Association of Short-Chain Fatty Acids in the Gut Microbiome With Clinical Response to Treatment With Nivolumab or Pembrolizumab in Patients With Solid Cancer Tumors

IMPORTANCE Immunotherapy using immune checkpoint inhibitors has been remarkably effective for treating multiple cancer types, and the gut microbiome is a possible factor affecting immune checkpoint inhibitor efficacy. However, the association between the gut microbiome and immune status of the tumor microenvironment remains unclear. Short-chain fatty acids (SCFAs) are major end product metabolites produced by the gut microbiota and have wide-ranging impacts on host physiology. OBJECTIVE To evaluate fecal and plasma SCFAs in patients with solid cancer tumors treated with programmed cell death-1 inhibitors (PD-1i). DESIGN, SETTING, AND PARTICIPANTS This was a prospective cohort biomarker study of patients with cancer who planned therapy with PD-1i at Kyoto University Hospital between July 2016 and February 2019. Data were analyzed from October 2019 to February 2020. EXPOSURES Patients who were treated with nivolumab or pembrolizumab were classified into 2 groups based on their treatment response using Response Evaluation Criteria in Solid Tumors version 1.1: responders who achieved an objective response and nonresponders. Dietary information in terms of intake frequency was obtained. Concentrations of SCFAs in fecal and plasma samples collected before PD-1i administration were measured using ultra-high-performance liquid chromatography coupled with tandem mass spectrometry. MAIN OUTCOMES AND MEASURES The concentration of SCFAs and progression-free survival. RESULTS Among 52 patients enrolled, the median (range) patient age was 67 (27-84) years, and 23 (44%) were women. Median (range) duration of follow-up of the survivors after administration of PD-1i was 2.0 (0.4–4.1) years. The overall response rate was 28.8%. High concentrations of some SCFAs were associated with longer progression-free survival. These included fecal acetic acid (hazard ratio [HR], 0.29; 95% CI, 0.15-0.54), propionic acid (HR, 0.08; 95% CI, 0.03-0.20), butyric acid (HR, 0.31; 95% CI, 0.16-0.60), valeric acid (HR, 0.53; 95% CI, 0.29-0.98), and plasma isovaleric acid (HR, 0.38; 95% CI, 0.14-0.99). CONCLUSIONS AND RELEVANCE Results of this study suggest that fecal SCFA concentrations may associated with PD-1i efficacy; thus, SCFAs may be the link between the gut microbiota and PD-1i efficacy. Because fecal examinations are completely noninvasive, they may be applicable for routine monitoring of patients. JAMA Network Open. 2020;3(4):e202895. doi:10.1001/jamanetworkopen.2020.2895 Key Points Question Are short-chain fatty acids associated with clinical outcomes in patients with solid cancer tumors treated with programmed cell death 1 inhibitors? Findings In this cohort study of 52 patients with solid tumors, high concentrations of fecal acetic acid, propionic acid, butyric acid, and valeric acid were significantly associated with longer progression-free survival. Meaning The findings suggest that fecal short-chain fatty acid concentrations may be a potential biomarker to identify patients with solid tumors who could benefit from treatment with programmed cell death 1 inhibitors. Author affiliations and article information are listed at the end of this article. Open Access. This is an open access article distributed under the terms of the CC-BY License. JAMA Network Open. 2020;3(4):e202895. doi:10.1001/jamanetworkopen.2020.2895 (Reprinted) April 16, 2020 1/12


Introduction
Immunotherapy using immune checkpoint inhibitors (ICIs), including programmed cell death 1 inhibitors (PD-1i) and cytotoxic T-lymphocyte antigen 4 inhibitors, given as monotherapies, has consistently demonstrated a long-term survival benefit with durable responses and disease stabilization in patients with untreated or previously treated advanced melanoma. [1][2][3] Immune checkpoint inhibitors have been remarkably effective across multiple cancer types. However, the response rate of PD-1i for solid cancer was relatively low. An optimal biomarker of the response to ICIs is critically needed for clinical decision-making.
Studies of various tumor types [4][5][6] have suggested that the gut microbiome profile is a possible factor associated with efficacy of ICIs. Several preclinical and clinical studies have supported an association between the gut microbiome and the efficacy of ICIs, but how this association functions in the tumor microenvironment remains unclear. Short-chain fatty acids (SCFAs) are major end product metabolites produced by the gut microbiota and have wide-ranging impacts on host physiology. The SCFAs have been confirmed to modulate immune cell response. The objective of this study was to evaluate fecal SCFAs in patients with solid cancer tumors treated with a PD-1i.

Methods
This was a prospective study of patients with cancer who were treated with PD-1i at Kyoto University Patients received either nivolumab (2 mg/kg every 3 weeks, 3 mg/kg every 2 weeks, or 240 mg every 2 weeks) or pembrolizumab (200 mg every 3 weeks). All patients were asked about their average frequency and amount of intake of various foods and their dietary habits during the 1 year preceding the onset of their current cancer. Dietary information, including beef or pork, chicken, fish, beans, green vegetables, cabbage, potato, radish, pumpkin, mushroom, seaweed, fruit, and yogurt, was obtained in terms of intake frequency. The following information was obtained from medical records and radiological images: dates of treatment initiation and final date of treatment, age, sex, Eastern Cooperative Oncology Group performance status, primary cancer site, metastatic site(s), laboratory test results at the start of treatment, treatment efficacy according to the Response Inc. Methanol, acetonitrile, and formic acid were obtained from Fujifilm Wako Chemical Co. Purified water was obtained using a PURELAB flex 5 system (ELGA Co). Stock solutions were adjusted using methanol. Concentrated solutions of mixed standard solutions were diluted as required by adding methanol for derivatization.

Analysis of SCFAs in Fecal and Plasma Samples
We used the Waters Acquity H Class ultra-performance liquid chromatography system (Waters Co).

Statistical Analysis
All patient characteristics were classified as categorical variables with the exception of age, which was analyzed as a continuous variable. The Mann-Whitney test was performed to evaluate the SCFAs affecting the treatment response. The SCFA concentrations were dichotomized according to the cutoff value using a recursive partitioning analysis. A recursive partitioning analysis was performed to select for the variables in the SCFAs. Dietary information was categorized into 2 groups: less than 3 times/week and 3 times/week or more. Logistic regression analysis was used to obtain odds ratios (ORs) and 95% confidence interval to measure associations. Progression-free survival was defined from the time of treatment initiation to disease progression or death due to other causes. Time-toevent distributions were estimated using the Kaplan-Meier method. To evaluate the impact of each

JAMA Network Open | Oncology
Short-Chain Fatty Acids in the Gut Microbiome and Response to PD-1 Inhibitors factor on progression-free survival, univariate and multivariate Cox proportional hazards analyses were applied; therefore, the measure of association in this study was the hazard ratio (HR) together with the 95% CI. Variables in multivariate analysis were age (<65 years vs Ն65 years), primary cancer (melanoma vs nonmelanoma), number of organ metastases (<3 vs Ն3), and those that had selected in a recursive partitioning analysis. P < .05 using 2-tailed testing was considered indicative of statistical significance. Statistical analyses were performed using Stata version 15

The Associations Between SCFA Concentrations and PD-1i Efficacy
For the analytical validation, our previous study was reported using biological samples. 7 Patient characteristics and the concentrations of fecal and plasma SCFAs in the responder and nonresponder groups are summarized in Table 2. There were no significant differences between the responder and  The overall response and disease control rates of all patients were 28.8% and 46.1%, respectively.

The Associations Between SCFA Concentrations and Dietary Habits
The ORs for the associations between fecal SCFA concentrations and dietary intake are shown in Table 4. The fecal BA concentration was significantly associated with a high frequency of mushroom intake (OR, 6.17; 95% CI, 1.20-31.7); the fecal VA concentration was significantly associated with high frequencies of green vegetable intake (OR, 4.50; 95% CI, 1.30-15.5), cabbage intake (OR, 3.76; 95% CI, 1.01-13.9), and mushroom intake (OR, 3.66; 95% CI, 1.04-12.9); and the fecal hydroangelic acid concentration was significantly associated with a high frequency of green vegetable intake (OR, 4.25; 95% CI, 1.19-15.1). There was no significant association of the fecal concentration of AA, PA, IBA, IVA, CA, or SA with the intake of any food evaluated.

JAMA Network Open | Oncology
Short-Chain Fatty Acids in the Gut Microbiome and Response to PD-1 Inhibitors

JAMA Network Open | Oncology
Short-Chain Fatty Acids in the Gut Microbiome and Response to PD-1 Inhibitors

Discussion
To our knowledge, this is the first study to find an association between SCFAs and the efficacy of PD-1i treatment for solid cancers. Several reports explored gut microbiome profiling in association with the response to anti-PD-1 antibody treatment in patients with various solid tumors. Although the mechanisms of the association between the gut microbiome and the response to ICIs remain unclear, gut microbial metabolites are thought to link the gut microbiome to systemic immunity. Short-chain fatty acids are the end products of gut microbe-mediated metabolism and the most abundant product of anaerobic fermentation of dietary fibers in the intestine.
Several reports have evaluated gut microbiome profiling in association with the response to PD-1i treatment, but the results differ among those reports. 4-6 Even 2 of those studies 5,6 that were conducted in the same country, both in melanoma patients, identified different bacteria species associated with the response to PD-1i treatment. However, the species associated with the PD-1i response, such as Faecalibacterium prausnitzii, Akkermansia muciniphila, Collinsella aerofaciens, and Bifidobacterium adolescentis, are BA-or PA-producing bacteria. [8][9][10] Our study found that high concentrations of fecal or plasma SCFAs were associated with a response to PD-1i treatment and longer progression-free survival. Short-chain fatty acids exhibit immunomodulatory functions in the host, affecting CD4 + T cells and antigen-presenting cells. Butyric

JAMA Network Open | Oncology
Short-Chain Fatty Acids in the Gut Microbiome and Response to PD-1 Inhibitors acid has been shown to induce FOXP3 + CD4 + Treg differentiation. [11][12][13] However, a recent report found that BA and other SCFAs increase the expression of IFNγ and granzyme B in CD8 + cytotoxic T lymphocytes and interleukin-17-secreting CD8 + T cells. 14 Another recent report showed that SCFA administration led to the recovery of an impaired immune response. 15 Another function of SCFAs, particularly BA and VA, is inhibition of histone deacetylases 14,16,17 ; histone deacetylase inhibition was found to upregulate PD-1 ligands in melanoma cells, enhance the immunotherapy response, inhibit apoptosis of CD4 + T cells within tumors, upregulate the antitumor immune response, and suppress tumor growth. [18][19][20] Histone deacetylase inhibitors have been shown to possess potent immunomodulatory activity. The combination of an ICI and histone deacetylase inhibitors has shown promising results. 21,22 Fecal SCFAs are rapidly absorbed by the colon in a concentration-dependent manner. Butyric acid, in particular, is a local nutrient for mucosal cells. Propionic acid is absorbed by the portal vein and mainly used by the liver as a major hepatic gluconeogenic substrate. 23 Thus, the concentrations of SCFAs in the peripheral serum are lower than those in the hepatic vein or portal vein. 24,25 Dietary intake is closely related to fecal excretion of several substances. The composition of gut microbiome changes by changing the food components and dietary habits. 26  Furthermore, a high frequency of mushroom intake was associated with longer progression-free survival. However, dietary information in this study was not during treatment using ICI, and was during the 1 year preceding the onset of their current cancer. Further study is needed in order to Abbreviations: AA, acetic acid; BA, butyric acid; CA, caproic acid; HA, hydroangelic acid; IBA, isobutyric acid; IVA, isovaleric acid; NA, not available; PA, propionic acid; SA, succinic acid; SCFA, short-chain fatty acid; VA, valeric acid.
a Odds ratios are shown for consumption of each food 3 times per week or more compared with fewer than 3 times per week (reference).