M cells of mouse and human Peyer''s patches mediate the lymphatic absorption of an Astragalus hyperbranched heteroglycan

The gut cell wall is considered an impenetrable barrier to orally administrated polysaccharides. We recently reported a selective lymphatic route for Radix Astragali polysaccharide RAP to enter Peyer''s patches (PPs) to trigger immune responses. However, how RAP enters PPs is unclear.

Available online 4 August 2022

0144-8617/© 2022 The Authors Published by Elsevier Ltd This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/)

M cells of mouse and human Peyer's patches mediate the lymphatic

absorption of an Astragalus hyperbranched heteroglycan

Quanwei Zhanga, Shuang Haob, Lifeng Lia, Man Liua, Chuying Huoa, Wanrong Baoa,

Huiyuan Chenga, Hauyee Funga, Tinlong Wonga, Wenjie Wua, Pingchung Leungc,

Shunchun Wangd, Ting Lie, Ge Zhanga, Min Lia, Zhongzhen Zhaoa, Wei Jiaa, Zhaoxiang Biana,

Timothy Mitchisonf, Jingchao Zhangb,*, Aiping Lyua,*, Quanbin Hana,*

aSchool of Chinese Medicine, Hong Kong Baptist University, Hong Kong 999077, Hong Kong, China

bThe First Affiliated Hospital, Zhengzhou University, Zhengzhou 450000, China

cState Key Laboratory of Research on Bioactivities and Clinical Applications of Medicinal Plants, The Chinese University of Hong Kong, Hong Kong 999077, Hong Kong,

China

dInstitute of Chinese Materia Medica, Shanghai University of Traditional Chinese Medicine, Shanghai 201203, China

eState Key Laboratory for Quality Research in Chinese Medicine, Macau University of Science and Technology, 999078, Macau

fLaboratory of Systems Pharmacology and Department of Systems Biology, Harvard Medical School, Boston 02115, United States

A R T I C L E I N F O

Keywords:

Radix Astragali polysaccharide

Microfold cell

Lymphatic absorption

Transcytosis

Immune regulation

A B S T R A C T The gut cell wall is considered an impenetrable barrier to orally administrated polysaccharides We recently reported a selective lymphatic route for Radix Astragali polysaccharide RAP to enter Peyer's patches (PPs) to trigger immune responses However, how RAP enters PPs is unclear Herein, we screened the intestinal epithelial cells of mice and found that the follicle-associated epithelium cells were specifically bound with FITC-RAP

Further studies in vitro and in vivo revealed that RAP was efficiently transported by microfold (M) cells We

also confirmed that M cell-transported RAP directly contacted dendritic cells More importantly, for the first time, we verified this interesting M cell-mediated transcytosis of RAP in the human distal ileum Mechanistically,

we identified M cells to be the transporter cells that independently deliver RAP into the lymphatic system to trigger immune responses This interesting transcytosis mechanism might apply to many other immunomodu-latory polysaccharides orally dosed to human body

1 Introduction

The gut wall barrier to macromolecules remains an unsolved

chal-lenge for developing orally-delivered macromolecular therapeutics

(Moroz et al., 2016) Recently proposed solutions have focused on

highly engineered systems such as self-orienting microinjectors

(Abramson et al., 2019), but such complexity creates challenges for

manufacture and regulation (Hussain, 2016) In puzzling over this, we

noticed that polysaccharides from traditional herbal medicines quickly

affect the immune system after oral dosing (Jiang et al., 2010;

Sche-petkin & Quinn, 2006; Yu et al., 2018) Our previous study revealed that

the Radix Astragali polysaccharide RAP enters Peyer's patches (PPs) in

the small intestine intact, and subsequently initiates mucosal immune

responses by targeting follicle dendritic cells (FDCs) (Zhang et al.,

2022) In other words, this appears to be a lymphatic route for

polysaccharides crossing the gut wall barrier

Radix Astragali (RAP, Huang qi in Chinese) is one of the most widely used traditional Chinese medicine (TCM) It could promote Yang and replenish Qi, which is considered the primitive application of the herb's immune regulation effects Being first recorded and classified as the top

grade in Shennong's Materia Medica Classic, RA exists various therapeutic

activities, such as antioxidant, anti-cancer, anti-diabetic, and immuno-modulatory effects (Chen et al., 2020; Fu et al., 2014; Liu et al., 2011) Statistics showed that RA was found in about 200 Chinese medicine prescriptions, and water decoctions were the major ways to use RA in clinics (Cheng et al., 2019) Polysaccharides of RA, as one of the most important compounds in its water decoction, play a critical role in im-mune functions of RA (Li et al., 2020; Zheng et al., 2020) A better

un-derstanding of how polysaccharides of RA work in vivo would be helpful

for the development and application of RA-based prescriptions

* Corresponding authors

E-mail addresses: zhangjingchao126@126.com (J Zhang), aipinglu@hkbu.edu.hk (A Lyu), simonhan@hkbu.edu.hk (Q Han)

Contents lists available at ScienceDirect Carbohydrate Polymers

journal homepage: www.elsevier.com/locate/carbpol

https://doi.org/10.1016/j.carbpol.2022.119952

Received 6 June 2022; Received in revised form 16 July 2022; Accepted 1 August 2022

The hyperbranched heteroglycan RAP (1334 kDa) was purified from

Radix Astragali (Yin et al., 2012) Its chemical structure was

charac-terized by monosaccharide composition, partial acid hydrolysis,

methylation analysis, GC–MS, NMR spectra, SEM, and AFM microscopy

It was composed of Rha, Ara, Glc, Gal, and GalA in a molar ratio of

0.03:1.00:0.26:0.37:(0.28) As shown in Supplementary Fig 1 in our

newly reported study (Zhang et al., 2022), its average molecular weight

and purity remain unchanged, and the 1H NMR spectrum, 13C NMR

spectrum, and monosaccharides composition further confirmed its

sta-ble chemistry As a major component, it looks similar to the reported

heteroglycans from the same herb medicine (Zheng et al., 2020)

Although our recent study revealed a selective route for RAP entering

PPs to trigger immune responses quickly (Zhang et al., 2022), it did not

explain how RAP actually crosses the cell wall Transcytosis by

micro-fold (M) cells (Dillon & Lo, 2019) is one possible mechanism These

specialized epithelial cells, located on immune sensors of the small

in-testine PPs, deliver pathogens and food molecules from the gut lumen to

the lymphatic system (Neutra et al., 1996; Okumura & Takeda, 2017)

The M cell-targeted transcytosis has been a promising route for oral

vaccination, especially for bioactive macromolecules (Kim & Jang,

2014; Saraf et al., 2020) Considering that RAP quickly triggers immune

responses in PPs 1 h after oral administration, and M cell-mediated

transport can be completed in 15 min (Sakhon et al., 2015), we

hy-pothesize that M cells are the transporter for RAP to pass through the gut

cell wall to initiate immunoregulation in vivo

In this study, we collected comprehensive evidence in vitro and in

vivo, on both animal and human subjects, that supports this hypothesis

We first screened the intestinal epithelial cells using flow cytometry and

found that RAP was specifically bound to the follicle-associated

epithelial (FAE) cells We further tested RAP's transport on an in vitro

FAE model A control nanoparticle and the specific M-cell marker GP2

were used to validate the specific transport function of M-like cells in

this model Immunofluorescence staining assay, flow cytometry, and

high-performance permeation chromatography were performed to

detect the transport of intact FITC-RAP in the lower layer of transwell

inserts In order to exclude the influence of FITC, we also tested

unla-beled RAP ELISA assay was used to assess the effects of transported RAP

on macrophages The results suggested that M-like cells transported RAP

could reach a significant concentration of 1 μg/mL In the in vivo

ex-periments, the FAE M cells were characterized by two specific cell

markers, GP2 and NKM 16-2-4, in immunofluorescence staining The

RAP transport by M cells was clearly demonstrated in 3D confocal

mi-crographs This interesting transport of RAP by M cells was also verified

in human subjects Taken together, evidence showed that M cells were

the specific transporter cells

2 Materials and methods

2.1 Human subjects, animals, and cells

For the study of human subjects, this study was performed following

the established ethical guidelines It was approved (2019-KY-373) by the

research ethics committee of the School of Medicine, Zhengzhou

Uni-versity, Zhengzhou, China

C57 BL/6 and BALB/c mice were purchased from the Chinese

Uni-versity of Hong Kong Five- to eight-week-old mice were used All animal

experiments followed the Animals Ordinance guidelines, Department of

Health, Hong Kong SAR

Caco-2 cells, Raji B cells, and RAW264.7 cells were obtained from

American Type Culture Collection (ATCC)

2.2 Antibodies and reagents

PE-conjugated rabbit anti-mouse glycoprotein 2 (GP2) antibody

(2F11-C3), rat anti-mouse GP2 antibody (2F11-C3), and mouse anti-

human GP2 antibody (3G7-H9) was from MBL International CD11c

primary antibody, goat anti-rabbit or rat IgG antibody conjugated with Alexa Fluor 647, and goat anti-rat IgG antibody conjugated with Alexa488 was purchased from Abcam Matrigel basement membrane matrix phenol red-free, phenol-sulfuric acid, 4′ ,6-diamidino-2-phenyl-indole DAPI, anti-microfold cell antibody (NKM 16-2-4), and methyl sulfoxide were obtained from Merck Fluoresbrite® Multifluorescent Microspheres (0.20 μm) were purchased from Polysciences DMEM, fetal calf serum (FBS), 4 % paraformaldehyde (PFA), and all ELISA kits were purchased from Thermo Fisher Scientific

2.3 Preparation of polysaccharides and FITC-labeled polysaccharides

RAP was previously prepared from the water extract of the dried

roots of Astragalus membranaceus (Yin et al., 2012) and stored at 4 ◦C RAP is a hyperbranched heteroglycan with a molecular weight of 1334 kDa The quality was stable compared to multiple newly prepared RAP samples This sample displayed high consistency in the HPGPC molec-ular distribution patterns and the HPLC oligo-saccharide profile pro-duced by partial hydrolysis (Zhang et al., 2022) As reported, RAP was labeled with fluorescein isothiocyanate isomer I (FITC) (Li et al., 2019)

2.4 Isolation of intestinal epithelial cells

The small intestine and Peyer's patches (PPs) were isolated from five mice and were washed five times with cold PBS All the samples were stirred in 20 mL of 0.5 mM EDTA/PBS at 37 ◦C with 300 g for 20 min, followed by vigorous shaking for 15 s Samples were washed three times and filtered with 70 μm nylon filter mesh, and then we would get cell suspensions Cell suspensions were prepared as shown in the method of flow cytometry

2.5 In vitro M-like cell model

An in vitro M-like cell model was used to evaluate the transport of

RAP (Beloqui et al., 2017) In brief, Caco-2 cells were cultured in flasks

in 10 % (v/v) fetal calf serum DMEM at 37 ◦C in a 10 % CO2 water- saturated atmosphere Caco-2 cells (6 × 106) grew on Transwell poly-ester inserts (0.3 μm pore size, 12 mm diameter) which were coated with

100 μg/mL Matrigel basement membrane matrix prepared in pure cold DMEM without phenol red Raji B cells (5 × 105) were resuspended in DMEM and added to the basolateral chamber of 7-day-old Caco-2 cell monolayers The co-cultures were maintained for 5 days Monoculture of Caco-2 cells, cultivated as above but without the Raji B cells, was used as control Inserts were used for all the following experiments, including assessment of M cell functionality by nanoparticle transport measure-ment using flow cytometry, M cell demonstration in co-cultured monolayers using an immunofluorescence staining assay, and determi-nation of transported RAP using chromatographic analysis and ELISA assay

2.6 Assessment of M cell transport functionality

The transport function of this in vitro M cell model was validated

using nanoparticles (Beloqui et al., 2017) Fluoresbrite multifluorescent microspheres (green, 200 nm diameter) were used in transport experi-ments The nanoparticles (400 μL, 5 × 109/mL) were added to the upper layer of mono-culture and co-culture inserts and then incubated for 2 h

at 37 ◦C under 5 % CO2 Solutions of the lower inserts were collected and measured using a cytometer FACSAria (BD Biosciences)

2.7 Identification of M-like cells using immunofluorescence staining

Three co-culture and monoculture inserts were used to detect the presence of M-like cells by immunofluorescence staining All inserts were washed three times with PBS and fixed with 4 % paraformaldehyde (PFA) for 15 min and then blocked for 1 h in 5 % of normal goat serum in

PBS Rat anti-human GP2 antibody was used as the primary antibody

and incubated overnight at 4 ◦C After washing three times with PBS, all

inserts were incubated with Alexa Fluor 647-conjugated goat anti-rat

secondary antibody for 1 h in the dark DAPI was used to stain cell

nuclei for 15 min in the dark All inserts were washed three times with

PBS, and the polyester membranes were cut and mounted using an anti-

fade mounting medium Slides were examined with a Leica TCS SP8

confocal laser scanning microscope

2.8 Treatment of FITC-RAP in the M-like cell model

FITC-RAP (2 mg/mL) in 400 μl DMEM was added to the upper layers

of monoculture and co-culture inserts and then incubated for 2 h at 37 ◦C

under 5 % CO2 All inserts were removed, and lower solutions were

collected for chromatographic analysis and flow cytometry analysis The

lower solution collected from the co-culture group treated with

unla-beled RAP was further used to treat RAW264.7 cells in a 96-well plate

for 24 h, and the supernatants were sampled for IL-6 and IL-12 detection

using ELISA kits

2.9 High-performance gel permeation chromatography coupled with

fluorescence detector (HPGPC-FLD) analysis

500 μL lower layer solution/insert were collected and detected using

HPGPC-FLD assay according to our reported method (Li et al., 2019) In

brief, all solutions were centrifuged at 15,000 rpm for 10 min The

separation was performed on a TSK GMPWXL column (300 × 7.8 mm i

d., 10 μm) system operated at 40 ◦C using an Agilent-1100 HPLC system

equipped with FLD Ammonium acetate aqueous solution (20 mM) was

used as a mobile phase at a 0.6 mL/min flow rate The excitation

wavelength and emission wavelength of FLD were 495 and 515 nm,

respectively

2.10 Flow cytometry analysis

200 μL lower layer solution was collected and measured using

FACSAria (BD Biosciences) cytometer To identify the transporter of

FITC-RAP in the intestinal epithelial cells (IECs) in vivo, cell suspension

of IECs was stained with PE-conjugated GP2 for 25 min at RT and

washed two times The stained cells were rinsed twice, resuspended in

PBS, and analyzed by a cytometer FACSAria All data were analyzed

with FlowJo V10 software

2.11 ELISA for quantitative analysis of cytokines

The supernatants of RAW264.7 cells collected from the above

method were centrifuged at 3000g for 10 min According to the

manu-facturer's instructions, cytokines Interleukin 6 (IL-6) and IL-12 were

determined using ELISA kits

2.12 Immunofluorescence staining

Frozen sections of PPs from mice treated with/without FITC-RAP

were washed three times with PBS and blocked with 5 % normal goat

serum in PBS for 1 h Sections were incubated with anti-mouse GP2 and

anti-mouse NKM 16-2-4 antibodies overnight at 4 ◦C PPs sections were

washed three times with PBS and then treated with Alexa Fluor 647

secondary goat anti-rabbit antibody, Alexa Fluor 488-conjugated anti-

rat antibody, and Alexa Fluor 594-coujugated anti-rat antibody for 1 h

at RT, followed by three consecutive PBS washes The cell nuclei were

stained with 1 μg/mL 4′,6-diamidino-2-phenylindole (DAPI) for 15 min

Sections were washed three times with PBS and mounted with an anti-

fade mounting medium Images were captured with a Leica TCS SP8

confocal laser scanning microscope

2.13 Whole-mount staining

For whole-mount staining, PPs were isolated from the small intestine

of mice treated with/without FITC-RAP After being washed three times, the PPs were fixed in 4 % paraformaldehyde overnight at 4 ◦C and then permeabilized with 0.1 % Triton X-100 for 30 min PPs were blocked with 5 % normal goat serum in PBS for 1 h Primary antibodies, including rat anti-mouse glycoprotein 2 (GP2) and rabbit anti-mouse NKM 16-2-4, and 1 μg/mL DAPI were used to incubate PPs for 1 h at

RT, followed by rinsing 3 × 5 min with PBS The second goat anti-rat IgG antibody conjugated with Alexa Fluor 647 and goat anti-rabbit IgG antibody conjugated with Alexa Fluor 488 were further stained for 1 h at

RT PPs were washed 3 × 5 min with PBS and mounted with the anti- fade reagent Images were captured by a confocal laser microscope (SP8, Leica Microsystems)

2.14 H&E staining

Frozen sections of PPs from mice were washed three times with PBS

In H&E staining, the sections were stained with Harris' hematoxylin, then treated with 1 % acid alcohol, and finally stained with 1 % eosin Excess eosin was removed by washing in running water Sections were mounted on slides with permount Images of H&E staining were taken with light microscopy

2.15 Study of human subjects

Twenty volunteers who needed surgery were enrolled in this study The age of the participants ranged from 20 to 50 years All distal ilea of the small intestine were collected after their surgeries Some of them were not qualified because they had no PPs Most of them only have one

PP 32 PPs isolated from twenty samples of distal ilea were used in this study In brief, the small intestines with PPs were stored in ice-cold PBS with 5 % FBS Each small intestine segment was ligated with a surgical suture for the ligated loop assay It was then injected with 10 mL FITC- RAP solution (5 mg/mL) followed by incubation at 37 ◦C in an atmo-sphere of 5 % CO2 for 2 h 20 PPs were collected for whole-mount staining, and 12 PPs were for immunofluorescence staining

2.16 Statistical analysis

Each experiment was independently repeated at least three times with consistent results As noted in figure legends, all data are shown as mean ± SD Statistical differences between each experimental group

were analyzed by a t-test or one-way ANOVA Significant differences with P < 0.05 were considered significant

3 Results

3.1 RAP specifically appears in FAE cells isolated from Peyer's patches

Our newly published data showed that intact Radix Astragali poly-saccharide RAP entered Peyer's patches (PPs) within 1 h to initiate im-mune responses (Zhang et al., 2022) We further screened the intestinal epithelial cells that may have been involved in the lymphatic transport Firstly, we confirmed the fast transport of intact RAP into PPs by an intestinal loop assay Images taken by the living system showed that within 1 h after sample loading to the loop, PPs were found to contain

FITC-RAP (P < 0.001, Fig 1A and B) High-performance gel permeation chromatography confirmed that the fluorescence signal came from intact FITC-RAP (Fig 1C) After confirming the transport, we isolated epithelial cells from mice orally treated with FITC-RAP Before analysis

by flow cytometry, we firstly confirmed the gate of intestinal epithelial cells (IECs) and excluded the CD45+ cells as shown in Supplementary Fig 1 With the similar gate strategy of IECs, flow cytometry analysis of these isolated cells indicated that only the cells from the FAE of PPs-no

other IECs-specifically accumulated FITC-RAP (P < 0.001, Fig 1D–G)

Therefore, we concluded that the transport of intact RAP into PPs is

associated with the FAE cells

3.2 Transport of intact RAP by M-like cells in vitro

We further tested the possibility that M cells may be the transporter

using an in vitro FAE cell model As shown in Fig 2A, we first evaluated

the expression of the M cell marker and validated its transporter

func-tion using control nanoparticles (Beloqui et al., 2017) Confocal images

of inserts showed that M-like cells were successfully characterized with

M cell-specific marker GP2 (Fig 2B) The flow cytometry analysis showed that the model group containing Caco-2 cells co-cultured with Raji B cells was permeable to nanoparticles, while the Caco-2 monolayer was not (Fig 2C and D) Thus, the induction of M-like cells, the integrity

of Caco-2 monolayers, and the transport function of M-like cells were confirmed

In the tests with FITC-RAP, specifically in the immunofluorescence staining assay, we found that FITC-RAP was perfectly co-localized with GP2+ M-like cells (Fig 3A) To figure out how much FITC-RAP was transported by the M-like cell model, we used flow cytometry (FCM) to detect the amount of FITC-RAP in the lower layer supernatant Flow

Fig 1 Selective transport of RAP in PPs (A) Images of PPs (n = 8) collected from mice in a ligated intestinal loop mouse model after FITC-RAP treatment using IVIS

Lumina XR Small Animal Imaging System (B) Fluorescence intensity of the region of interest (ROI) of the ex vivo images taken above (C) HPGPC-FLD chromatograms

of PPs with and without the FITC-RAP treatment for 1 h (D and E) Flow cytometry dots (D) and percentage (E) of intestinal epithelial cells (IECs) in cell suspension isolated from PPs and small intestine segments without PPs (SI) These tissues were collected from mice treated with FITC-RAP (10 mg/kg) by oral administration (F and G) Flow cytometry histograms (F) and percentage (G) of FITC-RAP-binding IECs gated from D Error bar indicates SD (n = 10) Data are shown as mean ± SD and

are representative of at least three independent experiments Significant difference ***P < 0.001, ****P < 0.0001

cytometry analysis of the lower layer of transwell inserts indicated much

more FITC-RAP was detected in the co-culture group, suggesting the

polysaccharide RAP was transported by M-like cells in the in vitro model

(Fig 3B and C) Furthermore, we also detected whether FITC-RAP was

degraded in the transport process using HPGPC assay The HPGPC

chromatograms clearly showed that the FITC-RAP in the lower layer

supernatants had the same molecular weight with primary FITC-RAP,

suggesting the transported FITC-RAP remained intact (Fig 3D) By

contrast, the monolayer of Caco-2 cells was not permeable for FITC-RAP

We also tested unlabeled RAP to exclude the effect of FITC using the

M-like cell model Our previous studies have demonstrated that RAP

strongly induced cytokine production by macrophages, such as

RAW264.7 cell line, in a dose-dependent manner (Li et al., 2017; Wei

et al., 2016; Wei et al., 2019) Therefore, cytokine production of

mac-rophages here was used to detect the transported RAP in the lower layer

of the co-culture by comparing with RAP treatment at different

con-centrations RAP standard solutions of varying concentrations (0.1, 1,

10, and 100 μg/mL) were used as control The results indicated that in

terms of both IL-6 and IL-12 production, the transported RAP in the co-

culture group exhibited significant inducing effects (Fig 3E and F),

which were comparable with 1 μg/mL of RAP standard solution It is suggested that M-cell's transport could be so efficient that the trans-ported RAP could reach a concentration as high as 1 μg/mL

3.3 Transport of RAP by M cells in vivo

The role of the M cell as the key transporter cell was further verified

in vivo Firstly, we need to identify M cells in PPs tissues Using whole-

mount staining of PPs, we found the dome zone of FAE in PPs (Fig 4A), which is the location of M cells We compared two reported specific M-cell markers, GP2 and NKM 16-2-4, using confocal micro-scopy and found that M cells were successfully labeled with both markers (Fig 4B–D) (Hase et al., 2009; Nochi et al., 2007) We also prepared frozen sections of PPs and identified GP2+NKM16-2-4+ M cells in the FAE of PPs (Fig 4E) The H&E staining image showed the structure of the FAE of PPs (Fig 4F), where M cells are located in Thus, these two specific markers, GP2 and NKM 16-2-4, were well validated

for definitely identifying M cells in vivo

After identifying M cells in vivo, we tracked orally administrated

FITC-RAP With the same method in Fig 1D, we gated IECs for verifying

Fig 2 Establishment and validation of the in vitro M cell-like model (A) Protocol used to establish the in vitro M cell model and its application (B) Confocal

microscopic images of co-culture transwell inserts stained with anti-GP2 mAb and subsequently with Alexa 647-conjugated secondary antibody (red) DAPI (blue) was used as a DNA-specific stain Scale bars, 50 μm (C) Flow cytometry detection of nanoparticles in the supernatant collected from the lower layer of monoculture

and co-culture groups (n = 3) (D) Percentage of nanoparticles transported by the M cell model Data are shown as mean ± SD Significant difference ****P < 0.0001 All in vitro experiments repeated at three times, respectively

the transport role of M cells using the flow cytometry analysis As shown

in Fig 5A and B, results of flow cytometry showed that GP2+ cells in

IECs of PPs were much higher than that in IECs of SI, suggesting that

GP2+ M cells are located in the FAE of PPs Further analysis of these

GP2+ M cells indicated that GP2+ M cells instead of other IECs

accu-mulated FITC-RAP (Fig 5C and D) The specific transport of FITC-RAP

by GP2+ M cells was further demonstrated by confocal microscopic

images of fixed and stained whole-mount preparation (Fig 5E) of PPs

isolated from mice 1 h after FITC-RAP treatment Consistently, three-

dimensional confocal micrographs demonstrated the accumulation of

GP2 around FITC-RAP, which was internalized into M cells from the

apical side to the basal side (Fig 6A–C and Supplementary movie 1)

Taken together, these data showed that M cells of PPs in the small

in-testine could independently transport RAP

In our previous study, we reported that follicle dendritic cells (FDCs)

were RAP's direct target immune cells in the PPs Here, we further

confirmed this finding using an immunofluorescence staining assay to

observe the co-localization of RAP, M-cell and FDCs As shown in Fig 6D

and Supplementary movie 2, GP2+ M cells in FAE are localized with

FITC-RAP in the FAE of PPs, and FITC-RAP is simultaneously localized with CD11c+ DCs in the sub-epithelial dome, which is consistent with the results of our previous flow cytometry analysis (Zhang et al., 2022) These results precisely show that RAP will directly contact FDCs after being transported into PPs

3.4 Transport of RAP by M cells in the human distal ileum

As M cells exist in both mice and humans, we hypothesize that this interesting M-cell mediated transcytosis may also occur in the human body So, we collected live distal ileum explants from surgery patients and screened them for PPs (Fig 7A and B) and incubated the ileum containing PPs with FITC-RAP M cells are located in the follicle asso-ciated epithelium (FAE) of PPs (Sakhon et al., 2015) In photos of stained whole-mount of sections, the FAE zone of PPs is denoted by a white circle (Fig 7C) To confirm the location of M cells and M cell-associated RAP transport, we first collected PPs and stained with GP2 cell marker

As shown in Fig 7D, FITC-RAP was recruited of puncta, most of which were positive for the M-cell marker GP2, suggesting GP2+ M cells

Fig 3 Transport of RAP in the M cell-like model (A)

Confocal microscopic images of co-culture transwell inserts treated with FITC-RAP (green) and then stained with GP2 mAb and subsequently with Alexa 647-conjugated secondary antibody (red) DAPI (blue) was used as a DNA-specific stain Scale bars,

25 μm (B and C) Flow cytometry histogram of FITC- RAP (B) and transport percentage (C) in the super-natant collected from lower layers of monoculture and co-culture groups (n = 9 per group) (D) HPGPC- FLD chromatograms of FITC-RAP in the supernatants collected from lower layers of monocultures and co- cultures (n = 9 per group) (E and F) IL-6 (E) and IL-12 (F) secretion of RAW264.7 cells treated with supernatants collected from the lower layer of monocultures and co-cultures (n = 9 per group) 2 h after loading 200 μL RAP (2 mg/mL), compared with cytokines produced from RAW 264.7 cells treated with RAP at different concentration (0.1, 1, 10, and

100 μg/mL) Data are shown as mean ± SD and are representative of three independent experiments

Significant difference *P < 0.05, ****P < 0.0001, ns

=no significance

located in the FAE of PPs could transport FITC-RAP like what happed in

mice The two signals were not coincidental, which might be due to the

dynamic transcytosis procedure Three-dimensional images further

confirmed and demonstrated the dynamic transport of FITC-RAP in M

cells from the apical surface to the subepithelial dome of PPs (Fig 7E

and Supplementary movie 3) Confocal microscopic images of frozen

sections of PPs confirmed that FITC-RAP has been transported into PPs

(Fig 7F) These findings demonstrate that RAP is transported by human

M cells into PPs

4 Discussion

An in vitro M-like cell model has been used to evaluate whether M

cells transport polysaccharides (Feng et al., 2020; Li et al., 2021; Xiang

et al., 2020; Zheng et al., 2022), but the model needs validation First, it

does not include a nanoparticle control to test the transport function

(Rieux et al., 2005; Rieux et al., 2007) The results would be doubtable

without control Using the well-studied nanoparticle as a control will be

helpful to judge whether the model is successfully established Second,

the presentation of M-like cells should be verified using a specific cell

marker Actually, this cell model is often directly used to test for

transcytosis without determining if M-like cells existed in this model

(Gibb et al., 2021) Even when the M-like cells were labeled in a few of

these tests, the cell marker in labeling was not specific For example,

Ulex europaeus agglutinin-1 (UEA-1) was used as the marker of M-like

cells (Jiang et al., 2019), but it had been found to be non-specific for M

cells (Terahara et al., 2008; Zheng et al., 2022) The current study is the

first in which GP2, which has been identified as a specific M-cell marker,

was used to test for polysaccharide transcytosis In in vivo studies, we

even used two cell markers, GP2 and NKM 16-2-4 (Hase et al., 2009;

Nochi et al., 2007), to validate the existence of M cells Third, it should

be determined if the transported polysaccharides were already degraded Few reports addressed this because the detection solely relied

on the fluorescence signal As we found previously when we tracked the orally dosed FITC-RAP in the caecum and colon using HPGPC-FLD, the fluorescence signal might come from degraded chemicals Fourth, the unlabeled polysaccharide should be tested by using other biochemical assays In the current report, we also observed a weak HPLC signal in the lower layer of the monoculture insert loaded with FITC-RAP But that loaded with unlabeled RAP was completely inactive in the ELISA assay Therefore, the weak signal observed FITC-RAP monoculture control might be attributable to a false positive result It has been reported that

this kind of weak signal always exists in the in vitro M-like cell model and

is hard to remove (Ahmad et al., 2017; Beloqui et al., 2017) It is also possible that FITC labeling changed the transcytosis property of RAP in

the in vitro cell model; this possibility deserves further investigation It is

suggested that the fluorescence signal is unreliable and needs to be validated with other analyses An isotype control of FITC might be helpful to address this concern

The popularly used in vitro M-like cell model has its own limitations

This model does not simulate all factors in the transport of intestinal antigens The small intestinal epithelial cells are heterogenous and act as

a passive barrier to prevent the transit of antigens and pathogens (Nagler-Anderson, 2001; Romero et al., 2015; Vancamelbeke & Ver-meire, 2017) Some of the epithelial cells are specialized for producing mucus as a vital physicochemical barrier (Grondin et al., 2020; Pela-seyed et al., 2014) This model only addresses the similar transcytosis

function Herein, using an in vivo model, we screened the small

in-testine's epithelial cells and proved for the first time that GP2+ M cell

was the specific transporter cell for polysaccharide RAP The in vivo study showed several advantages First, the in vitro cell model includes multiple time-consuming operations over 3 weeks, while the in vivo

Fig 4 M cells in the small intestine recognized by

the specific markers GP2 and NKM 16-2-4 (A) Confocal microscopic images of whole-mount PPs stained with DAPI (blue) DAPI (blue) was used as a DNA-specific stain The location of FAE in PPs is marked by a dotted white circle Scale bars, 250 μm (B) Confocal microscopic images of the FAE area of whole-mount PP stained with anti-NKM 16-2-4 mAb and anti-GP2 mAb and subsequently with Alexa 488- labeled (green) and Alexa 647-labeled (red) second-ary antibody Scale bars, 250 μm (C) Confocal microscopic image of the FAE area of whole-mount

PP stained with anti-NKM 16-2-4 mAb and anti-GP2 mAb Scale bars, 75 μm (D) The enlarged area was marked in C by a dotted white frame Scale bars, 25

μm (E) Confocal microscopic images of frozen PP tissue isolated from wild type BALB/c mice stained with anti-NKM 16-2-4 mAb and anti-GP2 mAb and subsequently with Alexa 488-conjugated (green) and Alexa 647-conjugated (red) secondary antibody Scale bars, 25 μm (F) Cryosections of ileal PPs stained with H&E and visualized by light microscopy The location

of FAE M cells in PPs is marked by a dotted white line Scale bars, 100 μm Abbreviations: FAE, follicle- associated epithelium; SED, sub-epithelial dome; LF, lymphoid follicle

assay could be quickly completed in 2–3 h Second, the in vivo study

involved the real, multi-cellular organ/tissue environment rather than a

single type of cells Third, the in vivo study enabled further observation

of which cell the polysaccharide directly targeted in PPs Thus, the in

vivo approach we established here might be a better option to investigate

polysaccharides' intestinal lymphatic absorption through M-cell

transcytosis

The M cell cell-mediated transcytosis of RAP might be receptor-

mediated There are three possible pathways for the M cell-mediated

antigen transcytosis: non-specific transcytosis, receptor-mediated

transcytosis, and extension of DCs through M cell transcellular pores

(Dillon & Lo, 2019; Kim & Jang, 2014; Lelouard et al., 2012; Mabbott

et al., 2013; Neutra et al., 2001) In our previous study, Dendrobium

officinale polysaccharide DOP showed an identical destiny after oral

administration (Li et al., 2019) Being indigestible and nonabsorbable, it

ended in regulating gut microbiota as it degraded to short-chain fatty

acids DOP was used as a control in the tests of the lymphatic route for

RAP to enter PPs (Zhang et al., 2022) The results showed that only RAP

rather than DOP entered PPs It is suggested that this lymphatic route is

selective Here M cells have been identified as the specific transporter

cell in this route, so we hypothesize that RAP's transcytosis might be

mediated by some specific receptors of M cells TLR4 is one possible

receptor because it has been reported to mediate the RAP-induced

signaling pathway of macrophages (Wei et al., 2016) GP2 is another possibility because it is not only the specific M-cell marker but also the receptor mediating the antigen sampling of PPs (Hase et al., 2009) Furthermore, GP2 is common in both mice and humans, and we demonstrated that human M cells have the same function in transporting RAP Therefore, we hypothesize that TLR4/GP2 is the receptor that mediates the M-cell transcytosis of RAP This hypothesis deserves testing

The M cell-mediated lymphatic route demonstrated here with human subjects is scientific evidence for the potential medicinal value of many immunomodulatory polysaccharides Whether these polysaccharides are found to be effective for mice is not necessarily relevant, although mouse studies are common There are big differences between the mice model and the human clinical trial Many promising drug candidates that fail are called “mouse drugs” because they are eventually found to

be effective in mice but not humans Despite this risk, mice models are still the best choice for most medicine R&D projects due to the concerns

of cost and regulations Therefore, there is a huge research gap between the mice study and the clinical trial of polysaccharide-based medicines Efficacy in mice is not a guarantee of efficacy in humans Another problem is studies of mechanisms It is hard to investigate the underlying mechanisms in clinical trials Here we demonstrated that the efficient M cell-mediated lymphatic route is common between mice and humans,

Fig 5 Transport of RAP by M cells in vivo (A and B) Flow cytometry density plot (A) and percentage (B) of GP2+ M cells in intestinal epithelial cells (IECs), which

are gated in FSC-SSC flow cytometry dots by analyzing cells isolated from PPs and small intestine without PPs (SI) The gate strategy of IECs is same with the IECs gate in Fig 1D These tissues (n = 10) are collected from mice treated with FITC-RAP (10 mg/kg) by oral gavage (C and D) Flow cytometry histogram (C) and percentage (D) of FITC-RAP-binding GP2+ M cells gated from A (E) After the ligated intestinal loop assay with FITC-RAP, PPs were stained with anti-GP2 mAb and DAPI (blue) Uptake of FITC-RAP (green) by GP2 + M cells (red) Scale bars, 25 μm Data are shown as mean ± SD and are representative of three independent

experiments Significant difference **P < 0.01, ****P < 0.0001

Fig 6 Follicle dendritic cells (FDCs) are the direct target immune cell of M cell-transported RAP in PPs (A) The X-Y angle of three-dimensional confocal image of

whole-mount PP stained for anti-GP2 mAb (red) and DAPI (blue) PPs were isolated from the ligated intestinal loop assay with FITC-RAP (green) The three- dimensional movie was shown in Movie S1 (B) Clockwise 45 degree-rotation image of A (C) Clockwise 45 degree-rotation and amplified image of B In this image, FITC-RAP was transcytosed by an M cell from the apical side to the M cell pocket which marked by a dotted line (D) Confocal images of PP frozen sections (dome zone), collected from the ligated intestinal loop assay treated with FITC-RAP (green), and stained with GP2 (red), CD11c (yellow), and DAPI (blue) Three- dimensional image of PP indicating that FITC-RAP (green) was transported into a sub-FAE layer and contacted CD11c+ dendritic cells (yellow) Arrows indicate FITC-RAP The three-dimensional movie was shown in Supplementary movie 2

successfully bridging this research gap RAP has been verified to have

diverse beneficial effects, such as antitumor, but it has never been tested

on humans Considering that most of RAP's bioactivities are associated

with the immune system and that RAP has been known to quickly enter

PPs and directly target FDCs to trigger immune responses, we

hypothesize that RAP follows the same lymphatic route to exhibit similar immune-related beneficial effects in human bodies In other words, RAP's bioactivities and mechanisms observed in the mice model should also apply to human bodies And this route may be the way many other active polysaccharides work This finding may attract great

Fig 7 Transport of RAP by M cells of human ileum PPs (A) Section of human distal ileum with PPs indicated by a black arrow (B) Enlarged view of the distal ileum

tissue section in A, with a black arrow pointing to a PP (2.5×) (C) Confocal microscopic images of the whole-mount PP domes stained with DAPI (blue), as a DNA- specific stain The dome zone, which is the location of FAE in PPs, is marked by a dotted white circle Scale bars, 100 μm (D) Confocal microscopic images of the whole-mount PP domes stained with anti-human GP2 antibody and DAPI, showing that FITC-RAP (green) was bound to GP2+ M cells (red) Scale bars, 100 μm (E) Three-dimensional confocal microscopic images of FAE FITC-RAP (green) was internalized by human GP2+ M cells (red) The three-dimensional movie was shown

in Supplementary movie 3 Scale bar, 8 μm (F) Confocal microscopic image of frozen sections of human ileal PPs stained with DAPI (blue) FITC-RAP (green) was found in the SED of PPs Scale bars, 25 μm Abbreviations: FAE, follicle-associated epithelium; SED, sub-epithelial dome; LF, lymphoid follicle