The Use of Phytochemicals to Improve the Efficacy of Immune Checkpoint Inhibitors: Opportunities and Challenges

Abstract

Immune checkpoint inhibitors (ICIs) have revolutionized cancer therapy and reshaped medical oncology practice over the past decade. However, despite unprecedented and durable clinical responses, most patients eventually fail to respond to ICI therapy due to primary or acquired resistance. There is a great need for complementary alternative medicine, such as botanicals and nutritional supplements, because of their capability to modulate a myriad of molecular mechanisms to prevent immunotherapy resistance and reduce its adverse effects. Mounting evidence suggests that phytochemicals, biologically active compounds derived from plants, can favorably regulate key signaling pathways involved in tumor development and progression. In addition, phytochemicals have been found to exert anticancer effects by altering the expression of checkpoint inhibitors of the immune response. The immunomodulatory activity of phytochemicals in the tumor microenvironment has recently received immense interest. Based on these immunomodulatory activities, phytochemicals could be candidates for combination with ICIs in future clinical studies. The current review focuses on the available evidence for combining phytochemicals with a discussion on the promising opportunities to enhance the efficacy of immune checkpoint inhibitors and potential challenges resulting from these combinations.

1. Introduction

Immune checkpoint inhibitors (ICIs) have dramatically changed the oncology landscape and become a vital part of cancer care [1]. ICIs have demonstrated efficacy in several scenarios, such as monotherapy, combination with chemotherapy or targeted therapy, and first-line and subsequent treatment lines in almost all cancers [2,3,4]. The monoclonal antibodies against PD-1 and PD-L1 are the foundations of modern immunotherapy, and currently, eight anti-PD-1/PD-L1 agents are licensed in several different indications [5,6]. Still, many patients do not respond to ICIs, and resistance is inevitable in previously responding patients. Baseline and on-treatment immune dysregulation and immune exhaustion are the main proposed reasons for this treatment resistance [7]. There is an urgent need for novel approaches, such as combinations, to modify tumor microenvironments and mobilize the immune system to fight against tumors more effectively.

Phytochemicals are plant-based bioactive chemicals with diverse health-promoting effects, including cancer prevention. Removal of antioxidant stress, prevention of DNA damage, and promotion of apoptosis after DNA damage with phytochemicals are hypothesized to be the main mechanisms of the anticancer effects of phytochemicals [8,9,10]. However, besides the antioxidant effects, the impacts of phytochemicals on the immune system could be instrumental for anticancer effects, and the immunomodulatory effects of phytochemicals in tumor microenvironments have recently received a lot of interest [11,12]. The preclinical evidence has demonstrated that plant-based polysaccharides, phenols, and silibinin could improve immunosuppression in the tumor microenvironments by inhibiting myeloid-derived suppressor cell accumulation and CD4-T lymphocyte expansion, as well as change the PD-L1 expression levels in tumor microenvironments [13]. Additionally, metabolomic studies have pointed out the possible role of phytochemicals as the mediators of the immune system and gut microbiome cross-talk, as evidenced by the increased response rates in ICI-treated melanoma patients with increased anacardic acid levels. Beneficial commensal bacteria such as Faecalibacterium prausnitzii and Bacteroides thetaiotamicron were also enriched in ICI responders, supporting an association between phytochemicals and a healthier gut microbiome and driving a better immune response [14].

Based on these immunomodulatory properties, phytochemicals could be candidates for combination therapies with anti-PD-1/anti-PD-L1 agents. Both preclinical data and data from metabolome and microbiome analyses have pointed out the role of phytochemicals in mediating immunotherapy efficacy. However, the evidence is limited, with conflicting results. Therefore, we aim to review the available evidence on the phytochemical and ICI combinations and to discuss the opportunities to improve ICI efficacy and possible challenges resulting from these combinations.

2. Literature Search and Review Structure

We conducted a literature review from the PubMed, Medline, and Embase databases to perform filtering of published studies. The MeSH search terms were “immunotherapy” OR “immune checkpoint” AND “PD-1” OR “PD-L1” OR “CTLA-4” OR “phytochemical” OR “phytonutrient” OR “microbiome”. Additionally, we conducted PubMed searches for the individual phytochemical names. We included original articles in the English language that evaluate the effects of phytochemicals on immune checkpoints, ICI and phytochemical combinations, and the association between the microbiome and ICI efficacy.

The review consisted of four sections. The first section was about “Preclinical Evidence Evaluating the Effects of Phytochemicals on Immune Checkpoints”, for which we reviewed the available evidence on the effects of phytochemicals on immune checkpoints in cell line and animal models. In the second section of the review, “The Preclinical Studies Evaluating the Efficacy of Phytochemical and Immune Checkpoint Inhibitor Combinations”, we reviewed the studies evaluating phytochemical and ICI combinations. Due to a lack of human data, we included preclinical studies only. In “The Clues from the Microbiome Studies on the Benefit of Phytochemicals in Immunotherapy Efficacy”, we reviewed the association between phytochemicals and the microbiome and the therapeutic opportunities that stemmed from phytochemicals’ effects on the microbiome. In the last part of the review, we discussed the further clinical perspectives and challenges related to phytochemical and ICI combinations in the clinic and areas needing improvement to develop effective combinations.

3. Preclinical Evidence Evaluating the Effects of Phytochemicals on Immune Checkpoints

Although the effects of phytochemicals on the immune system and tumor microenvironment have been extensively studied [12,15], the impact of phytochemicals on the immune checkpoints began receiving increased interest following ICIs’ entrance into clinical practice. Phytochemicals have been suggested to have the potential to modulate the therapeutic effects of immune checkpoint inhibitors through the regulation of several signaling pathways (Figure 1). Preclinical studies with variable phytochemicals were conducted on several tumor types, including NSCLC, breast cancer, colorectal cancer, hepatocellular cancer, melanoma, head and neck cancer, and glial tumors (Table 1). While most studies have focused on phytochemicals’ effects on the PD-1 and PD-L1 pathways, one study reported the blockage of both PD-1/PD-L1 and the CTLA-4/CD80 interactions by the flavonoids eriodictyol fisetin and quercetin liquiritigenin [16]. The phytochemicals mainly caused decreases in the tumor PD-L1 expression. In contrast, Z-guggulsterone, a phytosterol, increased PD-L1 mRNA expression and transcription in a dose-dependent manner via the activation of the AKT and ERK1/2 signaling pathways [17]. While the authors suggested that increased PD-L1 expression could be an opportunity to develop combinations with synergism, the increased PD-1 expression secondary to AKT and ERK1/2 activations could reduce ICI efficacy as a result of the pivotal roles of AKT and ERK1/2 activation in resistance to immunotherapy in clinical studies [18,19].

Interestingly, Phoenix dactylifera extract (including gallic acid, coffeic acid, and ellagic acid phytochemicals) increased cardiac and kidney PD-1 expressions in a mouse model exposed to Adriamycin [20]. The Phoenix dactylifera extract led to attenuated cardiotoxicity and nephrotoxicity secondary to decreased oxidative stress and apoptotic pressure. The authors suggested that increased PD-1 levels could protect from cardiotoxicity and nephrotoxicity via reduced oxidative stress and apoptotic pressure [20]. If their observations could be confirmed in clinical trials, the phytochemicals could be used for toxicity prevention in ICI-treated patients. Additionally, increased PD-1 expression and corresponding immunosuppression could be exercised as a strategy to prevent or mitigate immune-related adverse events by immunotherapy.

The STAT pathway was the other commonly affected pathway from phytochemicals, with five studies reporting STAT pathway aberrations in addition to changes in the PD-1/PD-L1 pathway [21,22,23,24,25]. While STAT1 and STAT2 involve the constitution of the anti-tumor response via cross-talks with IFN, STAT3 has pro-oncogenic and immunosuppressive properties [26]. Phytochemicals created a consistent inhibitory effect on the STAT3 pathway, and STAT3 down-regulation was among the drivers of PD-L1 suppression in several studies [22,25]. Contrary to the inhibition of immune-suppressive STAT3 by phytochemicals, Xu et al. demonstrated an inhibitory effect on STAT1 by apigenin, a flavonoid, and observed a greater decrease in PD-L1 by apigenin than curcumin secondary to more potent inhibition of STAT1 by apigenin [23]. These observations show that different phytochemicals could act differently on STAT proteins (immune-activating vs. immune-suppressive), and various classes of phytochemicals (i.e., flavonoid and non-flavonoid polyphenols) could have different potency on the tumor immune milieu (Figure 2). Further delineation of these interactions for individual phytochemicals is paramount to designing successful phytochemical and immunotherapy combinations.

Table 1. The combined effect of dietary phytochemicals and immune checkpoint inhibitors on different types of cancer.

Authors	Phytochemical Group	Phytochemical Compound	Immune Checkpoint Inhibitor	Cancer Type	Cancer Model	Mechanism of Effect	Outcome
Shao et al. (2017) [27]	Non-flavonoid polyphenols	Curcumin	Anti-PD-L1 antibody (200 µg)	Bladder Cancer	MB49 bladder carcinoma tumor-bearing C57BL/6 mice	- Increase intratumoral CD8+ T-cell infiltration - Elevated the level of IFN-γ in the blood - Decrease the number of intratumoral MDSCs	Prolonged survival of intraperitoneal metastasized bladder cancer
Dent et al. (2020) [28]	Non-flavonoid polyphenols	Curcumin	Anti-PD-1 antibody (50 mg/kg)	CRC	CT26 colorectal tumor-bearing C57/BL6 mice	- Reduce the expression of PD-L1, PD-L2, and ODC - Elevated the level of MHCA	Reduce tumor growth
Guo et al. (2021) [29]	Non-flavonoid polyphenols	Curcumin	Anti-PD-1 antibody (10 mg/kg)	HCC	Hep3B hepatocellular tumor-bearing BALB/c female nude mice	- Reduce surface PD-L1 expression - Activate lymphocytes - Inhibit immune evasion - Down-regulate TGF-β1 expression	Lowered the HCC growth rate and improved tumor microenvironment
Gong et al. (2021) [30]	Non-flavonoid polyphenols	Curcumin	Anti-PD-1 antibody (10 mg/kg)	CRC	MC-38 colorectal tumor-bearing C57BL/6 mice	N/A	- Reduce tumor growth and tumor volume -Reduce the risk of tumor recurrence (0% vs. 50% in the combination and IO monotherapy arms)
Hayakawa et al. (2021) [31]	Non-flavonoid polyphenols	Curcumin	Anti-PD-1 Ab and Anti-PD-L1 Ab (200 µg/100 µL/mouse)	CRC	MC-38 colorectal tumor-bearing C57BL/6 mice CT26 colorectal tumor-bearing BALB/c mice	Inhibit STAT3 expression induced by exogenous IL-6	Reduce tumor growth
Liu et al. (2021) [32]	Non-flavonoid polyphenols	Curcumin	Anti-PD-L1 antibody (10 µg/mL)	HNSCC	Human HNSCC cell lines (SNU1076, SNU1041, SCC15, and FaDu)	Reinvigorates defective T-cells through multiple (PD-1 and TIM-3) and multi-level (IC receptors and their ligands) IC axis inhibition	Reduce the tumor volume and weight
Kang et al. (2020) [33]	Non-flavonoid polyphenols	Gallic acid	Anti-PD-1 antibody (5 μg/mL)	NSCLC	A549 and H292 NSCLC cell lines	- Reduce expression levels of PD-L1 - Increase IFNγ levels	Decrease NSCLC cell viability
Lasso et al. (2020) [34]	Non-flavonoid polyphenols	Gallotallin	Anti-PD-L1 antibody (200 µg)	Melanoma	B16-F10 melanoma tumor-bearing C57BL/6 mice	- Increase IFNγ levels - Increase the number of activated CD4+ and CD8+ T-cells - Decrease the number of MDSCs - Increase PD-L1 expression	Decrease in tumor size
Zhang et al. (2019) ([35]	Non-flavonoid polyphenols	Resveratrol	Anti-PD-L1 antibody (100 µg)	Ovarian Cancer	Human ovarian carcinoma cell lines (SKOV3 and A2780) and murine ovarian carcinoma cell line (ID8)	Induction of tumor cell apoptosis	Reduce tumor growth
Jiang et al. (2021) [25]	Flavonoid	Luteolin Apigenin	Anti-PD-1 antibody (10 mg/kg)	NSCLC	KRAS-mutant human lung cell lines (H358, H460, H2122, and A549) and mice in vivo mode	Down-regulated the IFN-γ-induced PD-L1 expression by suppressing the phosphorylation of STAT3	Reduce the tumor volume and weight
Liu et al. (2020) [36]	Flavonoid	Bilberry Anthocyanin	Anti-PD-L1 antibody (200 µg)	CRC	MC-38 colorectal tumor-bearing C57BL/6 mice	Modulate the gut microbiome	Tumor growth delay
Mo et al. (2021) [37]	Flavonoid	Icaritin	Anti-PD-1 antibody (10 mg/kg)	HCC, CRC, and melanoma	Human hepatocellular carcinoma, colorectal cancer and melanoma cell lines (HEPA1–6, MC-38, and B16F10) and mice in vivo model	Down-regulate PD-L1 expression and reduced nuclear translocation of NF-κB p6.	Reduce the tumor volume and weight
Tang et al. (2019) [38]	Flavonoid	Melafolone	Anti-PD-1 Ab (200 µg/100 µL/mouse)	Lung Cancer	Lewis lung carcinoma or CMT tumor-bearing C57BL/6 mice	Down-regulate VEGF, TGF-β, and PD-L1 through COX-2 and EGFR inhibition	Promoted survival Tumor growth inhibition
Jiang et al. (2019) [39]	Terpenes	Lycopene	Anti-PD-1 antibody (6 mg/kg)	Lung Cancer	Lewis lung carcinoma tumor-bearing C57BL/6 mice	- Increase IFNγ levels - Inhibit the expression of PD-L1 via activating JAK - Inhibit the phosphorylation of AKT	Reduce the tumor volume and weight
Han et al. (2019) [40]	Terpenes	Cryptotanshinone	Anti-PD-L1 antibody 10 µg)	HCC	HCC-bearing mice model	Develops long-term ani-tumor immunity and increased tumor infiltration of CD8+ T-cell	Tumor growth inhibition
Dong et al. (2018) [41]	Terpenes	Diosgenin	Anti-PD-1 antibody (200 µg)	Melanoma	B16-F10 melanoma tumor-bearing C57BL/6 mice	Enhances T-cell immune response by modulating intestinal microbiota and inducing T-cell infiltration	Tumor growth inhibition
Ye et al. (2021) [42]	Others	Agrocybe aegerita galectin	Anti-PD-1 Ab (200 µg intraperitoneal)	HCC	H22, HepG2, and RAW264.7 cell lines Male Balb/c mice	Increase CD4+ and CD8+ T-cells with combination	Tumor growth inhibition

Abbreviations: COX-2: Cyclooxygenase-2; EGFR: Epidermal growth factor receptor; HCC: Hepatocellular carcinoma; IFNγ: Interferon-gamma; IL-6: Interleukin 6; MDSCs: Myeloid-derived suppressor cells; NSCLC: Non-small-cell lung carcinoma; MHCA: Human major histocompatibility complex class I A; PD-L: Programmed death ligand; ODC: Ornithine decarboxylase; STAT3: Signal transducer and activator of transcription 3; TNF: Tumor necrosis factor; VEGF: Vascular endothelial growth factor.

Table 1. The combined effect of dietary phytochemicals and immune checkpoint inhibitors on different types of cancer.

Authors	Phytochemical Group	Phytochemical Compound	Immune Checkpoint Inhibitor	Cancer Type	Cancer Model	Mechanism of Effect	Outcome
Shao et al. (2017) [27]	Non-flavonoid polyphenols	Curcumin	Anti-PD-L1 antibody (200 µg)	Bladder Cancer	MB49 bladder carcinoma tumor-bearing C57BL/6 mice	- Increase intratumoral CD8+ T-cell infiltration - Elevated the level of IFN-γ in the blood - Decrease the number of intratumoral MDSCs	Prolonged survival of intraperitoneal metastasized bladder cancer
Dent et al. (2020) [28]	Non-flavonoid polyphenols	Curcumin	Anti-PD-1 antibody (50 mg/kg)	CRC	CT26 colorectal tumor-bearing C57/BL6 mice	- Reduce the expression of PD-L1, PD-L2, and ODC - Elevated the level of MHCA	Reduce tumor growth
Guo et al. (2021) [29]	Non-flavonoid polyphenols	Curcumin	Anti-PD-1 antibody (10 mg/kg)	HCC	Hep3B hepatocellular tumor-bearing BALB/c female nude mice	- Reduce surface PD-L1 expression - Activate lymphocytes - Inhibit immune evasion - Down-regulate TGF-β1 expression	Lowered the HCC growth rate and improved tumor microenvironment
Gong et al. (2021) [30]	Non-flavonoid polyphenols	Curcumin	Anti-PD-1 antibody (10 mg/kg)	CRC	MC-38 colorectal tumor-bearing C57BL/6 mice	N/A	- Reduce tumor growth and tumor volume -Reduce the risk of tumor recurrence (0% vs. 50% in the combination and IO monotherapy arms)
Hayakawa et al. (2021) [31]	Non-flavonoid polyphenols	Curcumin	Anti-PD-1 Ab and Anti-PD-L1 Ab (200 µg/100 µL/mouse)	CRC	MC-38 colorectal tumor-bearing C57BL/6 mice CT26 colorectal tumor-bearing BALB/c mice	Inhibit STAT3 expression induced by exogenous IL-6	Reduce tumor growth
Liu et al. (2021) [32]	Non-flavonoid polyphenols	Curcumin	Anti-PD-L1 antibody (10 µg/mL)	HNSCC	Human HNSCC cell lines (SNU1076, SNU1041, SCC15, and FaDu)	Reinvigorates defective T-cells through multiple (PD-1 and TIM-3) and multi-level (IC receptors and their ligands) IC axis inhibition	Reduce the tumor volume and weight
Kang et al. (2020) [33]	Non-flavonoid polyphenols	Gallic acid	Anti-PD-1 antibody (5 μg/mL)	NSCLC	A549 and H292 NSCLC cell lines	- Reduce expression levels of PD-L1 - Increase IFNγ levels	Decrease NSCLC cell viability
Lasso et al. (2020) [34]	Non-flavonoid polyphenols	Gallotallin	Anti-PD-L1 antibody (200 µg)	Melanoma	B16-F10 melanoma tumor-bearing C57BL/6 mice	- Increase IFNγ levels - Increase the number of activated CD4+ and CD8+ T-cells - Decrease the number of MDSCs - Increase PD-L1 expression	Decrease in tumor size
Zhang et al. (2019) ([35]	Non-flavonoid polyphenols	Resveratrol	Anti-PD-L1 antibody (100 µg)	Ovarian Cancer	Human ovarian carcinoma cell lines (SKOV3 and A2780) and murine ovarian carcinoma cell line (ID8)	Induction of tumor cell apoptosis	Reduce tumor growth
Jiang et al. (2021) [25]	Flavonoid	Luteolin Apigenin	Anti-PD-1 antibody (10 mg/kg)	NSCLC	KRAS-mutant human lung cell lines (H358, H460, H2122, and A549) and mice in vivo mode	Down-regulated the IFN-γ-induced PD-L1 expression by suppressing the phosphorylation of STAT3	Reduce the tumor volume and weight
Liu et al. (2020) [36]	Flavonoid	Bilberry Anthocyanin	Anti-PD-L1 antibody (200 µg)	CRC	MC-38 colorectal tumor-bearing C57BL/6 mice	Modulate the gut microbiome	Tumor growth delay
Mo et al. (2021) [37]	Flavonoid	Icaritin	Anti-PD-1 antibody (10 mg/kg)	HCC, CRC, and melanoma	Human hepatocellular carcinoma, colorectal cancer and melanoma cell lines (HEPA1–6, MC-38, and B16F10) and mice in vivo model	Down-regulate PD-L1 expression and reduced nuclear translocation of NF-κB p6.	Reduce the tumor volume and weight
Tang et al. (2019) [38]	Flavonoid	Melafolone	Anti-PD-1 Ab (200 µg/100 µL/mouse)	Lung Cancer	Lewis lung carcinoma or CMT tumor-bearing C57BL/6 mice	Down-regulate VEGF, TGF-β, and PD-L1 through COX-2 and EGFR inhibition	Promoted survival Tumor growth inhibition
Jiang et al. (2019) [39]	Terpenes	Lycopene	Anti-PD-1 antibody (6 mg/kg)	Lung Cancer	Lewis lung carcinoma tumor-bearing C57BL/6 mice	- Increase IFNγ levels - Inhibit the expression of PD-L1 via activating JAK - Inhibit the phosphorylation of AKT	Reduce the tumor volume and weight
Han et al. (2019) [40]	Terpenes	Cryptotanshinone	Anti-PD-L1 antibody 10 µg)	HCC	HCC-bearing mice model	Develops long-term ani-tumor immunity and increased tumor infiltration of CD8+ T-cell	Tumor growth inhibition
Dong et al. (2018) [41]	Terpenes	Diosgenin	Anti-PD-1 antibody (200 µg)	Melanoma	B16-F10 melanoma tumor-bearing C57BL/6 mice	Enhances T-cell immune response by modulating intestinal microbiota and inducing T-cell infiltration	Tumor growth inhibition
Ye et al. (2021) [42]	Others	Agrocybe aegerita galectin	Anti-PD-1 Ab (200 µg intraperitoneal)	HCC	H22, HepG2, and RAW264.7 cell lines Male Balb/c mice	Increase CD4+ and CD8+ T-cells with combination	Tumor growth inhibition

Abbreviations: COX-2: Cyclooxygenase-2; EGFR: Epidermal growth factor receptor; HCC: Hepatocellular carcinoma; IFNγ: Interferon-gamma; IL-6: Interleukin 6; MDSCs: Myeloid-derived suppressor cells; NSCLC: Non-small-cell lung carcinoma; MHCA: Human major histocompatibility complex class I A; PD-L: Programmed death ligand; ODC: Ornithine decarboxylase; STAT3: Signal transducer and activator of transcription 3; TNF: Tumor necrosis factor; VEGF: Vascular endothelial growth factor.

4. Preclinical Studies Evaluating the Efficacy of Phytochemical and Immune Checkpoint Inhibitor Combinations

Several studies have evaluated the efficacy of ICI plus phytochemical combinations using variable phytochemicals in different cancer models (Table 2). The first report of a phytochemical and ICI combination strategy (bisdemethoxycurcumin with an anti-PD-L1 antibody) was published by Shao et al. in 2017 in a mouse bladder cancer model [27]. The researchers observed T-lymphocyte-based immune activation and removal of immune exhaustion via a decrease in the intratumoral myeloid-suppressor cells, leading to increased survival [27]. Later studies most frequently used curcumin, a non-flavonoid polyphenol, as the experimental phytochemical [28,29,30,31] and primarily focused on the colorectal cancer models [28,30,31,36], followed by NSCLC [33,38,39]. The intratumoral expansion of CD8+ or CD4+ lymphocytes and the increased levels of IFN-γ were consistent findings throughout the studies [34,42].

The decrease in PD-L1 levels was observed in five studies conducted with curcumin [28,29], lycopene [39], gallic acid [33], and melafolone [38] and could contribute to the synergism between ICI and phytochemicals by releasing the inhibitory breaks of immune checkpoints. In contrast, an increase in PD-L1 expression level was reported by Lasso et al. with gallatonin-rich Caesalpinia spinosa and anti-PD-L1 antibody combination [34]. The previous observations of PD-L1 expression increase with resveratrol and piceatannol in breast and colorectal cancer cell lines support the variable effects of different phytochemicals on immune checkpoint expressions [43], although the reason for contrasting effects on PD-L1 expression with the phytochemicals from the same class (non-flavonoid polyphenols) is yet to be defined.

While it could be problematic to expect a synergism between ICIs and phytochemicals, we think the differential effects of phytochemicals on PD-L1 expression levels could be beneficial to aid individualized treatment planning according to tumor microenvironments in different tumors. While decreasing immunosuppression secondary to decreased PD-L1 expression would be beneficial in a tumor-agnostic manner, increased ICI efficacy with increased tumoral PD-L1 expression levels was observed in NSCLC [33], melanoma [34], and bladder cancer [27] patients. Both the pretreatment with phytochemicals and the addition of phytochemicals at times of progression to increase PD-L1 expression should be evaluated as a treatment strategy, especially in tumors with an increased benefit with increased PD-L1 expressions.

Table 2. Summary of phytochemicals targeting PD-L1/PD-1 and CTLA-4 pathways in different types of cancer.

Authors/Year	Phytochemical Group	Phytochemical Compound	Source	Cancer Type	Cancer Model or Toxicity	Mechanism of Action
Liao et al. (2018) [21]	Non-flavonoid polyphenols	Curcumin	Turmeric	HNSCC	4-NQO induced C57BL/6 tongue squamous cell carcinoma mice	Decrease PD-L1 and p-STAT3Y705 protein expression
Deng et al. (2020) [44]	Non-flavonoid polyphenols	Curcumin	Turmeric	HCC	HepG2 hepatocellular tumor-bearing BALB/c female nude mice	Decreased the protein expression of PD-L1 Inhibiting the TLR4/NF-κB signaling pathway angiogenesis.
Lucas et al. (2018) [43]	Non-flavonoid polyphenols	Resveratrol	Red wine, Grapes, Passion fruit	Breast cancer	Cal51 triple-negative breast cancer and SW620 colon cancer	Increase the expression level of PD-L1 via HDAC3/p300-mediated nuclear factor (NF)-κB signaling
Verdura et al. (2020) [45]	Non-flavonoid polyphenols	Resveratrol	Red wine, Grapes, Passion fruit	Breast Cancer	JIMT-1 and MDA-MB-231 breast cancer cells	Increased PD-L1 dysfunction
Yang et al. (2021) [46]	Non-flavonoid polyphenols	Resveratrol	Red wine, Grapes, Passion fruit	NSCLC	Human lung adenocarcinoma cell lines (A549 and H1299)	Activation of SirT1 deacetylase leads to disassembly of the destruction complex, thereby enhancing the binding of β-catenin/TCF to the PD-L1 promoter
Coomb et al. (2016) [47]	Flavonoid	Apigenin	Parsley, onions, grapefruit, oranges,	Breast Cancer	Triple-negative MDA-MB-468 BC cells, HER2(+) SK-BR-3 BC cells, and 4T1 mouse mammary carcinoma cells	Inhibit IFNγ-induced PD-L1 upregulation
Xu et al. (2018) [23]	Flavonoid	Apigenin	Apple, artichoke, basil, celery, cherry, grapes	Melanoma	B16-F10 melanoma tumor-bearing C57BL/6 mice	Inhibit the IFN-γ-induced activation of STAT Decreased expression levels of PD-L1
Choi et al. (2020) [48]	Flavonoid	Apigenin	Salvia plebeia	CRC	hPD-L1 knock-in MC38 tumor-bearing humanized PD-1 mouse model	Blocking of PD-1/PD-L1 interaction
Rawangkan et al. (2018) [49]	Flavonoid	Epigallocatechin gallate (EGCG)	Green tea	NSCLC	4-(methylnitrosamino)-1-(3-pyridyl)-1- butanone induced nonsmall-cell lung cancer A/J mice	Down-regulate IFN-γ- and EGF-induced PD-L1 expression
Sellam et al. (2020) [50]	Flavonoid	Silibinin	Silybum marianum	HNSCC	Nasopharyngeal carcinoma cell line	Down-regulation in PD-L1 expression by interfering with HIF-1α/LDH-A
Rugamba et al. (2021) [51]	Flavonoid	Silibinin	Silybum marianum	NSCLC	A549, H292, and H460 cell lines	Suppresses the mRNA expression of PD-L1 and EMT regulators via inhibition of STAT3 phosphorylation
Wudtiwai et al. (2021) [52]	Flavonoid	Hesperidin	Orange peel and other citrus species	Oral Cancer	Human OSCC cell lines (HN6 and HN15)	Inhibition of phosphorylation of STAT1 and STAT3 down-regulates IFN-γ -induced PD-L1 expression
Ke et al. (2019) [24]	Flavonoid	Baicalin	Scutellaria baicalensis	HCC	H22 hepatocellular tumor-bearing BALB/c mice or BALB/c-nu/nu mice	Decrease STAT3 activity Down-regulate IFN-γ-induced PD-L1 expression
Song et al. (2022) [53]	Flavonoid	Baicalin	Scutellaria baicalensis	CRC	Human colon cancer cell lines (HCT-116 and CT26) and mice in vivo model	Inhibition of NF-κB signaling pathway down-regulated PD-L1 expression and MDSC, and up-regulated CD4 + and CD8 + T-cells
Liu et al. (2021) [54]	Flavonoid	Licochalcone A	Glycyrrhiza glabra	CRC	Human cancer cell lines (A549, HeLa, Hep3B, and HCT116) and mice in vivo model	PD-L1 expression was down-regulated by inhibition of NF-κB and Ras/Raf/MEK signaling pathways.
Hao et al. (2019) [55]	Flavonoid	Icaritin	Epimedium	Melanoma	B16-F10 melanoma tumor-bearing C57BL/6 mice MC-38 colorectal tumor-bearing C57BL/6 mice	Reduce frequency of MDSCs Down-regulate PD-L1 expression
Mazewski et al. (2019) [56]	Flavonoid	Cyanidin-3-O-glucoside	Blueberry, raspberry, black rice, cherry	CRC	Human colorectal cancer cells	Decreased PD-1 and PD-L1 protein expression
Chen et al. (2022) [57]	Flavonoid	Myricetin	Cranberry, blueberry, lemon, garlic	Lung cancer	Human lung cancer cell lines (NCI-H1650, NCI-H460, and A549)	Inhibition of IFN-γ-induced PD-L1 and IDO1 expression in cancer cells by inhibiting the JAK-STAT-IRF1 axis
Kim et al. (2020) [58]	Flavonoid	Kaempferol	Geranii herba	N/A	PD-1 Jurkat and PD-L1/aAPC CHO-K1 cells	Inhibiting PD-1/PD-L1 Interaction
Sahyon et al.(2020) [20]	Flavonoid	Gallic acid	Phoenix dactylifera	N/A	Adriamycin-induced cardiotoxicity and nephrotoxicity	Increased cardiac and kidney PD-1 protein percentage
Xing et al. (2018) [22]	Terpenes	Fraxinellone	Dictamnus dasycarpus	Lung cancer	Human A549 lung cancer cell line	Inhibit PD-L1 expression by downregulating the STAT3 and HIF-1α signaling pathways
Huang et al. (2019) [59]	Terpenes	Platycodin	Platycodon grandiflorus	Lung cancer	NCI–H1975 and NCI–H358 lung cancer cell lines	Down-regulate the protein level of PD-L1
Zhang et al. (2019) [60]	Terpenes	Triptolide	Tripterygium wilfordii	Glioma	Glioma cells	Down-regulated IFN-γ-induced PD-L1 expression
Kuo et al. (2021) [61]	Terpenes	Triptolide	Tripterygium wilfordii	HNSCC	OSCC cell line SAS (JCRB0260) and mice in vivo model	Reduces IFN-γ-related JAK2-STAT1 pathway and decreases PD-L1 expression
Tian et al. (2021) [17]	Phytosterol	Z-guggulsterone	Commiphora mukul tree	NSCLC	Lewis lung carcinoma tumor-bearing C57BL/6 mice	Inducing PD-L1 upregulation partly mediated by FXR, Akt, and Erk1/2 signaling pathways
Wang et al. (2020) [62]	Saponins	Panaxadiol	Panax ginseng	CRC	HCT116, SW620, HT29, and HEK293 colon cancer cell line and mice in vivo model	Reduces PD-L1 expression by suppressing HIF-1α and STAT3
Deng et al. (2020) [44]	Saponins	Ginsenosides	Panax	HCC	HepG2 hepatocellular tumor-bearing BALB/c female nude mice	Decreased the protein expression of PD-L1 Inhibiting the TLR4/NF-κB signaling pathway angiogenesis.
Bedi et al. (2019) [63]	Alkaloid	Camptothecin	Camptotheca acuminita	CRC	SW620, HCT116, and RKO colon cancer cells	Reduces PD-L1 expression and upregulates the secretion of pro-tumorigenic cytokines
Hunakova et al. (2019) [64]	Isothiocyanates	Isothiocyanate	Broccoli, Brussels sprouts, cabbage, cauliflower, horseradish	Breast Cancer	Human triple-negative Breast Carcinoma MDA-MB-231 Cells	Decrease expression levels of PD-L1
Chang et al. (2019) [65]	Astragalus membranaceus extract	Extract	Astragalus membranaceus	Breast cancer, CRC	Mouse breast cancer 4T1 and colorectal cancer CT26	Down-regulate PD-L1 expression by suppressing the AKT signaling pathway
Li et al. (2019) [16]	Rhus verniciflua Stokes extract	Eriodictyol, fisetin, quercetin, liquiritigenin	Rhus verniciflua Stokes	-	N/A	Blocked both the PD-1/PD-L1 and the CTLA-4/CD80 interactions
Safonova et al. (2020) [66]	Tussilago farfara extract	Rhamnogalacturonane I and neutral polysaccharides complex	Tussilago farfara	Lung cancer	Lewis lung carcinoma tumor-bearing C57BL/6 mice	Reducing expression levels of PD-1 and PD-L1 Inhibiting PD-1/PD-L1 Interaction
Ryan et al. (2022) [67]	Black raspberry extract	Extract	Black raspberry	HNSCC	Nitroquinoline-1-oxide (4NQO) induced head and neck cancer C57BL/6 mice	Decreased levels of PD-L1 expression

Abbreviations: 4-NQO: 4-nitroquinoline-1-oxide; CTLA4: Cytotoxic T-Lymphocyte Associated Protein 4; EGF: Epidermal growth factor; ERK1/2: Extracellular signal-regulated kinase 1/2; FXR: farnesoid X receptor; HIF-1:hypoxia-inducible factor-1; HNSCC: Head and neck squamous cell carcinoma; IFNγ: Interferon-gamma; LDH: Lactate dehydrogenase; MDSCs: Myeloid-derived suppressor cells; NSCLC: Non-small-cell lung carcinoma; NF-κB: Nuclear factor kappa-light-chain-enhancer of activated B cells; PD-L: Programmed death ligand; STAT3: Signal transducer and activator of transcription.

Table 2. Summary of phytochemicals targeting PD-L1/PD-1 and CTLA-4 pathways in different types of cancer.

Authors/Year	Phytochemical Group	Phytochemical Compound	Source	Cancer Type	Cancer Model or Toxicity	Mechanism of Action
Liao et al. (2018) [21]	Non-flavonoid polyphenols	Curcumin	Turmeric	HNSCC	4-NQO induced C57BL/6 tongue squamous cell carcinoma mice	Decrease PD-L1 and p-STAT3Y705 protein expression
Deng et al. (2020) [44]	Non-flavonoid polyphenols	Curcumin	Turmeric	HCC	HepG2 hepatocellular tumor-bearing BALB/c female nude mice	Decreased the protein expression of PD-L1 Inhibiting the TLR4/NF-κB signaling pathway angiogenesis.
Lucas et al. (2018) [43]	Non-flavonoid polyphenols	Resveratrol	Red wine, Grapes, Passion fruit	Breast cancer	Cal51 triple-negative breast cancer and SW620 colon cancer	Increase the expression level of PD-L1 via HDAC3/p300-mediated nuclear factor (NF)-κB signaling
Verdura et al. (2020) [45]	Non-flavonoid polyphenols	Resveratrol	Red wine, Grapes, Passion fruit	Breast Cancer	JIMT-1 and MDA-MB-231 breast cancer cells	Increased PD-L1 dysfunction
Yang et al. (2021) [46]	Non-flavonoid polyphenols	Resveratrol	Red wine, Grapes, Passion fruit	NSCLC	Human lung adenocarcinoma cell lines (A549 and H1299)	Activation of SirT1 deacetylase leads to disassembly of the destruction complex, thereby enhancing the binding of β-catenin/TCF to the PD-L1 promoter
Coomb et al. (2016) [47]	Flavonoid	Apigenin	Parsley, onions, grapefruit, oranges,	Breast Cancer	Triple-negative MDA-MB-468 BC cells, HER2(+) SK-BR-3 BC cells, and 4T1 mouse mammary carcinoma cells	Inhibit IFNγ-induced PD-L1 upregulation
Xu et al. (2018) [23]	Flavonoid	Apigenin	Apple, artichoke, basil, celery, cherry, grapes	Melanoma	B16-F10 melanoma tumor-bearing C57BL/6 mice	Inhibit the IFN-γ-induced activation of STAT Decreased expression levels of PD-L1
Choi et al. (2020) [48]	Flavonoid	Apigenin	Salvia plebeia	CRC	hPD-L1 knock-in MC38 tumor-bearing humanized PD-1 mouse model	Blocking of PD-1/PD-L1 interaction
Rawangkan et al. (2018) [49]	Flavonoid	Epigallocatechin gallate (EGCG)	Green tea	NSCLC	4-(methylnitrosamino)-1-(3-pyridyl)-1- butanone induced nonsmall-cell lung cancer A/J mice	Down-regulate IFN-γ- and EGF-induced PD-L1 expression
Sellam et al. (2020) [50]	Flavonoid	Silibinin	Silybum marianum	HNSCC	Nasopharyngeal carcinoma cell line	Down-regulation in PD-L1 expression by interfering with HIF-1α/LDH-A
Rugamba et al. (2021) [51]	Flavonoid	Silibinin	Silybum marianum	NSCLC	A549, H292, and H460 cell lines	Suppresses the mRNA expression of PD-L1 and EMT regulators via inhibition of STAT3 phosphorylation
Wudtiwai et al. (2021) [52]	Flavonoid	Hesperidin	Orange peel and other citrus species	Oral Cancer	Human OSCC cell lines (HN6 and HN15)	Inhibition of phosphorylation of STAT1 and STAT3 down-regulates IFN-γ -induced PD-L1 expression
Ke et al. (2019) [24]	Flavonoid	Baicalin	Scutellaria baicalensis	HCC	H22 hepatocellular tumor-bearing BALB/c mice or BALB/c-nu/nu mice	Decrease STAT3 activity Down-regulate IFN-γ-induced PD-L1 expression
Song et al. (2022) [53]	Flavonoid	Baicalin	Scutellaria baicalensis	CRC	Human colon cancer cell lines (HCT-116 and CT26) and mice in vivo model	Inhibition of NF-κB signaling pathway down-regulated PD-L1 expression and MDSC, and up-regulated CD4 + and CD8 + T-cells
Liu et al. (2021) [54]	Flavonoid	Licochalcone A	Glycyrrhiza glabra	CRC	Human cancer cell lines (A549, HeLa, Hep3B, and HCT116) and mice in vivo model	PD-L1 expression was down-regulated by inhibition of NF-κB and Ras/Raf/MEK signaling pathways.
Hao et al. (2019) [55]	Flavonoid	Icaritin	Epimedium	Melanoma	B16-F10 melanoma tumor-bearing C57BL/6 mice MC-38 colorectal tumor-bearing C57BL/6 mice	Reduce frequency of MDSCs Down-regulate PD-L1 expression
Mazewski et al. (2019) [56]	Flavonoid	Cyanidin-3-O-glucoside	Blueberry, raspberry, black rice, cherry	CRC	Human colorectal cancer cells	Decreased PD-1 and PD-L1 protein expression
Chen et al. (2022) [57]	Flavonoid	Myricetin	Cranberry, blueberry, lemon, garlic	Lung cancer	Human lung cancer cell lines (NCI-H1650, NCI-H460, and A549)	Inhibition of IFN-γ-induced PD-L1 and IDO1 expression in cancer cells by inhibiting the JAK-STAT-IRF1 axis
Kim et al. (2020) [58]	Flavonoid	Kaempferol	Geranii herba	N/A	PD-1 Jurkat and PD-L1/aAPC CHO-K1 cells	Inhibiting PD-1/PD-L1 Interaction
Sahyon et al.(2020) [20]	Flavonoid	Gallic acid	Phoenix dactylifera	N/A	Adriamycin-induced cardiotoxicity and nephrotoxicity	Increased cardiac and kidney PD-1 protein percentage
Xing et al. (2018) [22]	Terpenes	Fraxinellone	Dictamnus dasycarpus	Lung cancer	Human A549 lung cancer cell line	Inhibit PD-L1 expression by downregulating the STAT3 and HIF-1α signaling pathways
Huang et al. (2019) [59]	Terpenes	Platycodin	Platycodon grandiflorus	Lung cancer	NCI–H1975 and NCI–H358 lung cancer cell lines	Down-regulate the protein level of PD-L1
Zhang et al. (2019) [60]	Terpenes	Triptolide	Tripterygium wilfordii	Glioma	Glioma cells	Down-regulated IFN-γ-induced PD-L1 expression
Kuo et al. (2021) [61]	Terpenes	Triptolide	Tripterygium wilfordii	HNSCC	OSCC cell line SAS (JCRB0260) and mice in vivo model	Reduces IFN-γ-related JAK2-STAT1 pathway and decreases PD-L1 expression
Tian et al. (2021) [17]	Phytosterol	Z-guggulsterone	Commiphora mukul tree	NSCLC	Lewis lung carcinoma tumor-bearing C57BL/6 mice	Inducing PD-L1 upregulation partly mediated by FXR, Akt, and Erk1/2 signaling pathways
Wang et al. (2020) [62]	Saponins	Panaxadiol	Panax ginseng	CRC	HCT116, SW620, HT29, and HEK293 colon cancer cell line and mice in vivo model	Reduces PD-L1 expression by suppressing HIF-1α and STAT3
Deng et al. (2020) [44]	Saponins	Ginsenosides	Panax	HCC	HepG2 hepatocellular tumor-bearing BALB/c female nude mice	Decreased the protein expression of PD-L1 Inhibiting the TLR4/NF-κB signaling pathway angiogenesis.
Bedi et al. (2019) [63]	Alkaloid	Camptothecin	Camptotheca acuminita	CRC	SW620, HCT116, and RKO colon cancer cells	Reduces PD-L1 expression and upregulates the secretion of pro-tumorigenic cytokines
Hunakova et al. (2019) [64]	Isothiocyanates	Isothiocyanate	Broccoli, Brussels sprouts, cabbage, cauliflower, horseradish	Breast Cancer	Human triple-negative Breast Carcinoma MDA-MB-231 Cells	Decrease expression levels of PD-L1
Chang et al. (2019) [65]	Astragalus membranaceus extract	Extract	Astragalus membranaceus	Breast cancer, CRC	Mouse breast cancer 4T1 and colorectal cancer CT26	Down-regulate PD-L1 expression by suppressing the AKT signaling pathway
Li et al. (2019) [16]	Rhus verniciflua Stokes extract	Eriodictyol, fisetin, quercetin, liquiritigenin	Rhus verniciflua Stokes	-	N/A	Blocked both the PD-1/PD-L1 and the CTLA-4/CD80 interactions
Safonova et al. (2020) [66]	Tussilago farfara extract	Rhamnogalacturonane I and neutral polysaccharides complex	Tussilago farfara	Lung cancer	Lewis lung carcinoma tumor-bearing C57BL/6 mice	Reducing expression levels of PD-1 and PD-L1 Inhibiting PD-1/PD-L1 Interaction
Ryan et al. (2022) [67]	Black raspberry extract	Extract	Black raspberry	HNSCC	Nitroquinoline-1-oxide (4NQO) induced head and neck cancer C57BL/6 mice	Decreased levels of PD-L1 expression

Abbreviations: 4-NQO: 4-nitroquinoline-1-oxide; CTLA4: Cytotoxic T-Lymphocyte Associated Protein 4; EGF: Epidermal growth factor; ERK1/2: Extracellular signal-regulated kinase 1/2; FXR: farnesoid X receptor; HIF-1:hypoxia-inducible factor-1; HNSCC: Head and neck squamous cell carcinoma; IFNγ: Interferon-gamma; LDH: Lactate dehydrogenase; MDSCs: Myeloid-derived suppressor cells; NSCLC: Non-small-cell lung carcinoma; NF-κB: Nuclear factor kappa-light-chain-enhancer of activated B cells; PD-L: Programmed death ligand; STAT3: Signal transducer and activator of transcription.

5. Clues from Microbiome Studies on the Benefit of Phytochemicals in Immunotherapy Efficacy

The commensal bacteria residing in the gastrointestinal system and their genome is called the gastrointestinal (GI) microbiome. The GI microbiome plays significant roles in self-defense and inflammation [68]. Recently, the GI microbiome has emerged as a predictor of ICI efficacy, and several studies have demonstrated better survival in ICI-treated patients enriched with beneficial commensal bacteria, including Akkermansia muciniphila, Bacteroides spp., and Faecalibacterium prausnitzii. Akkermansia muciniphila in particular was associated with a response to ICIs in four studies conducted on patients with three different tumors (RCC, NSCLC, and HCC) [69,70,71,72]. Whether the increases in these bacteria are secondary to a bystander effect or these bacteria play roles in ICI efficacy is highly debated. However, the restoration of ICI efficacy with Akkermansia muciniphila transplantation in ICI non-responders mice [70] and the recent report of improved ICI efficacy with microbiome and ICI combination in metastatic renal cell carcinoma patients [73] point to an anti-tumor role of the microbiome rather than it only being a biomarker.

Several phytochemicals such as curcumin and phenols could increase levels of beneficial commensal bacteria (

Table 3. Summary of phytochemicals as modifiers of the gut microbiome in the immunotherapy studies (adapted from Guven DC et al. [77]).

Lead Author, Year	Target Population/Patient Number (n)	Main Findings	Candidate Phytochemical	Phytochemical Enriched Nutrient	Bacteria
Vetizou et al. (2015) [78]	Melanoma/25	Increased levels of Bacteroides thetaiotaomicron and Bacteroides fragilis in ICI responders Improved response to CTLA-4 blockade with FMT from patients with increased fecal Bacteroides spp. levels	Polyphenol/coumarin [79]	Soluble fiber-rich diet [74]/High coffee-consumption [80]	Bacteroides thetaiotaomicron Bacteroides fragilis
Sivan et al. (2015) [81]	Melanoma/Mice	Increased levels of Bacteroides spp. in ICI responders Similar tumor control as PD-L1 blockade with oral supplementation of Bifidobacterium spp.	Resveratrol [82]	Grapes, wine, and peanuts [83]	Bifidobacterium spp.
Dubin et al. (2016) [84]	Melanoma/34	Lower risk of colitis in Bacteroides spp. enriched patients	Polyphenol/coumarin [79]	Soluble fiber-rich diet [74]/High coffee-consumption [80]	Bacteroides spp.
Chaput et al. (2017) [85]	Melanoma/26	Longer progression-free and overall survival in Faecalibacterium spp. enriched patients	Anthocyanin [86]	Black Raspberries [86]	Faecalibacterium spp.
Gopalakrishnan et al. (2018) [87]	Melanoma/112	Higher alpha diversity and increased Ruminococcaceae levels in the feces of ICI responders Higher buccal and fecal levels of Bacteroidales in ICI non-responders	Polyphenols [88] (Naringenin and quercetin) Anthocyanin [89]	Onion, apple, broccoli [90] Black Raspberries [89]	Ruminococcaceae Bacteroidales
Matson et al. (2018) [91]	Melanoma/42	Increased Bifidobacterium longum in the feces of ICI responders Tumor control, augmented T-cell responses, and improved efficacy of anti-PD-L1 blockade with oral supplementation of responders’ feces to germ-free mice	Resveratrol [92]	Grapes, wine, and peanuts [83]	Bifidobacterium longum
Routy et al. (2018) [69]	NSCLC/140 RCC/67	Increased levels of Akkermansia muciniphila in ICI responders Restoration of the efficacy of PD-1 blockade in antibiotic pretreated mice after FMT from ICI responders	Curcumin [93]/EGCG [94]	Prebiotic nondigestible fiber-rich diet [95]	Akkermansia muciniphilia
Peters et al. (2018) [96]	Melanoma/26	Lower risk of progression in patients with higher community diversity Lower risk of progression in patients enriched with Faecalibacterium prausnitzii	Anthocyanin [86]	Black Raspberries [86]	Faecalibacterium prausnitzii
Fukuoka et al. (2018) [97]	NSCLC/14 Gastric cancer/24	Higher alpha diversity and Ruminococcaceae levels in ICI responders	Ellagitannins [98]	Pomegranates, nuts [99]	Ruminococcaceae
Derosa et al. (2018) [70]	RCC/85	Increased abundance of Akkermansia muciniphila and Bacteroides salyersiae in non-resistant renal cell carcinoma patients	Curcumin [93]/EGCG [94]	Prebiotic nondigestible fiber-rich diet [95]	Akkermansia muciniphila
		Restoration of the efficacy of the ICI with Akkermansia muciniphila and Bacteroides salyersiae transplantation to mice with unfavorable/dysbiotic profile	Not reported		Bacteroides salyersiae
Maia et al. (2018) [100]	RCC/20	Increased abundance of Roseburia and Faecalibacterium spp. in ICI responders	Anthocyanin [86] Polyphenols [101]	Black Raspberries [86] Resistant Starch [101]	Faecalibacterium prausnitzii Roseburia
Botticelli et al. (2020) [102]	NSCLC/11	Increased fecal Akkermansia muciniphila, Bifidobacterium longum, and Faecalibacterium prausnitzii levels in ICI responders	Curcumin [93]/EGCG [94] Resveratrol [82], Anthocyanin [103] Anthocyanin [86]	Prebiotic nondigestible fiber-rich diet [95] Grapes, wine, and peanuts [83], Chinese purple sweet potato cultivar [103] Black Raspberries [86]	Akkermansia muciniphila Bifidobacterium longum Faecalibacterium prausnitzii
Liu et al. (2020) [36]	CRC/mice	The addition of anthocyanins in the α-PD-L1 treatment showed an overrepresentation of Lachnospiraceae and Ruminococcaceae	Anthocyanin [86]	Black Raspberries [86]	Ruminococcaceae and Lachnospiraceae
Chung et al. (2021) [71]	HCC/8	Increased fecal Akkermansia levels in ICI responders	Curcumin [93]/EGCG [94]	Prebiotic nondigestible fiber-rich diet [95] Grapes, wine, and peanuts [83]	Akkermansiaceae
Grenda et al. (2022) [72]	NSCLC/47	Increased fecal Akkermansia levels in ICI responders	Curcumin [93]/EGCG [94]	Prebiotic nondigestible fiber-rich diet [95] Grapes, wine, and peanuts [83]	Akkermansiaceae

Abbreviations: EGCG: Epigallocatechin gallate; FMT: Fecal microbiota transfer; ICI: Immune checkpoint inhibitors; NSCLC: Non-small-cell lung carcinoma; RCC: Renal cell carcinoma.

) and correct the dysbiosis created by oxidative stress, as shown in alcoholic liver disease and fatty liver disease models. Similarly, phytochemicals could correct dysbiosis in patients treated with ICIs. These phytochemicals are most commonly found in fiber-rich diets (Table 3). Frankel et al. reported indirect evidence of this strategy’s possible benefit in melanoma patients [14]. Metabolomic analyses in 39 ICI-treated melanoma patients revealed that the anacardic acid levels were increased in responders. Most of these patients (five out of six) had dietary habits that explained the high anacardic acid levels. Based on these points, we think using phytochemicals to correct dysbiosis and increase levels of bacteria associated with ICI efficacy should be exploited in clinical trials. Additionally, further research is needed to delineate the optimal fiber type to consume for ICI efficacy. Recently, Nakajima et al. reported that a soluble fiber diet increased gut Bacteroides fragilis, previously associated with ICI response, while the insoluble fiber-rich diet reduced the Bacteroides fragilis levels in a mice model [74]. In addition, Li et al. previously demonstrated the enrichment of beneficial commensal Actinobacteria and Akkermansia in obese mice fed with nondigestible fructans, which are found in higher concentrations in bananas, compared to mice fed with cellulose [75,76]. Until the results of these trials become available, the recommendation of eating soluble fiber-rich diets and diets rich in phytochemicals that increase beneficial commensal bacteria could benefit ICI-treated patients.

6. Future Perspectives

Phytochemicals have several effects on promoting anti-tumor immunity and modulating the tumor microenvironment, including immune checkpoints. Based on the available preclinical evidence, phytochemicals have the potential to transform immunologically cold tumors into hot tumors and improve immunotherapy efficacy. This strategy could be especially promising for tumors that have consistently garnered less benefit from ICIs, including sarcomas and brain tumors [104,105]. The available preclinical evidence and the demonstration of higher ICI response rates in melanoma patients with increased anacardic acid levels in metabolomic studies support the progression of phytochemical and ICI combinations to clinical studies [14]. Patient-derived xenograft models would be suitable venues to assess the clinical use of phytochemicals in transforming tumor immune profiles and would be beneficial for further clinical studies.

Combining chemotherapy and ICIs and combining two ICIs (CTLA-4 and PD-1/PD-L1) has become the standard of care in the first-line treatment of NSCLC, gastric cancer, RCC, and melanoma, with improved response rates and survival [106,107,108,109,110]. The increased immune activation in the tumor microenvironment was among the main drivers of the increased efficacy of these combinations [111]. However, the efficacy of adding phytochemicals to these combinations has not been investigated yet. We think that phytochemicals could add benefits to these combinations, and further research focusing on the efficacy of adding phytochemicals to chemotherapy-ICI and ICI-ICI combinations could have implications for current clinical practice.

Another knowledge gap concerns the efficacy of using phytochemical and ICI combinations in adjuvant settings. Although it would be harder to measure the efficacy of phytochemical and ICI combinations in adjuvant settings, animal models with radiated tumors or chemoprevention models using ICI and phytochemical combinations should be designed considering the well-known efficacy of phytochemicals in chemoprevention.

Lastly, the phytochemical structure of the traditional herbs with well-known immunomodulatory effects should be thoroughly delineated to identify more candidates to combine with ICIs. For example, Artemisia products are used in China to fight malaria, allergies, and auto-immune diseases [112]. The recently conducted phytochemical analysis of Artemisia annua L. demonstrated the presence of several flavonoids, hydroxycoumarins, and phytosterols, phytochemicals with possible anticancer effects on immune checkpoints [113]. Furthermore, recent gene expression analyses in HCC about Artemisia scoparia demonstrated the increased expression of BIRC5 and secondary expression of CTLA-4 and LAG-3 immune checkpoints and a possible immune activation with this herbal medicine [114]. Further studies evaluating the phytochemical constituent of herbal medications and corresponding effects on the immune checkpoints are needed.

7. Conclusions

In conclusion, plant-based phytochemicals could be beneficial adjuncts to ICIs with improved immune activation and tumor microenvironment modulation. Further research is needed considering the difficult scenarios in the clinical practice and the areas needing improvement to answer more clinically oriented questions.