS) in patients receiving antibiotics concomitantly with atezolizumab. Exploratory analysis suggests OS differences were detected across all classes of antibiotics used (29).

Very recently, a systematic review and meta-analysis conducted by Lurienne et al. was published (30). They concluded that the use of antibiotics before or during ICI administration reduces OS outcomes, especially their use shortly before or after starting ICI (30). Table 1 summarizes the key findings of studies assessing the impact of concomitant antibiotic therapy in cancer prognosis in patients treated with ICI.

TABLE 1.

Summary about studies assessing the impact of concomitant antibiotic therapy in cancer prognosis in patients treated with ICIs^a

Reference	No. of patients	Type of cancer	PFS (ATB vs no ATB)	OS (ATB vs no ATB)
14	249	NSCLC, renal, and urothelial	4.1 vs 3.5 mo (P = 0.017)	20.8 vs 11.5 mo (P < 0.001)
24	360	NSCLC and renal	7.4 vs 1.9 mo (P < 0.01)	30.6 vs 17.3 mo (P  = 0.03)
4	291	NSCLC, melanoma, and renal	6.3 vs 3.1 mo (P  = 0.003)	21.7 vs 10.4 mo (P  = 0.002)
7	90	NSCLC	4.1 vs 1.2 mo (P = 0.049)	NR vs 7.9 mo (P  = 0.037)
3	157	NSCLC	P < 0.0001 if high AIER	P = 0.0004 if high AIER
Pooled analysis
28	2,740	Several	P < 0.001	P < 0.001
29	169	NSCLC	No significant differences in PFS	14.1 vs 8.5 mo (P = 0.01)

^aNSCLC, non-small-cell lung cancer; PFS, progression-free survival; ATB, antibiotic; OS, overall survival; AIER, antibiotic-immunotherapy exposure ratio.

Corticosteriods.

Corticosteroids are used to minimize inflammation and decrease immune activity with effect on T-cell function (31). They also affect the microbiome (30). For example, in a paper published in 2015, the authors explained that mouse gut microbiota revealed phylogenetic shifts after dexamethasone treatment. They observed an increase in the Actinobacteria, Bifidobacterium, and Lactobacillus levels compared with controls (32). More research is needed to clearly establish the modification of the microbiota after corticosteroids.

Corticosteroids are used for mild to severe IrAE (19). Moreover, some patients could need corticosteroids during ICI treatment as part of best supportive care (2, 31, 33). However, there is some doubt whether corticosteroids could influence ICI response.

The baseline or early use of corticosteroids at the time of initiating ICI has an influence on ORR, PFS, and OS results (2, 33–38). Curiously, in one study there were no significant differences in overall survival between patients who received steroids to treat IrAE compared with patients who did not develop IrAE (36). It has been postulated that baseline corticosteroids could blunt a proliferative burst of CD8-positive T cells that are otherwise needed for ICI effective response (33). Nevertheless, it should be noted that patients with early use of steroids tend to have worse ECOG, increased numbers of metastases, or metastases in the brain (2), all of which are independent predictors of unfavorable outcomes.

Recently, the concomitant use of ICI with chemotherapy in first-line NSCLC has been approved. It seems that the use of corticosteroids as premedication does not have a detrimental effect. This suggests that corticosteroid dose or duration thresholds are associated with decreased ICI response (31). Another explanation is that the benefit of combination has a synergic effect that offsets the detrimental corticosteroid action (33).

It has been speculated that glucocorticoid-induced tumor necrosis factor receptor-related-protein (GITR) could improve immune response, as it is an anti-glucocorticoid agent (39). In a phase 1/2a clinical trial using human GITR agonist IgG1 monoclonal antibody BMS-986156, which included 292 patients with advanced solid tumors, the researchers tried implementing this protein with dose-escalation in monotherapy and in combination with nivolumab. The study concluded that its efficacy with nivolumab is comparable to historical data and had a manageable safety profile (40).

Probiotics.

The gut microbiome can be altered by probiotics and is defined as a single or combination of bacterial species that, when administered in adequate amounts, confer a health benefit to the host (4, 41). Indeed, several clinical trials exploring the impact of probiotic therapy for cancer patients are currently ongoing (4). Probiotics could enhance the response to immunotherapy by modulating the individual species of the microbiome (18). Additionally, they can modulate production of anti-inflammatory cytokines, which are implicated in carcinogenesis and can activate phagocytes to destroy early-stage cancer cells (42).

Most probiotics used for therapy are lactic acid-producing bacteria, such as Lactobacillus and Bifidobacterium. Other possible probiotics are Streptococcus, Bacillus, Enterococcus, and Saccharomyces spp., with some safety limitations (41). The modulation capacity of Lactobacillus spp. and Bifidobacterium spp. in immunotherapy cancer responses has been observed in several studies (43). Sivan et al. assessed melanoma growth in mice that received a combination of Bifidobacterium with anti-PD-L1 therapy. They observed that this combination abolished tumor outgrowth (43). Lactobacillus spp. and Bifidobacterium spp. have been deeply studied in experimental models and appear to be safe (44, 45). These species produce lactic and acetic acids, decreasing luminal pH, which hampers the overgrowth of other undesired species (45). Furthermore, they increase the expression of tight junction proteins, improving barrier function in the gut (45).

CONCLUSIONS

It has been widely demonstrated that concomitant use of immunotherapy with antibiotics or corticosteroids implies changes in the expected response to ICI. Early use of antibiotics just prior to or immediately after the initiation of ICI has a negative impact on survival. A similar result occurs with the baseline or early use of corticosteroids at the beginning of ICI treatment. Therefore, it is important to make careful use of them and consider the influence that they may have on ICI response. Regarding probiotics, this is an area of potential development that can help modulate the effect of immunotherapy. Their relationship with the gut microbiome has been extensively demonstrated. This link is likely one of the modulating mechanisms of the response to immunotherapy.