Comment by Marc Rivard, M.D. and raelian bishop:
Immunotherapy is a very promising approach for cancers.
However, some people respond less well to it. This small study
demonstrates the potential role of the intestinal microbiota.
Indeed, by performing a fecal transplant from healthy patients
to cancer patients, the latter respond better to immunotherapy.
In certain immunotherapy protocols, if a person has taken
antibiotics before the sessions, a fecal transplant is often
carried out. Antibiotics save lives, but you have to be certain
that there is a real medical indication to take them. Always ask
your doctor if it is really necessary to take antibiotics. They
are often prescribed too quickly.
Fecal
microbiota transplantation improves anti-PD-1 inhibitor efficacy
in unresectable or metastatic solid cancers refractory to
anti-PD-1 inhibitor
FMT
with anti-PD-1 shows benefits in advanced solid
cancers resistant to anti-PD-1
•
FMT
induced partial response to anti-PD-1 in R7 recipients
with enhanced immunity
•
Lactobacillus
salivariusandBacteroides plebeiusmay
inhibit T cell activity
•
Prevotella merdaeImmunoactis
boosts T cell activity and reduces tumor growth
Summary
The
gut microbiome significantly influences immune responses
and the efficacy of immune checkpoint inhibitors. We
conducted a clinical trial (NCT04264975) combining an
anti-programmed death-1 (PD-1) inhibitor with fecal
microbiota transplantation (FMT) from anti-PD-1 responder
in 13 patients with anti-PD-1-refractory advanced solid
cancers. FMT induced sustained microbiota changes and
clinical benefits in 6 of 13 patients, with 1 partial
response and 5 stable diseases, achieving an objective
response rate of 7.7% and a disease control rate of 46.2%.
The clinical response correlates with increased cytotoxic
T cells and immune cytokines in blood and tumors. We
isolatedPrevotella
merdaeImmunoactis from a responder to
FMT, which stimulates T cell activity and suppresses tumor
growth in mice by enhancing cytotoxic T cell infiltration.
Additionally, we foundLactobacillus salivariusandBacteroides plebeiusmay
inhibit anti-tumor immunity. Our findings suggest that FMT
with beneficial microbiota can overcome resistance to
anti-PD-1 inhibitors in advanced solid cancers, especially
gastrointestinal cancers.
Immune
checkpoint inhibitors (ICIs) targeting cytotoxic T
lymphocyte antigen (CTLA-4) or programmed death-1/programmed
death-ligand 1 (PD-1/PD-L1) pathways have transformed cancer
treatment across a wide range of cancer types. However, ICIs
have not universally demonstrated effective outcomes across
all cancer types, and their efficacy in those where they
have shown success, such as gastroesophageal, liver,
bladder, and head and neck cancers, remains modest.
Although recent research has progressed in
elucidating predictive biomarkers, understanding resistance
mechanisms in immuno-oncology, and improving treatment
outcomes by combining ICIs with other therapies, overcoming
primary and secondary resistance remains significantly
challenging.
The
gut microbiome has emerged as a key player in shaping immune
responses and affecting the efficacy of ICIs. Studies have
shown that the presence of specific commensals in patient
stool samples correlates with clinical response to ICIs.
Preclinical experiments have demonstrated that
a specific gut microbiome or fecal microbiota
transplantation (FMT) using fecal material from patients
responding to ICIs can induce tumor regression, augment
T cell responses, and improve the anti-tumor efficacy of
ICIs.
Recent pilot clinical trials have shown that
the combination of anti-PD-1 inhibitors and FMT could
potentially overcome resistance to anti-PD-1 therapy by
altering the gut microbiome and reprogramming the tumor
microenvironment in a subset of patients with
anti-PD-1-refractory melanoma.
In this study, we aimed to evaluate the
potential of FMT in overcoming resistance to anti-PD-(L)1
inhibitors in patients with advanced solid cancers
refractory to these inhibitors (NCT04264975) and identify
specific commensal bacteria that play a causal role in the
efficacy of FMT and ICI treatments.
Results
Study
design and participant characteristics
This
FMT trial was a prospective, single-arm, and single-center
study of the combination of FMT and an anti-PD-(L)1
inhibitor (Figure 1A).
The study included two distinct participant categories:
donors and recipients (Figures 1A
andS1).
Donors were characterized by a durable complete response
(CR) or partial response (PR) according to Response
Evaluation Criteria in Solid Tumors (RECIST) v1.1 for at
least 6 months following anti-PD-(L)1 monotherapy for
unresectable or metastatic solid tumors. Recipients, on
the other hand, were individuals with confirmed disease
progression during anti-PD-(L)1-based therapy,
monotherapy, or combination therapy for unresectable or
metastatic solid tumors. To eliminate the native gut
microbiota, recipients were given oral antibiotics of
amoxicillin clavulanate for a duration of 5 days before
FMT (only prior to the first FMT). FMT procedures were
conducted via colonoscopy, followed by the continuation or
reintroduction of an anti-PD-(L)1 inhibitor (administered
at the standard dose and schedule) until either
unacceptable toxicity or disease progression occurred.
Subsequent FMT sessions from the same or different donors
were permitted, at the investigator's discretion, with a
time interval of 2–4 weeks prior to the initial response
assessment or in cases where clinical response was not
evident. Sequential samples of blood, stool, and tumor
biopsies were collected both before and after FMT.
Response assessment was carried out every 6–8 weeks using
computed tomography (CT) scans, following RECIST v1.1.
Figure 1Clinical responses
to FMT combined with nivolumab in patients
with advanced solid cancer refractory to
nivolumab
Between
January 2019 and August 2020, we recruited 13 FMT
recipients with metastatic gastric cancer (GC) (n = 4), esophageal
squamous cell carcinoma (ESCC) (n = 5), and
hepatocellular carcinoma (HCC) (n = 4) (Tables 1andS1).
The median age was 60 years (range, 38–76), and all had
undergone heavy prior treatment (median of 10 systemic
treatments, range, 4–34). All patients had confirmed
disease progression on nivolumab monotherapy, with six
(46.2%) displaying primary resistance and seven (53.8%)
displaying secondary resistance. Enrollment occurred
immediately after confirming disease progression, and FMT
was administered while continuing nivolumab. All
recipients’ tumors were microsatellite stable (MSS), with
seven (53.8%) having PD-L1-positive (combined positive
score [CPS] ≥ 1) and four (30.8%) having PD-L1 CPS ≥ 5. We
recruited six FMT donors: four with HCC and one each with
GC and ESCC. These donors achieved and maintained a CR (n = 4) or PR (n = 2) for at least
1 year with nivolumab or pembrolizumab, except one with an
ongoing response of 7.9 months (Table 2).
All donors’ tumors were MSS and PD-L1-negative, except for
one donor with low PD-L1 expression (CPS of 7). The median
tumor mutation burden (TMB) in the donors was 10.9
mutations per megabase (range, 9.4–37.5).
Table 1Profiles
of recipients in FMT study (n = 13)
Administration
of study treatment and treatment-related AEs
All
patients were treated with a combination of nivolumab with
FMT, which involved continuing nivolumab treatment after
failure to prior nivolumab therapy. A median cycle of
nivolumab after FMT was five cycles (range, 1–27). Of the
13 recipients, 7 received subsequent FMTs from the same (n = 4) and/or
different donors (n =
4) after the first FMT. Among them, one recipient first
underwent subsequent FMTs from the same donor and then
from a different donor.
Regarding
safety profiles, treatment-related adverse events (AEs)
were minimal (Table S2).
Out of the thirteen patients, seven patients (53.8%)
experienced at least one treatment-related AE, all of
which were either grade 1 or 2, except for one patient who
experienced grade 3 immune-related gastritis. The most
common treatment-related AEs were skin pruritus (n = 5, 38.5%),
followed by skin rash (n =
2, 15.4%), and hypothyroidism (n = 2, 15.4%).
Endocrine AEs were managed with hormone replacement
therapy, and immune-related gastritis (CTCAE; Common
Terminology Criteria for Adverse Events, grade 3) was
managed with systemic corticosteroids.
Efficiency
of FMT and clinical response after FMT
To
evaluate the efficiency of FMT engraftment, we performed
16S rRNA sequencing analysis on stool samples obtained
from both recipients and donors. Analysis using
Bray-Curtis distance (BCD) revealed that the microbial
compositions of most recipients underwent significant
changes from their baseline composition after FMT (Figures 1B
andS2).
To quantify these changes, we calculated the engraftment
distance (ED) based on the BCDs between baseline of
recipient and donor samples (Figure S3).
A decrease in the ED from baseline indicates a change in
microbial composition. Since all recipients in our FMT
trial exhibited reduced ED, this demonstrates that our FMT
trial induced alterations in the microbial compositions of
all recipients (Figure 1C).
Remarkably, the microbial compositions of recipient#7 (R7)
closely resembled those of the respective donor as early
as 1 day after both the first and second FMT procedures,
and this similarity persisted for 282 days following the
second FMT procedure (Figure 1B).
Among
the thirteen recipients, one achieved a PR, and five
exhibited stable disease (SD) following FMT in conjunction
with continued nivolumab treatment. These outcomes,
representing an objective response rate of 7.7% (1/13) and
a disease control rate of 46.2% (6/13), underscore the
combined therapeutic impact of FMT and ongoing
immunotherapy (Figure 1D).
Notably, R7, who was initially metastatic HCC with primary
resistance to nivolumab, achieved a PR following FMT with
continued immunotherapy and exhibited a significant
reduction in tumor size (Figures 1D
and 1E). Our analysis demonstrates that FMT with
immunotherapy can effectively modify the gut microbiome of
recipients, leading to significant clinical efficacy,
particularly exemplified by the remarkable case of R7.
Immune
changes in systemic and tumor microenvironments and
clinical outcomes following FMT
We
focused on R7 and observed longitudinal changes in several
factors. Initially, R7 showed progressive disease (PD)
with a 22.4% increase in target lesions 6 weeks post the
first FMT from donor #1 (D1) (Figures 2A
and 2B). After the second FMT from donor #5 (D5), tumor
progression slowed, with only a 13% increase in target
lesions compared to the first PD assessment. 8 weeks after
the second FMT, there was a significant 30.5% reduction in
tumor size (Figures 2A
and 2B, top). Eventually, R7 achieved a PR with a maximum
tumor size reduction of 47.7% from baseline before first
FMT (Figures 2A
and 2B). Tumor markers, including serum α-fetoprotein
(AFP) and protein induced by vitamin K absence or
antagonist-II (PIVKA-II), also significantly decreased
8 weeks post-second FMT (Figure 2B,
bottom).
Figure 2Analysis of
longitudinal tumor reduction and
immune-related changes in recipient #7
Using
CyTOF (Cytometry by time of flight), we evaluated the
systemic immune changes with the response to FMT combined
with nivolumab. R7 showed gradual increase in CD8+T cells
and CD8+CCR7−CD27−terminal
effector T (Tte) cells and a decrease in CD4+CD25+CD127−CCR4+regulatory
T (Treg) cells from baseline (Figures 2C
and 2D). At 23 weeks post-second FMT, Tte cells increased
by 255% and Treg cells decreased by 81% (Figure 2C).
These immune changes correlated with R7’s clinical
response and were more pronounced after the second FMT (Figure 2D).
This immune activation was absent or modest in other
patients (Figure S4).
Using the Simoa HD-1 Analyzer and Simoa SP-X, we found
elevated levels of interferon (IFN)-γ, tumor necrosis
factor alpha (TNF-α), interleukin (IL)-2, IL-7, IL-15, and
IL-6 in R7 specifically after the second FMT but not the
first FMT, which correlated with the clinical response (Figure 2E).
On the other hand, R6, R9, and R14, who achieved a SD,
showed an upward trend in levels of some of these
cytokines, including IL-6, following FMT (Figure S5).
This suggests that although FMT did not lead to a
significant reduction in tumor size for these patients, it
still resulted in elevated level of these cytokines to
some extent. In our assessment of tumor-infiltrating
immune cells, which play a critical role in the response
to anti-PD-1 therapy, using multiplex immunohistochemistry
(IHC), we observed notable changes in R7. 4 weeks after
the second FMT, R7 showed a significant increase in
tumor-infiltrating cytotoxic T cells (from 14 to 361
cells/mm2)
and major histocompatibility complex (MHC)-II+M1
macrophages (from 25 to 174 cells/mm2),
while the number of CD4+FOXP3+CD8−CD20−Treg
cells remained extremely low (from 5 to 4 cells/mm2).
In contrast, no comparable immune activation was noted
subsequent to the first FMT (Figures 2F
and 2G). This highlights a noteworthy immune activation
specifically following the second FMT. R7’s enhanced
immune response was not mirrored in all patients. In R8
(best response, PD) and R14 (best response, SD), although
increases in cytotoxic T cells were also observed, M1
macrophage increases were absent or minimal (Figure S6).
Overall, R7 showed notable tumor reduction and enhanced
immune response following the second FMT, with significant
changes in systemic and tumor-infiltrating immune cells.
Following
the second FMT and continued immunotherapy, R7 displayed
significant tumor reduction and an enhanced immune
response. The CT scan showed diffuse gastric wall
thickening, while esophagogastroduodenoscopy (EGD)
revealed multiple gastric ulcers, confirmed as acute
lymphocytic gastritis leading to a PR at 8th week
post-treatment. However, 11 weeks after the second FMT,
the patient reported progressive epigastric pain and
nausea, diagnosed as immune-related gastritis via EGD and
abdominal CT scan (Figures S7A
and S7B). The CT scan showed diffuse gastric wall
thickening, while EGD revealed multiple gastric ulcers,
confirmed as acute lymphocytic gastritis (Figure S7B).
This suggested an overactive immune response in the
intestine. In order to manage the patient’s symptoms,
systemic corticosteroid treatment was initiated
approximately 12 weeks after the second FMT and continued
long term due to fluctuating symptoms. Despite the
potential negative impact of prolonged steroid use, R7
experienced disease progression with a PFS
(Progression-Free Survival) of 8.7 months post-second FMT.
A subsequent FMT was planned using the same D5 fecal
material. Unfortunately, the patient sustained a femoral
fracture requiring surgery. After recovery, the third FMT
resulted in SD, but the fourth FMT failed to halt tumor
growth (Figures S7A
and S7C). Multiplex IHC analysis of tumor biopsies before
and after the third FMT showed a mild increase in
cytotoxic T cells, MHC-II+M1
macrophages, and helper T cells, though not as pronounced
as after the second FMT (Figure S7D).
Therefore, in our subsequent analysis aimed at identifying
potential causative bacterial strains, we intended to
focus on examining the changes that occurred during the
second FMT while excluding the third and fourth FMTs.
Uncovering
bacterial strains associated with clinical outcomes
after FMT
To
identify the specific microbiota responsible for the
clinical benefits of FMT combined with an anti-PD-1
inhibitor, we analyzed microbial changes in R7 using 16s
rRNA sequencing. Our analysis showed that following the
first FMT from D1,Bacteroideslevels
increased from 2.87% to 6.4%, whilePrevotellalevels
decreased from 19.75% to 0% (Figure 3A).
Conversely, after the second FMT from D5,Bacteroidesdecreased
from 6.4% to 0.92%, whilePrevotellaincreased
from 0% to 38.48% (Figure 3A).
This shift highlighted a reciprocal trend betweenPrevotellaandBacteroides(Figure 3B).
To pinpoint the bacteria contributing to the favorable
efficacy of FMT with anti-PD-1 therapy, we identified
bacteria significantly more prevalent in D5 than in D1 and
those that significantly emerged in R7 post-second FMT
compared to post-first FMT. This led to the identification
of 47 common bacteria (Figure 3C).
Excluding bacteria present in R7's baseline stool sample,
we narrowed the list to 34 unique bacteria.Prevotella merdae(type
strain; sp. Marseille-P4119)
andPrevotella stercorearanked
highest based on their LDA (Linear Discriminant Analysis)
scores (Figure 3C;Table S3).
These species were significantly more abundant in D5 and
R7 post-second FMT compared to D1 and R7 post-first FMT,
respectively (Figures 3D
andS8A).
Given the prominence ofP. merdaein
the stool samples after the second FMT and its high LDA
score, we prioritized further investigations on this
species, identifying it asP. merdaesp.
Marseille-P4119 by using clade-specific marker gene
analysis through shotgun metagenomic sequencing.
Figure 3Discovery of key
bacteria influencing clinical outcomes
post-FMT in recipient #7
To
explore the possibility of discovering a new strain ofP. merdae, we
isolatedP. merdaefrom
R7's stool sample and performed whole genome sequencing.
Comparative genomic analysis revealed that the isolatedP. merdaebelongs
toP. merdae(Table S4),
but the average nucleotide identity withP. merdaesp.
Marseille-P4119 was 97.41%, suggesting a potential
distinction. Phylogenetic analysis identified over 1,980
single-nucleotide polymorphisms (SNPs) in conserved
markers, indicating significant dissimilarity from the
reference genome ofP. merdaesp.
Marseille-P4119 (Figure 3E).
These genomic differences in marker genes suggest the
discovery of a new strain, which we designated asP. merdaeImmunoactis.
We hypothesize that this strain contributed to the
favorable clinical and immune response in R7 after the
second FMT.
Following
a similar logical framework, we identified bacteria that
might negatively impact clinical responses. We focused on
bacteria more prevalent in D1 than in D5, and in R7’s
stool after the first FMT compared to the second. This
analysis identified 16 common bacteria (Figure 3F;Table S5).
Among these bacteria,Bacteroides plebeius(B. plebeius)
exhibited the highest LDA score and was more abundant in
D1 and R7 post-first FMT than post-second FMT,
respectively (Figures
3F andS8B).B. plebeiusis
suggested to potentially cause unfavorable clinical
responses due to its immunosuppressive effects.
Despite
undergoing the second FMT with fecal material from the
same D5, R6 experienced only a brief duration of SD with a
progression-free survival of 2.1 months. Further microbial
comparison between R6 and R7 revealed thatP. merdae, which
was suspected to be an immunity-boosting strain, was
well-established in both after the FMT from D5 (Figure S8C).B. plebeius, which
potentially exhibits immunosuppressive properties, was not
detected at all in R6 (Figure S8C).
This prompted us to investigate why R6 did not experience
the expected therapeutic effect despite the presence ofP. merdae. We
identified significant bacterial differences between R6
and R7. Comparing stools from R6's and R7’s second FMT
from same D5, we noted a higher presence ofLactobacillus salivarius(L. salivarius) in
R6’s stool samples compared to R7’s (Figure 3G;Table S6).
Interestingly,L. salivariuswas
more abundant in D1 compared to D5 and in R7’s stool
samples after the first FMT compared to the second FMT (Figures 3G
andS8D).
These findings suggest thatL. salivariusmay
be linked to unfavorable clinical responses due to its
potential immunosuppressive properties, similar toB. plebeius.
In
our further investigation, we examined whether the ratio
of three bacteria (P. merdae/(B. plebeius +L. salivarius)) is
associated with clinical significance. Initially, we
investigated whether this ratio had an influence on
survival probability in an independent cohort of patients
with unresectable or metastatic solid cancers who were
treated with nivolumab monotherapy (Table S7;
seeSTAR
Methods). This cohort was specifically established
for microbiome biomarker research and included a total of
72 patients. All the included patients had baseline fecal
samples available and had not received antibiotics within
the 4 weeks preceding treatment. Interestingly, our
observations confirmed that a higher ratio was associated
with prolonged survival (Figures S9A–S9C).
Furthermore, we observed a tendency of increasing
microbial ratios in R7 following the second FMT, which
positively correlated with the ratio of effector CD8+T cells
and Tregs populations (Figures 3H
andS9D).
In fact, both these microbial and immunological trends
intensified after the second FMT. These findings highlight
the intricate impact of the gut microbiota on the efficacy
of FMT and anti-PD-1 therapy. Further research is required
to investigate the specific roles of these bacteria and
their influence on clinical outcomes.
The
influence ofP. merdaeImmunoactis
on enhancing immune response and cytotoxic activity
Our
comprehensive metagenomic analysis suggests thatP. merdaeImmunoactis
may be pivotal in overcoming resistance to anti-PD-1
inhibitors, whileB. plebeiusandL. salivariuscould
have opposing effects. To explore the immune-related
effects of these bacteria, we treated human T cells with
their respective conditioned bacterial supernatants. When
referring to “supernatant” in the following context, it
denotes the conditioned bacterial supernatant.P. merdaeImmunoactis’
supernatant significantly boosted the proliferation of CD4+and
CD8+T cells
(Figures
4A and 4B), whereas those fromB. plebeiusandL. salivariushindered
T cell proliferation (Figures S10A
and S10B). Additionally, CD8+T cells
treated withP. merdaeImmunoactis’
supernatants showed increased IFN-γ secretion compared to
controls, whileB. plebeiusorL. salivariussupernatants
led to decreased IFN-γ secretion (Figure 4C).
To investigate whether the two non-efficacious strains
hindered the effectiveness ofP. merdaeImmunoactis,
we conducted a parallel experiment using a three-bacteria
mixture. This mixture's proportions were derived from
fecal samples of R6’s post-2nd FMT and from D5, where all
three strains coexisted. Since, on average, they were
present at about one-tenth the level ofP. merdae,
three-bacteria mixture was composed in a ratio ofP. merdaeImmunoactis:B. plebeius:L. salivarius =
10:1:1 (Figure S10C).
The three-bacteria mixture significantly reduced T cell
proliferation (Figure S10D),
indicating that the non-efficacious strains curbed the
T cell proliferation triggered byP. merdaeImmunoactis.
We
observed significant differences in IFN-γ secretion among
the three strains (Figure 4C),
prompting us to investigate changes in the IFN-γ pathway’s
expression. OT-1 cells, from
C57BL/6-Tg(TcraTcrb)1100Mjb/Crl called by OT-1 mouse, were
activated with Ova-peptide and treated with supernatants
from each strain. It showed significant increases in key
immune cytokines in the “Ova act +P. merdaeImmunoactis”
group. A comparison of IFN-γ pathway-related gene
expression revealed obvious differences; there were
increases withP. merdaetreatment
and decreases withB. plebeiustreatment,
except forIrf-1andStat-2(Figure S10E).
This indicates differential activation of the IFN-γ
pathway by the two strains, leading to distinct immune
response outcomes.
We
found a significant increase in granzyme expression
levels following treatment withP. merdaeImmunoactis
supernatant, suggesting a potential enhancement in
cytotoxicity (Figure
S11A). This led us to exploreP. merdaeImmunoactis’
impact on CD8+T cell
cytotoxic activity using an OT-1 cell cytotoxicity assay (Figure S11B).
T cells activated byP. merdaeImmunoactis’
supernatants induced significant apoptosis and death among
pre-seeded MC38 cells (Figures 4D
andS11C),
indicating its potential to augment CD8+T cell
cytotoxicity against cancer. In subsequentin vivoinvestigation,
the combination ofP. merdaeImmunoactis
treatment and the anti-PD-1 inhibitor resulted in a more
significant reduction in tumor volume compared to either
treatment alone (Figure 4E).
Conversely, the two non-efficacious strains significantly
increasedin vivotumor
growth, further supporting the specific anti-cancer
efficacy ofP. merdaeImmunoactis
(Figure S12A).
We also tested whether the two non-efficacious strains
could interfere with the synergistic effect induced byP. merdaeImmunoactis
with anti-PD-1. WhileP. merdaeImmunoactis
showed a significant synergistic effect, the
three-bacteria mixture accelerated tumor growth (Figure S12B).
Additionally, we observed significantly reduced tumor
growth with the combination ofP. merdaeImmunoactis
and anti-PD-1, even in the non-immunogenic 4T1 model (Figure S12C).
In summary,P. merdaeImmunoactis
enhances responsiveness to anti-PD-1 therapy by boosting
immunity, but the presence of non-efficacious strains may
diminish these synergistic effects.
To
evaluate the potential enhancement of immune responses byP. merdaeImmunoactis
administration, we analyzed the alterations in immune gene
expression and immune cell populations in tumor tissues
and lymph nodes. Tumors treated with bothP. merdaeImmunoactis
and the anti-PD-1 inhibitor exhibited increased expression
ofIfn-γandCxcl10while
showing decreased expression ofCcl22compared
to tumors treated with the anti-PD-1 inhibitor alone (Figure 4F).
Treatment with the combination ofP. merdaeImmunoactis
and the anti-PD-1 inhibitor resulted in a significant
increase in tumor-infiltrating T cells within tumor
tissues, while the proportion of Treg cells among the CD45+T cells
remained unchanged (Figure 4G).
Notably, the combined treatment of the anti-PD-1 inhibitor
andP. merdaeImmunoactis
significantly increased the population of CD8+T cells,
CD8+CD44+CD62L−(CD8+Effector)
T cells, and CD8+PD-1+T
cells within the tumor microenvironment, compared to
anti-PD-1 treatment alone (Figures 4H–4J).
Additionally, we observed an increased ratio of activated
natural killer (NK) (NKAct)
to exhausted NK (NKEx)
withP. merdaeImmunoactis
treatment (Figure 4K).
However, these cell populations remained unchanged in the
lymph nodes (Figure S12D).
Taken together, these findings collectively suggest the
potential ofP. merdaeImmunoactis,
particularly when combined with the anti-PD-1 inhibitor,
to stimulate immune cell proliferation and enhance
cytotoxic activity, thereby promoting a more effective
anti-tumor immune response within the tumor
microenvironment.
Discussion
Our
findings demonstrate that FMT incorporating effective
microbiota has the potential to overcome resistance to
anti-PD-1 inhibitors in advanced solid cancers. This study
particularly emphasizes the treatment of gastrointestinal
cancers, including GC, ESCC, and HCC, which are known to be
immunologically challenging to treat. Our study provides
proof-of-concept for the effectiveness of FMT in combination
with anti-PD-1 inhibitors in solid cancers beyond melanoma,
expanding upon previous pilot studies that showed the
efficacy of FMT combined with anti-PD-1 inhibitors in
patients with melanoma refractory to anti-PD-1 treatment.
Among
the 13 recipients, one with HCC achieved a PR and five had
SD, resulting in a 7.7% response rate and 46.2% disease
control rate. Concerns might arise whether this efficacy
stemmed from FMT or continued ICI beyond progression, known
for certain cancers like NSCLC (Non-small cell lung cancer)
and melanoma.
ICIs can induce atypical tumor responses,
including pseudoprogression or delayed response,
contributing to the benefit of continued treatment. However,
all patients in our study had documented disease progression
on anti-PD-1 therapy, confirmed by consecutive CT scans at
least 4 weeks apart, indicating ongoing PD. A tumor biopsy
before FMT did not show massive immune infiltration typical
of pseudoprogression, ruling out pseudoprogression in
anti-PD-1 monotherapy. Of five patients showing SD post-FMT
with nivolumab, one (R6) had HCC, while four (R2, R5, R9,
and R14) had ESCC. In these cancer types, the literature is
sparse on how commonly SD occurs and how long it
persists with continued ICIs beyond progression. A
small-sized retrospective study reported a 5.8% response
rate, a 38% SD rate, and 3.7 months of progression-free
survival in advanced HCC patients treated beyond progression
(n = 34).
However, this study included not only
anti-PD-(L)1 inhibitors but also an anti-PD-L1 plus an
anti-VEGF (Vascular endothelial growth factor) inhibitor,
and initial disease progression was not confirmed with
subsequent CT scans as in our study. In esophageal cancer,
there is no literature suggesting the benefit of continuing
ICIs beyond progression. Our four heavily pre-treated ESCC
patients (all 4th- or 5th-line setting) showed
progression-free survival of 2.8, 2.8, 2.7, and 4.6 months,
respectively. Tumor growth on nivolumab monotherapy,
confirmed on consecutive CT scans before FMT, was slowed
down after FMT plus nivolumab, suggesting an actual effect
of the combination. One HCC patient (R6) receiving FMT plus
nivolumab as 3rd-line therapy also showed 3.1 months of
progression-free survival and tumor shrinkage after FMT plus
nivolumab. These findings suggest that the clinical benefit
observed after FMT in our study is not solely from the
continuation of the anti-PD-1 inhibitor but from the
beneficial effect of combining FMT with the anti-PD-1
inhibitor.
Previous
studies on FMT and cancer immunotherapy have provided
insights into the general impact of microbiome changes on
therapeutic outcomes, with some, such as Routy et al.,
Our study expands this field to include
non-melanoma cancers, such as HCC. However, in our study, a
response was observed only in HCC, making it challenging to
determine if this response is specific to HCC due to the
small sample size, which includes three cancer types and
only one HCC responder. Additionally, we involved six donors
because it was unclear which donor might possess the
causative microbiome for FMT efficacy before the FMT.
Furthermore, we attempted to match cancer types between
recipients and donors whenever possible. Consequently,
patients with GC and ESCC did not receive FMT using stool
samples from D5, the only donor who was an HCC patient and
induced a tumor response in the FMT trial. While additional
clinical studies are needed to further elucidate the
efficacy of combining FMT with ICIs in these cancer types,
our research nonetheless demonstrates the potential for
applying FMT in non-melanoma cancers. This highlights an
important step toward broader therapeutic applications and
the need for further investigation to optimize treatment
strategies.
Additionally,
we identified specific bacterial strains likeP. merdaeImmunoactis,L. salivarius, andB. plebeiusand
examined their distinct immunological roles in the context
of anti-PD-1 therapy. We discovered and isolated a strain,Prevotella merdaeImmunoactis,
from a patient (R7) who exhibited an excellent response to
FMT, suggesting it as a potential causative bacterium for
the therapeutic effects of FMT. Moreover, two
non-efficacious strains,L. salivariusandB. plebeius, were
identified and found to inhibit T cell activation,
potentially impairing the efficacy of FMT and anti-PD-1
inhibitors. We also confirmed that the ratio of these three
bacteria (P. merdae/(B. plebeius +L. salivarius)) has
clinical implications, including survival likelihood. This
nuanced understanding highlights the complex interplay
between beneficial and detrimental bacteria within the gut
microbiota in determining treatment outcomes. By isolating
and characterizing these specific strains, we demonstrated
the direct mechanisms through which these bacteria influence
the efficacy of immunotherapies, offering potential targets
for enhancing the immunogenicity of the tumor
microenvironment. This approach allowed us to move beyond
general associations to specific causal relationships,
providing a more granular view of microbiota-mediated
modulation of cancer treatment. This emphasizes the need for
targeted manipulation of the gut microbiome to optimize
therapeutic outcomes in cancer patients.
In
the case of R7, although there was a notable reduction in
tumor size and heightened immune response observed following
the second FMT with anti-PD-1 therapy, the later stages
presented significant challenges. About 11 weeks after the
second FMT, the patient experienced progressive epigastric
pain and nausea, diagnosed as immune-related gastritis, a
rare but known immune-related AE (irAE) in ICI-treated
patients.
The exact pathophysiology of irAEs, especially
organ-specific ones, remains unclear but seems to involve
more than just an overactive immune response. To manage
these symptoms, long-term systemic corticosteroid treatment
was initiated around the 12th week after the second FMT,
introducing an unknown variable to the treatment’s long-term
impacts. While planning the third FMT using fecal material
from the same D5, the patient sustained a femoral fracture
from a fall, necessitating surgery. Post-surgery, subsequent
FMTs with the same donor's fecal material did not yield the
anticipated results. The third FMT stabilized the disease,
but the fourth could not halt tumor growth. The diminished
efficacy of these FMTs can be attributed to prolonged
steroid use. Steroids, known for their immunosuppressive
properties,
Moreover, the synergistic interaction of
multiple strains might influence the outcome, suggesting
that other unidentified strains could have affected the
effectiveness of the third and fourth FMTs.
In
conclusion, our findings suggest the potential of FMT to
boost responsiveness to ICIs in advanced solid tumors,
extending beyond melanoma. By comparing responses among
patients, we identifiedP. merdaeImmunoactis
as an effective strain, contrasting with the ineffectiveB. plebeiusandL. salivarius. Our
analysis of the IFN-γ signaling pathway elucidated the
mechanisms of immune modulation by these strains. This study
not only demonstrates the broad applicability of FMT in
enhancing immunotherapy efficacy but also opens new paths
for targeted microbial interventions in cancer treatment.
STAR★Methods
Key
resources table
REAGENT
or RESOURCE
SOURCE
IDENTIFIER
Antibodies
InVivoMAb
rat IgG2a isotype control, anti-trinitrophenol
(clone 2A3)
Further
information and requests for resources and reagents
should be directed to and will be fulfilled by the lead
contact, Hansoo Park (hspa...@gist.ac.kr).
Materials
availability
P. merdaeImmunoactis,
identified in this study is available from Korean
Collection for Type Cultures (KCTC). Details inkey
resources table.
Data
and code availability
Clinical
metadata for matched sequencing files of this study are
available for academic use under confirmation oflead
contact.
Any
of computational code was not manually changed. All
original code is available from each development site
listed inkey
resources table. Details of setting are described
inmethod
details.
Any
additional information and re-analysis is available
under confirmation oflead
contactupon request.
Experimental
Model and Subject Detail
Ethical
approval and consent information
This
study was conducted in accordance with the Declaration
of Helsinki and was approved by the IRB for Asan Medical
Center and GIST (2018-0608 and 20210416-HB-60-07-04,
respectively). All participants provided written
informed consent for the study, including blood, stool,
and tumor tissue biopsy.
Clinical
study design and study population
This
study was conducted at Asan Medical Center and
registered inclinicaltrials.gov;
the registration number is NCT04264975.
Fecal
microbiota transplantation study
This
was a prospective, single-center study of concurrent
fecal microbiota transplantation (FMT) alongside an
anti-programmed death-1/programmed death-ligand 1
(PD-1/PD-L1) inhibitor. The primary objective was to
investigate whether the combination of FMT derived
from good responders to anti-PD-(L)1 inhibitors,
together with an anti-PD-(L)1 inhibitor could overcome
resistance to anti-PD-(L)1 inhibitors in patients with
advanced solid cancer refractory to these inhibitors.
Secondary objectives were to evaluate the safety
profile and examine the systemic and intratumor immune
effects of combining FMT with an anti-PD-(L)1
inhibitor, and identify the potential causative
bacteria that contribute to the efficacy of FMT via
analysis of the gut microbiota composition.
The
key eligibility criteria for recipients were
histologically confirmed solid tumors, except for
hepatocellular carcinoma (HCC), for which a clinically
confirmed diagnosis as per the American Association
for the Study of Liver Diseases (AASLD) was allowed
; confirmed disease progression during anti-PD-(L)1
based therapy (as monotherapy or combination) for
unresectable or metastatic solid tumors; age ≥19
years; Eastern Cooperative Oncology Group (ECOG)
performance status 0‒2; at least one measurable
lesion(s) per the Response Evaluation Criteria in
Solid Tumors (RECIST) v1.1
; no history of other active malignancy within the
previous 3 years before study entry (with the
exception of curatively treated non-melanoma skin
cancer, superficial bladder cancer, or carcinoma
in situ of prostate, cervix, breast, or stomach); no
active infection; no history of active autoimmune
disease requiring systemic treatment except for type 1
diabetes mellitus or autoimmune thyroid disease; no
concurrent condition requiring immunosuppressants; and
no contraindications to colonoscopy.
The
key eligibility criteria for donors were
histologically confirmed solid tumors, except for HCC,
for which clinically confirmed diagnosis as per the
AASLD was allowed
; ongoing durable complete or partial response per
RECIST v1.1 for at least 6 months with anti-PD-(L)1
monotherapy for unresectable or metastatic solid
tumors; age ≥19 years; no concurrent transmissible
diseases; no history or risk behaviors for infectious
disease such as Human Immunodeficiency Virus or other
viral infection, recent travel within 6 months to
countries with endemic diarrheal diseases or high risk
of traveler’s diarrhea; no history of chronic
gastrointestinal disease including inflammatory bowel
disease; and no recent ingestion of allergen with
known recipient allergy. Only donors who passed
serological and stool screening tests were considered
suitable to donate stool samples for FMT implant.
Biomarker
study for validation
This
study was a single-center, prospective,
non-interventional study to investigate the
association between the gut microbiome and the
clinical benefits of anti-PD-(L)1 inhibitors in
patients with unresectable or metastatic solid
cancers. The key eligibility criteria were
histologically confirmed solid tumors, except for HCC,
for which a clinically confirmed diagnosis as per the
AASLD was allowed
; unresectable or metastatic solid tumors treated with
anti-PD-(L)1 monotherapy; age ≥19 years; ECOG
performance status 0‒2; at least one measurable or
evaluable lesion(s) per RECIST v1.1
; no history of other active malignancy within the
previous 3 years before study entry (with the
exception of curatively treated non-melanoma skin
cancer, superficial bladder cancer, or carcinoma
in situ of prostate, cervix, breast, or stomach); no
active infection; no history of active autoimmune
disease requiring systemic treatment except for type 1
diabetes mellitus or autoimmune thyroid disease; and
no concurrent condition requiring immunosuppressants.
Patient
stool samples and peripheral blood samples were
collected at baseline before treatment with
anti-PD-(L)1 inhibitors and at every tumor response
evaluation, which was performed using CT scans every
6–8 weeks according to RECIST v1.1.
Study
treatment in the FMT trial
Preparation
of FMT implant
Feces
of 200-500 mg were collected from a donor immediately
after defecation in a fecal container provided by the
study personnel. The collected feces were transported
to the laboratory as soon as possible within 4 hours
after defecation. For each stool collected, screening
stool tests including stool ova and parasites, stool
culture forSalmonella,Shigella,Campylobacter,Yersinia,
Enterohemorrhagic Escherichia coli,Clostridium
difficile, and antibiotics resistant bacteria
including vancomycin-resistant enterococci
carbapenem-resistant Enterobacteriaceae, and
extended-spectrum beta-lactamases-producing
Enterobacteriaceae were also performed.
The
feces from the donor used for FMT were processed as
soon as possible within 6 hours of defecation. They
were diluted with a sterile solution (normal saline
0.9%) and then homogenized manually using a pipette or
in a blender with low speed for at least 30 seconds.
The fecal suspension was filtered through a clean
metal coffee strainer to remove food debris. The
filtrate was then centrifuged for 15 minutes at
6000 × gand
suspended with sterile solution (0.9% normal
saline/glycerin in a 90/10 v/v ratio) in one third of
the initial volume. The filtrate was packaged into
sterile vials (50 mL in each vial) and stored at
-80°C. On the day of FMT, the frozen suspensions were
thawed using a 37°C warm water bath starting from
about 2 hours before the FMT procedure.
Treatment
with FMT plus anti-PD-(L)1 inhibitor
Recipients
underwent an initial native microbiota depletion phase
where they received oral antibiotics (Augmentin®:
amoxicillin clavulanate; 625 mg po tid) for 5 days
(D-5 to D-1) before FMT treatment (D0). The last dose
of Augmentin was completed at least 12 hours (± 2
hours) prior to the administration of the FMT implant.
FMT was performed via colonoscopy, and the
gastrointestinal tract was prepared by bowel lavage
using 2 L of polyethylene glycol electrolyte solution
before the procedure according to the institute’s
protocol. Thawed fecal suspensions were administered
through colonoscopy, and patients were asked to retain
the material as long as possible after the procedure.
After
the administration of FMT, recipients received an
anti-PD-1 inhibitor at the standard dose on the day
after the FMT, which was continued until unacceptable
toxicity or disease progression. Subsequent FMT
sessions from the same or different donors were
permitted, at the investigator's discretion, with a
time interval of 2 to 4 weeks prior to the initial
response assessment or in cases where clinical
response was not evident. Response assessment was
carried out using computed tomography (CT) scans
conducted every 6 to 8 weeks, following RECIST v1.1.
Preparation
of bacteria
The
following bacterial strains were obtained from the
Korean Collection for Type Cultures (KCTC):Bacteroides plebeius(KCTC
5793) andLactobacillus
salivarius(KCTC 43133).Prevotella merdaeImmunoactis
was isolated as described below. All bacterial species
were cultured at 37°C under anaerobic conditions, which
were created in an anaerobic jar (Oxoid) with an
anaeropack (#A-06, MGC).B. plebeiuswas
cultured in autoclaved reinforced clostridial medium
(RCM) broth (#218081, BD Biosciences).L. salivariuswas
cultured in autoclaved De Man, Rogosa and Sharpe (MRS)
broth (#288130, BD Biosciences).P. merdaeImmunoactis
was cultured in autoclaved brain heart infusion (BHI)
broth (#237500, BD Biosciences) supplemented with 0.5%
yeast extract (#212750, BD Bioscience), 0.05% L-Cysteine
hydrochloride monohydrate (#C7880, Merck), 0.005% hemin
(#H9039, Sigma-Aldrich), and 0.0001% vitamin K1 (#95271,
Sigma-Aldrich). To eliminate the effects of culture
media on T cells, we prepared conditioned bacterial
supernatant; Overnight-cultured bacterial cultures
(1 mL) were inoculated into 9 mL of Roswell Park
Memorial Institute (RPMI) 1640 medium (#10-040-CV,
Corning) and cultured for 24 h. These cultures were then
centrifuged at 6,000 rpm for 5 min, and the supernatants
were filtered using a 0.2 μm syringe filter (#17823-K,
Sartorius). One milliliter of each supernatant was
stored at -80 °C until use. TheP. merdaeImmunoactis
culture was centrifuged at 6,000 rpm for 5 min, the
supernatant was discarded, and the pellet was
resuspended in PBS. The optical density of the
suspension at 600 nm (OD600)
was adjusted to OD600 =
6.
Mammalian
Cell culture
MC38
(mouse colon cancer cell line) and 4T1 (mouse breast
cancer cell line) were cultured in DMEM (#11965-118,
Thermo Fisher Scientific) supplemented with 10% heat-inactivated
FBS (#92916754, MP Biomedicals) and 1%
penicillin-streptomycin (Thermo Fisher Scientific) at
37°C in a 5% CO2atmosphere.
Cells were sub-cultured when they reached 70–80%
confluence.
Mice
experiments
All
animal experiments were conducted following the approved
protocols of the Institutional Animal Care and Use
Committee of GIST (GIST-2021-096 and GIST-2024-011). The
study followed the approved policies for maintaining and
handling the animals. For the syngeneic tumor model,
female C57B6/N mice (Orientbio) seven weeks-old were
subcutaneously injected with 2 × 105MC38
mouse colon cancer cells on day 0, and female Balb/c
mice (Orientbio) seven weeks-old were subcutaneously
injected with 5 × 1054T1
mouse breast cancer cells on day 0. Tumor size was
measured three times every week until the endpoint, and
tumor volume was calculated using the formula:
length × width2 × 0.5.
Each bacterial strain, three-bacteria mixture or PBS
(#10010-049, Thermo Fisher Scientific) was orally
administered daily, starting 14 days before tumor
inoculation and continuing for three weeks after tumor
inoculation. Each strain was prepared to be ingested in
an amount corresponding to an optical density value of
10 (OD600 =
10) per mice. Three-bacteria mixture was composed with
the ratio ofP. merdaeImmunoactis:B. plebeius:L. salivarius =
10 : 1 : 1 based on OD600value.
To investigate the effects of combination therapy,
2 mg/kg IgG isotype (clone 2A3, #BE0089, BioXCell) and
2 mg/kg anti-PD-1 monoclonal antibody (mAb) (clone
RMP1-14, #BE0146, BioXCell) were intraperitoneally
injected on days 7, 9, 11, 14, 16, and 18 after tumor
inoculation. PBS and IgG isotypes were used as controls
for bacterial administration and anti-PD-1 mAb
injections, respectively.
Method
Details
Multiplex
immunofluorescence staining
Multiplex
IHC staining, scanning, and analysis were performed by
PrismCDX Co., Ltd. (Gyeonggi-do, Republic of Korea). For
the staining, 4-μm thick sections were cut from
formalin-fixed, paraffin-embedded (FFPE) blocks. Slides
were heated for at least 1 hour in a dry oven at 60°C,
followed by multiplex immunofluorescence staining using
a Leica Bond Rx™ Automated Stainer (Leica Biosystems,
Newcastle, UK). The list of antibodies and fluorophores
used in the staining process is summarized inTable S8.
Briefly, slides were dewaxed using Leica Bond Dewax
solution (#AR9222, Leica Biosystems), followed by
antigen retrieval using Bond Epitope Retrieval 2
(#AR9640, Leica Biosystems) for 30 min. The staining was
proceeded in sequential rounds of blocking with antibody
diluent/block (ARD1001EA, Akoya Biosciences,
Marlborough, MA, USA), followed by primary antibody
incubation for 30 min, and incubation with OPAL polymer
HRP Ms+Rb (#ARH1001EA, Akoya Biosciences) for 10 min.
The antigen was visualized using tyramide signal
amplification (Akoya Biosciences) for 10 min, after
which the slide was treated with Bond Epitope Retrieval
1 (#AR9961, Leica Biosystems) for 20 min to remove bound
antibodies before the next step. The process from
blocking to antigen retrieval was repeated for each
antibody stain. Nuclei were counterstained with DAPI
(#62248, Thermo Fisher Scientific, Waltham, MA, USA)
after the final antigen retrieval step. The slides were
overlaid with coverslips and ProLong Gold antifade
reagent (#P36935, Invitrogen, Waltham, MA, USA).
Multispectral
imaging and analysis
Multiplex-stained
slides were scanned at 20× magnification using the
Vectra Polaris Automated Quantitative Pathology Imaging
System (Akoya Biosciences). Representative images for
training were selected in a Phenochart (Akoya
Biosciences) and an algorithm was created using the
inForm Image Analysis software (Akoya Biosciences).
Multispectral images were unmixed using the spectral
library in inForm software, and tumor tissues were
segmented based on the presence or absence of
cytokeratin (CK) expression. Each cell was segmented
based on DAPI staining, and phenotyping was performed
based on the expression compartment and intensity of
each marker. After designating the region of interest
(ROI) to be analyzed on the tissue slide, the algorithm
created in this way was applied in a batch-running
manner. The exported data were consolidated and analyzed
in the R software using the phenoptr (Akoya Biosciences)
and phenoptrReport packages (Akoya Biosciences).
Detection
of multiple cytokines (Simoa)
Simoa
HD-1 Assay
Plasma
levels of IL-2 and IL-15 were measured using the Simoa
reagent kit (Quanterix, MA, USA) on a Simoa HD-1
Analyzer (Quanterix, MA, USA) at PrismCDX Co., Ltd. by
an investigator who was blinded to clinical
information. Plasma samples were diluted 1:4 according
to Quanterix guidelines. All HD-1 consumables,
including wash buffers, 96-well plates, discs,
cuvettes, and sealing oil, were purchased from
Quanterix. The assay was performed on the fully
automated Simoa HD-1 using a 3-step protocol. In the
3-step assay, the target antibody coated with
paramagnetic beads was incubated with the sample so
that the target molecule in the sample was bound by
the bead-coated antibody. After washing, they were
mixed with a biotinylated detector antibody and
incubated to bind the detector antibody to the
captured target. After washing, the complex was mixed
with streptavidin-β-galactosidase (SBG) to allow SBG
to bind to the biotinylated detector antibody and
generate an enzyme label on the captured target. After
the final wash, the beads were resuspended in RGP
substrate solution and transferred to a Simoa disc.
Captured and labeled target antigens were measured
while β-galactosidase hydrolyzed the RGP substrate.
The average number of enzymes per bead (AEB) was
calculated from the fraction of active wells that
correlated with the analyte concentration. For
assessing plasma IL-2 and IL-15, all samples were
analyzed in duplicates, and the mean of the replicates
for each sample was calculated.
Simoa
SP-X Assay
Ten
plasma markers were measured using Simoa CorPlex Human
Cytokine 10-plex Panel 1 kit (Quanterix, MA, USA). The
ten targets were interferon gamma (IFN-γ), IL-1β,
IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-22, and
tumor necrosis factor alpha (TNFα). Each well of a
96-well microplate was pre-spotted with
protein-specific antibodies to capture the target
protein in the sample. After washing unbound proteins,
biotinylated detector antibodies were added to bind to
the secondary site of the target protein and washed to
remove the excess detector antibody. Finally,
streptavidin–horseradish peroxidase (SA–HRP) was added
to allow HRP to react with the substrate and generate
a luminescent signal detected by the SP-X Imaging
System (Quanterix, MA, USA). The amount of protein
present was quantified based on the signal intensity.
Mass
flow cytometry (CyTOF)
Peripheral
blood mononuclear cells (PBMCs) were stained using the
Maxpar Direct Immune Profiling Assay kit (Fluidigm)
according to the manufacturer’s instructions. PBMC
aliquots were prepared, and the buffer was exchanged
with Maxpar Cell Staining Buffer (Fluidigm). Human
TruStain FcX (BioLegend, San Diego, CA, USA) was added
to each tube to block Fc-receptor, followed by
incubation for 10 minutes at 19-24°C. PBMCs were then
directly transferred to a 5 mL tube containing a dry
antibody pellet (Fluidigm) and incubated for 30 min at
room temperature. After washing, PBMCs were fixed in
1.6% formalin for 10 min at room temperature. Finally,
PBMCs were incubated with Cell-ID Intercalator-Ir
(Fluidigm) in Maxpar Fix and Perm Buffer (Fluidigm) for
approximately 48 h at 4°C. Prior to acquisition, PBMCs
were washed twice with Maxpar Cell Staining Buffer
(Fluidigm) and filtered through a 40 μm cell strainer
before being acquired on a Helios mass cytometer
(Fluidigm). Mass cytometry data files were analyzed
using FCS Express 7 Flow software (CAT #402001, De Novo
Software, Pasadena, CA, USA). The clone and mass of each
antibody are described inTable S9.
Fecal
DNA extraction and sequencing
All
fecal samples were collected and stored in collection
tubes (#R1101, Zymo Research) until DNA extraction. DNA
was extracted from fecal samples using FastDNA® SPIN Kit
for Soil (MP Biomedicals) according to the
manufacturer’s instructions. The purity and quantity of
DNA were estimated using NanoDrop One Spectrophotometer
(Thermo Fisher Scientific). The bacterial 16S rRNA V3–V4
region was amplified using the Illumina’s 16S
Metagenomic Sequencing Library Preparation guide
(Illumina), using primers with adapter overhang
sequences added.
Forward primer: 5’-
TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG -3’,
Reverse primer: 5’-
GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC
-3’. The PCR reaction mix included 2 μL of genomic DNA,
0.5 μL of each primer, 12.5 μL of 2× KAPA HiFi HotStart
ReadyMix (Kapa Biosystems), and 9.5 μL of distilled
water, for a total volume of 25 μL. The PCR conditions
used were as follows: 95°C for 3 min for pre-denaturing
DNA; 25 cycles at 95°C for 30 s for denaturing DNA; 50°C
for 30 s for annealing; 72°C for 30 s for extension; and
72°C for 5 min for final extension. The PCR products
were purified using AMPure XP Beads (Beckman Coulter).
Dual index attachment and Illumina sequencing adapters
were performed using PCR products (5 μL), Illumina
Nextera XT Index Primer 1 (5 μL, N7xx), Nextera XT Index
Primer 2 (5 μL, S5xx), 2× KAPA HiFi HotStart Ready Mix
(25 μL), and nuclease-free water (10 μL) using the
following protocol: 95°C for 3 min; 8 cycles of 95°C for
30 s, 55°C for 30 s, and 72°C for 30 s; and 72°C for
5 min. To remove DNA debris and obtain exact DNA
fragments from PCR product, the PCR products were
cleaned using AMPure XP beads (Beckman), and then
quality control of the 16S metagenomic libraries was
performed using Agilent Technologies 2100 Bioanalyzer
(Agilent). Libraries were standardized and pooled for
sequencing on MiSeq platform (Illumina), according to
the standard Illumina sequencing protocol. The DNA from
the fecal samples was extracted using TruSeq Nano DNA
sample Prep Kit (Illumina, CA, USA) for shotgun
metagenomic sequencing, following the manufacturer’s
instructions by Theragene Bio (Gyeonggi-do, Republic of
Korea). The sequencing was performed by Theragene Bio
(Gyeonggi-do, Republic of Korea) on an Illumina platform
(version 1.9), and the libraries were strandardized and
according to the standard Illumina sequencing protocol.
Metagenomic
analysis
Illumina
adapter sequences of paired-end reads were removed using
Cutadapt version 4.1.
After trimming, the reads were processed
using QIIME2 version 2022.8. Briefly, the reads were
first assigned to each sample according to a unique
index, and pairs of reads from the original DNA
fragments were merged using an import tool in QIIME2.
Quality control and trimming were
performed to yield sequences with lengths of 230 bp and
210 bp for the forward and reverse reads, respectively.
The DADA2 software package
wrapped in QIIME2 was used to remove
low-quality bases from the reads. A consensus method
implemented in DADA2 was used to remove chimeras from
the FASTQ files. Alpha and beta diversities were
analyzed using core-metrics-phylogenetic analysis in the
QIIME2 diversity plugin. Alpha and beta diversities were
calculated using alpha- and beta-group significance in
the QIIME2 diversity plugin, respectively. Alpha
diversity was calculated using observed species
(observed features), and beta diversity was compared by
principal coordinate analysis using Bray–Curtis
distances (BCD). To assess FMT engraftment, we
calculated the "Engraftment Distance" using the formula
BCDtoDonor +
(1 - BCDtoBase).
The significance of the similarity between the groups
was evaluated using permutational multivariate analysis
of variance (PERMANOVA) with 999 permutations. Taxonomic
annotation was performed by mapping the training
reference set with the primers (forward,
5′-CCTACGGGNGGCWGCAG-3′; reverse,
5′-GACTACHVGGGTATCTAATCC-3′) and extracting the V3–V4
region using the NCBI reference database (https://www.ncbi.nlm.nih.gov/refseq/)
with manually added microbiota by using RESCRIPt.
and R Studio (2023.09.1+494). To address
the microbial composition analysis between the groups R6
and R7, we employed the Analysis of Composition of
Microbiomes (ANCOM).
Additionally, 16S rRNA and marker
gene-based shotgun metagenomic analyses were performed.
Briefly, quality checks of the sequenced reads were
conducted using kneaddata (https://github.com/biobakery/kneaddata)
with the hg37dec database (version 0.1.4) and --trimmomatic
Trimmomatic-0.39option.
Fecal
samples were suspended in anaerobic PBS and then seeded
on brain heart infusion (#237500, BD) agar plats
supplemented with 0.5 % yeast extract (#212750, BD),
0.05 % L-cysteine HCl (#C7880, Sigma-Aldrich), 0.0005 %
hemin (#51280, Sigma-Aldrich), 0.0001 % vitamin K1
(#95271, Sigma-Aldrich), 0.01 % kanamycin (#HKA01,
Rd-tech), and 5 % defibrinated sheep blood (#S1876,
KisanBio). The plates were incubated at 37°C under
anaerobic conditions using a GasPak 100 system (#260626,
BD). Colonies were randomly picked from the plates and
tested with PCR forP. merdaeImmunoactis
species-specific sequences with the primer set,
5’-GCTCAACCTGGGCATTGCA-3’ and
5’-CATGTTTTAGGGATTCGAGCG-3’′. DNA of isolatedP. merdaeImmunoactis
was prepared using TruSeq Nano DNA sample Prep Kit
(Illumina, CA, USA) for whole genome sequencing by
Theragene Bio (Gyeonggi-do, Republic of Korea).
Libraries were standardized and pooled for sequencing on
the Illumina platform (version 1.9), according to the
standardized Illumina sequencing protocol by Theragene
Bio (Gyeonggi-do, Republic of Korea). Sequenced reads
were quality-checked using BBDuk plug-in BBMap with--trimq=30 and --maq=30options,
by using tetra correlation search with
default option. Calculation of the genomic similarity
betweenP. merdaeImmunoactis
andP. merdaesp.
Marseille-P4119
Human
T cells were isolated from blood samples collected from
healthy donors in their 20s with informed consent. This
study was approved by the Institutional Review Board
(20210416-HB-60-07-04) of the Gwangju Institute of
Science and Technology (GIST), South Korea. PBMCs were
isolated from blood samples of 40–50 mL volume using
Ficoll® Paque Plus (#GE17-1440-02, Merck) through the
process of density gradient centrifugation. The CD4+and
CD8+T cells
were isolated separately using CD4+T
Cell Isolation Kit (#130-096-533, Miltenyi Biotec) and
CD8+T
Cell Isolation Kit (#130-096-495, Miltenyi Biotec),
respectively. The isolated T cells were resuspended in
RPMI 1640 medium (#10-040-CV, Corning) supplemented with
10% fetal bovine serum (FBS; #92916754, MP Biomedicals),
1% penicillin-streptomycin (#15140-122, Thermo Fisher
Scientific), 1× NEAA (#11140-050, Thermo Fisher
Scientific), 200 mM HEPES (#15630-080, Thermo Fisher
Scientific), 2 mM sodium pyruvate, 2 mM L-glutamine
(#25030-081, Thermo Fisher Scientific), and 1×
2-mercaptoethanol (#21985-023, Thermo Fisher
Scientific). Additionally, the T cells were stained with
carboxyfluorescein diacetate succinimidyl ester (CFSE;
#C34554, Thermo Fisher Scientific) for 30 min at 37°C in
a 5% CO2atmosphere.
96-well non-treated plates (#32096, SPL) were pre-coated
with 0.2 μg anti-CD3e antibody (#100340, BioLegend) in
each well and incubated overnight at 4°C. CFSE-stained
T cells were then added to the pre-coated plates at a
concentration of 2×105cells/well
and incubated either alone or with 10% bacterial
supernatant or three-bacteria mixture for 72 h at 37°C
in a 5% CO2atmosphere.
Three-bacteria mixture was composed with the ratio ofP. merdaeImmunoactis:B. plebeius:L. salivarius =
10 : 1 : 1 based on OD600value.
After 72 h, T cells were harvested and washed with FACS
buffer (PBS + 10% FBS + 2 mM EDTA), and their
proliferation was analyzed using a BD FACS CANTO II (BD
Bioscience). The supernatants from the T cells were also
collected and stored at -80°C until use for ELISA. The
IFN-γ levels in the supernatants were measured using a
human IFN-γ ELISA kit following the manufacturer’s
protocol (#88-7316-88, Thermo Fisher Scientific). FlowJo
software (Treestar) was used to analyze the data.
OT-1
T cell Cytotoxicity Assay
For
the cytotoxicity assay, ovalbumin-specific CD8+T cells
(OT-1 T cells) were used. Briefly, the spleen was
isolated from OT-1 mice (#003831,
C57BL/6-Tg[TcraTcrb]1100Mjb/J, The Jackson Laboratory)
and ground on a 70 μm cell strainer (#352350, Falcon)
with 3 mL PBS. Cells were resuspended in 10 mL of 1× RBC
lysis buffer (#420301, BioLegend) and incubated for
10 min at room temperature. OT-1 T cells were isolated
using a CD8+T
Cell Isolation Kit (#130-096-495, Miltenyi Biotec).
Isolated OT-1 T cells were resuspended in the T cell
medium. A 60 mm non-treated plate (#32096, SPL) was
pre-coated with 5 mg anti-CD3ϵ antibody (#100340,
BioLegend) and 0.5 mg anti-CD28 antibody (#102116,
BioLegend), and incubated overnight at 4°C. Further,
OT-1 T cells were distributed into pre-coated 60 mm
plate at 1–2×106cells/well
with 10% bacterial supernatant and incubated for 48 h at
37°C in a 5% CO2atmosphere.
After 48 h of incubation, OT-1 T cells were harvested
for RNA isolation, or co-cultured with Ova-treated MC38
cells at a 5:1 ratio in T cell media for 6 h. The cells
were then harvested and washed with FACS buffer.
Finally, the cells were stained using Annexin V–PI
staining kit (#ALX-850-020-K101, Enzo Life Science) and
analyzed using flow cytometry. MC38 cells were treated
using ovalbumin (323-339) (#O1641, Merck) and incubated
overnight at 37°C in a 5% CO2atmosphere
before co-culturing with OT-1 T cells.
Comparison
of relative gene expression (RT-PCR)
Activated
OT-1 cells were harvested, and RNA was extracted using
TRIzol reagent (#15596026, Invitrogen) following the
manufacturer's protocol. RNA quality control was
conducted by assessing the A260/A280 and A260/A230
ratios. Subsequently, cDNA synthesis was performed using
the TOPscript™ RT DryMIX kit (#RT200, Enzynomics),
followed by RT-PCR using the TOPreal™ qPCR 2X PreMIX kit
(#RT501M, Enzynomics) and a StepOnePlus (#4376600,
Applied Biosystems). Gene expression levels were
normalized to 1 in the "Ova act" group, and relative
expression levels were compared across other
experimental groups. Primer sequences and PCR conditions
are provided inTable S10.
Immune
cell profiling
The
lymph nodes and tumor tissues were isolated from mice
21 days after inoculation with MC38 cells. The immune
cells analyzed were T cells, NK cells, dendritic cells
(DC), Tregs, and macrophages. To isolate the lymphocytes
from the lymph nodes, they were ground on a 70 μm cell
strainer (#352350, Falcon) with 1 mL PBS (#10010-049,
Thermo Fisher Scientific), and the cells were washed
with PBS and counted. For the tumor tissue, it was cut
into small pieces and incubated in 5 mL dissociation
media for 40 min at 37°C. The dissociation media was
made using T cell media supplemented with 2.5 mg/mL
collagenase type I (#17100-017, Thermo Fisher
Scientific), 1.5 mg/mL collagenase type II (#17101-015,
Thermo Fisher Scientific), 1 mg/mL collagenase type IV
(#17104-019, Thermo Fisher Scientific), 50 μg/mL DNase
type I (#10104159001, Merck), and 0.25 mg/mL
hyaluronidase type IV-S (#H3884, Merck). After the
dissociation process, the samples were centrifuged, and
the supernatant was discarded. The cells were then
resuspended in 5 mL of 1× RBC lysis buffer (#420301,
BioLegend) and incubated for 10 min at room temperature.
The samples were filtered through a 70 μm cell strainer
and the number of cells was counted. To block the
Fc-receptor, the samples were incubated with anti-mouse
CD16/CD32 (BioLegend, #101330) for 10 min at room
temperature. Antibody mixtures were then added to the
samples for surface marker staining and incubated for
1 h at 4°C. In the case of Tregs and macrophages, the
samples were fixed and permeabilized using a
fixation/permeabilization buffer set (BioLegend,
#424401), and an intra Ab mixture was added for
intra-marker staining and incubated for 1 h at 4°C. The
stained cells were then acquired using CANTO II (BD
Biosciences) and BD-FACS Diva software v.8.0.2 (BD
Bioscience) and analyzed using FlowJo software (v.10,
TreeStar). The staining markers used for each cell type
were as follows: T cells were stained with CD45
(#103116, BioLegend), CD3 (#100218, BioLegend), CD4
(#100422, BioLegend), CD8a (#563068, BD Bioscience),
CD44 (#563970, BD Bioscience), CD62L (#104412,
BioLegend), and PD-1 (#109104, BioLegend); NK cells were
stained with CD45 (#103116, BioLegend), CD3 (#100218,
BioLegend), NK1.1 (#550627, BD Bioscience), and NKG2A
(#11-5896-82, eBioscience); DC were stained with CD11c
(#562782, BD Bioscience), CD11b (#566416, BD
Bioscience), and B220 (#103224, Biolegend); Tregs were
stained with CD45 (#103116, BioLegend), CD3 (#100218,
BioLegend), CD4 (#100422, BioLegend), CD8a (#563068, BD
Bioscience), CD25 (#101910, BioLegend), Foxp3 (#562996,
BD Bioscience), and IFN-γ (#163503, BioLegend); and
macrophages were stained with CD45 (#103116, BioLegend),
CD11b (#566416, BD Biosciences), F4/80 (#565411,
BioLegend), and TNF-α (#506308, BioLegend).
Quantification
and Statistical Analysis
Statistical
calculations were performed using Prism (Version 9.2.0,
GraphPad). Differences between two variables were analyzed
using the two-sided unpaired t-test or
Wilcoxon-Mann-Whitney test. Multiple variables were
assessed by one-way or two-way ANOVA with Tukey's multiple
comparison test. Correlation analysis was performed using
R version 4.1.0. Statistical significance was considered
if the p-value was lower than 0.05.
Acknowledgments
This
study was supported by grants from the Asan Institute for
Life Sciences, Asan Medical Center (2018IT0608 to S.R.P.),
the National Cancer Centre, Korea (NCC-1911267 to H.P. and
S.R.P.), the GIST Research Institute (GRI), funded by the
GIST (in 2020 and 2023 to H.P.), and the Bio and Medical
Technology Development Program (2022R1A2C2008976 and
RS-2023-00228315) from the Ministry of Science and ICT,
Korean Government.
Author
contributions
Conceptualization,
H.P. and S.R.P.; resources, S.-Y.K., E.-J.D., J.S.S.,
S.H.P., S.W.H., M.-N. Kim, S.-H.K., and SR.P.;
investigation, G.K., Yunjae Kim, Sujeong Kim, B.C., Seungil
Kim, and M.-N. Kweon; data curation, G.K., Yunjae Kim,
Seungil Kim, D.-J.B., and SR.P.; visualization, Yunjae Kim,
G.K., Sujeong Kim, and D.-J.B.; writing – original draft,
G.K., Yunjae Kim, Sujeong Kim, and S.R.P.; writing –
review & editing, Yunjae Kim, Sujeong. Kim, G.K., B.C.,
Yeongmin Kim, K.M., and S.R.P.; project administration,
Yunjae Kim, G.K., Sujeong Kim, H.P., and S.R.P.;
supervision, M.D.A., C.L., H.P., and SR.P.; funding
acquisition, H.P. and SR.P.
