Accepted Manuscript Title: Epithelial cells are a source of natural IgM that contribute to innate immune responses Author: Wenwei Shao Fanlei Hu Junfan Ma Chi Zhang Qinyuan Liao Zhu Zhu Enyang Liu Xiaoyan Qiu PII: DOI: Reference:
S1357-2725(16)30017-6 http://dx.doi.org/doi:10.1016/j.biocel.2016.01.017 BC 4784
To appear in:
The International Journal of Biochemistry & Cell Biology
Received date: Revised date: Accepted date:
30-8-2015 1-12-2015 22-1-2016
Please cite this article as: Shao, W., Hu, F., Ma, J., Zhang, C., Liao, Q., Zhu, Z., Liu, E., and Qiu, X.,Epithelial cells are a source of natural IgM that contribute to innate immune responses, International Journal of Biochemistry and Cell Biology (2016), http://dx.doi.org/10.1016/j.biocel.2016.01.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Manuscript
Epithelial cells are a source of natural IgM that contribute to innate immune responses Wenwei Shao1*, Fanlei Hu1, 2*, Junfan Ma3, Chi Zhang1, Qinyuan Liao1, Zhu Zhu1,
1
ip t
Enyang Liu3, Xiaoyan Qiu1, 3† Department of Immunology, School of Basic Medical Sciences, Peking University
Department of Rheumatology and Immunology, Peking University People’s
us
2
cr
Health Science Center, Beijing, 100191, China
Hospital, Beijing, China
Key Laboratory of Medical Immunology, Ministry of Health, Beijing, 100191,
an
3
†
These authors contributed equally to this work.
Corresponding author: Xiaoyan Qiu, MD, PhD, Center for Human Disease
ed
*
M
China
Genomics, Peking University, 38 Xue-yuan Road, 100191, Beijing, China. Tel.:
Ac
ce pt
86-10-82805477; Fax: 86-10-82801149; E-mail: [email protected].
Page 1 of 35
ABSTRACT Currently, natural IgM antibodies are considered to be the constitutively secreted products of B-1 cells in mice and humans. In this study, we found that
ip t
mouse epithelial cells, including liver epithelial cells and small intestinal epithelial cells (IECs), could express IgM that also showed natural antibody
cr
activity. Moreover, similar to the B-1 cell-derived natural IgM that can be
us
upregulated by TLR9 agonists (mimicking bacterial infection), the expression of epithelial cell-derived natural IgM could also be significantly increased by TLR9
an
signalling. More importantly, the epithelial cell-derived IgM was polyreactive,
M
and it could recognize single-stranded DNA (ssDNA), double-stranded DNA (dsDNA), lipopolysaccharide (LPS), and insulin with low affinity; additionally,
ed
TLR9 agonists could enhance it in a MyD88-dependent manner. Furthermore, epithelial cell-derived IgM could bind various bacteria; therefore, it could be
ce pt
involved in anti-infection responses. Together, these results highlight the fact that epithelial cells are an important source of natural IgM, in addition to that
Ac
produced by B-1 cells, and IgM contributes to the innate immune responses in local tissues, further demonstrating that the epithelium is a first line of defense in the protection against invading microbes.
Key words: Epithelial cells • Natural IgM • TLR9 agonists • Innate immune response
Page 2 of 35
INTRODUCTION Immunoglobulin (Ig) was previously thought to be produced by only B-lineage cells. However, in the last decades, the results from a series of studies have indicated that
ip t
non-B cells can also produce Igs. As early as 20 years ago, studies by our group showed that many types of non-B cancer cells, especially epithelial cancer cells, could
cr
also express Igs, including IgG, IgA and IgM (1-7, 8, 9). Moreover, the epithelial
us
cancer cells that expressed IgG showed growth factor-like activity that could promote cancer progression (1, 2, 10). Subsequently, other researchers have also demonstrated
an
these unusual findings (11-19). Recently, growing evidence has been found to support
M
the finding that normal non-B cells, including epithelial cells, endothelial cells, neurons, germ cells, and even monocytes, can also express Igs, such as IgG, IgA and
ed
IgM (4, 5, 13, 20, 21). All of these studies indicated that the classical concept suggesting that B cells are the only source of Igs should be updated. However, the
ce pt
detailed physiological significance of non-B cell-derived Igs remains unclear. Natural antibodies are preformed antibodies that are present, even in naive germ-free
Ac
mice, in the absence of any exogenous antigenic exposure. Consistent with their specificities for microbial antigens, natural antibodies play an important non-redundant role in the first line of defence against bacterial and viral infections. Additionally, natural antibodies have also been shown to have specificities for self-antigens and have therefore been proposed to provide important homeostatic "house-keeping" functions. Thus far, B-1 cells that express CD5 have been considered as the major source of natural antibodies, including natural IgM, IgG and IgA (22, 23).
Page 3 of 35
Particularly, B-1a cells, which are enriched in the peritoneal cavity, have been considered as the only source of natural IgM. However, the B-1a cells in the peritoneal cavity have been proven to poorly contribute to serum IgM levels (24, 25),
ip t
suggesting that in addition to B-1 cell, there are other cell types involved in the production of natural IgM. Recently, many epithelial cell lineages, including breast,
cr
colon, and epidermal squamous cells, were found to spontaneously express Igs,
us
including IgG, IgA and IgM. Importantly, epithelial cell-derived Igs showed natural antibody activity. For example, the IgG and IgA that are spontaneously secreted by
an
epidermal squamous cells showed anti-bacterial activity (9). A cervical cancer cell
single-stranded
DNA
(ssDNA),
M
line (HeLa)-derived IgM displayed natural IgM activity; it could recognize double-stranded
DNA
(dsDNA)
and
ed
lipopolysaccharide (LPS) and was upregulated by a TLR9 agonist (26). By integrating these facts, we hypothesize that in addition to B-1 cells, epithelial cells are an
ce pt
important source of natural antibodies, and they might play an important role in innate immunity via participating in humeral immunity.
Ac
The IgM isotype represents one of the major Ig classes in the body. IgM is also the earliest type of membrane-associated Ig expressed during B cell ontogeny and secreted during Ag-specific immune responses. IgM can be divided into two types: circulating IgM that exists independent of known immune exposure, which are referred to as natural IgM (nIgM), and immune IgM, which is generated in response to defined antigenic stimuli. In healthy adults, circulating polyclonal IgM is generally present at 1 to 2 mg/ml in the blood. As the first line of defence against invading
Page 4 of 35
microbes or stress damage, polyclonal nIgM can directly recognize a wide range of pathogen-associated molecular patterns (PAMPs) on different microbial pathogens or damage-associated molecular patterns (DAMPs), such as ssDNA and dsDNA, which
ip t
are released by damaged or necrotic tissues. nIgM usually displays polyreactivity and triggers a rapid humeral immune response (2-7, 11).
cr
In this study, using liver and small intestine epithelial cells in adult Balb/c mice as
us
models, we further revealed IgM expression and expression of its functional transcripts in epithelial cells that were sorted by flow cytometry. Moreover, we
an
showed that IgM expression and secretion were significantly upregulated by CpG
M
ODN, a synthetic analogue of bacterial DNA. Importantly, we further identified the natural antibody properties of this IgM.
ed
RESULTS
IgM expression in the epithelial cells of Balb/c mice
ce pt
We first determined whether IgM is present in epithelial cells, including epithelial cells from the pancreas, liver, lung, kidney, stomach, uterus and small intestine of
Ac
mice. Immunohistochemistry studies showed that IgM was present in the cytoplasm of these epithelial cells (Fig. 1A). To further elucidate that IgM could be spontaneously expressed by this cell type, epithelial cells were isolated and sorted from the liver and small intestine of mice according to their cytokeratin (CK) expression profiles using flow cytometry (CK18 for liver epithelial cell, and CK8 for IECs) (Fig. 1B, D), and the IgM gene rearrangement and transcription levels were assessed. To exclude the possibility of B cell contamination, the sorted epithelial cells
Page 5 of 35
was evaluated for the absence of transcripts of CD19, CD20, and CD138. Then, specific primers were used to amplify the constant and variable regions of the Ig κ and μ chains, respectively. This analysis demonstrated the presence of rearranged Ig µ and
ip t
Ig transcripts in liver epithelial cells and IECs (Fig. 1C, E). IgM was also detected in sorted liver epithelial cells and IECs with goat anti-mouse IgM by Western blot
cr
analysis (Fig. 1F). FACS staining with PE-conjugated goat anti-mouse IgM further
us
confirmed the existence of the IgM protein (Fig. 1G).
Sequence analysis of the variable regions of the Ig µ and transcripts
an
The IgM gene transcripts and repertoires of the liver epithelial cells and IECs from six
M
Balb/c mice were detected. Spleen cell-derived IgM gene transcripts served as the positive control. The sequencing results demonstrated that 30 sequences of 32 VHDJH
ed
rearrangements and all of the 42 VκJκ rearrangements from the liver epithelial cells showed a functional and typical V(D)J rearrangement pattern. Additionally, nIgM
ce pt
from B-1 cells are often without N-region additions and are germline-encoded or with minimal somatic hypermutations. The sequence analysis revealed that although
Ac
almost all of the liver epithelial cell-derived Ig µ transcripts had N-regions, all of the Ig transcripts showed no N-region additions. Moreover, both of the Ig µ and Ig transcripts possessed germline or near germline sequences. Different from the rearrangement diversity that was observed in the spleen cells, the liver epithelial cell-derived Ig µ variable regions showed conserved usage in different individuals (Table 1). VIGHV1-37*01/D
IGHD2-13*01/JIGHJ2*01,
the predominate VDJ rearrangement
pattern in two of the five mice, was the preferential usage in the epithelial cells.
Page 6 of 35
Similarly, the rearrangement patterns of the Ig variable region in the liver epithelial cells revealed conserved usage in different individuals (Table 2). We also obtained the transcripts of the Ig variable region in the IECs, which showed similar
IGKV9-124*01J IGKJ2*01
ip t
characteristics to that in the liver epithelial cells (supplementary Table 1). V recombination occurred in both the liver epithelial cells and the
cr
IECs but not in the spleen cells.
us
A TLR9 agonist stimulated epithelial cells to secrete IgM in mice
The TLR9 signalling pathway is an important regulator for IgM secretion in B cells.
an
To determine whether IgM from epithelial cell has effects on bacterial invasion
M
defence, the epithelial cell IgM expression levels were detected when mice were stimulated with the TLR9 agonist, CpG, which mimics bacterial infection. TLR9
ed
expression in liver and small intestine epithelial cells was detected (data not shown) by IHC. Additionally, the TLR9 transcripts in these cells were further confirmed by
ce pt
RT-PCR (Fig. 2A). Subsequently, the TLR9 signalling pathway regulating IgM secretion in epithelial cells was activated carried out by the CpG treatment. As
Ac
expected, the mouse serum IgM levels were significantly elevated (Fig. 2B). Importantly, the result revealed that the IgM transcription after CpG stimulation (but not with a GpC control or PBS) was significantly upregulated in flow cytometer-sorted liver epithelial cells and IECs (Fig. 2 C, D, E, F). The Western blot results further revealed that the IgM levels in the liver and small intestine were increased after stimulation with CpG (Fig. 2G). Moreover, the CpG induction punctuated the accumulation of IgM in the extracellular matrix of the liver epithelial
Page 7 of 35
cells and in the small intestinal lumen (Fig. 2H). ELISPOT results further confirmed that the CpG treatment promoted IgM secretion of these epithelial cells (Fig. 2I). Epithelial cell-derived IgM displayed natural IgM properties
ip t
The fact that IgM was spontaneously expressed in epithelial cells guided us to explore its potential functions as a natural antibody. Natural antibodies are often polyreactive
cr
and can bind to multiple antigens with low affinity. Natural antibody activities of IgM
us
produced from several epithelial tissues of mice, including liver, pancreas, lung and small intestine, were analyzed. Using an ELISA assay, we found that, similar to the
an
B-1 cell-derived nIgM, the epithelial cell-derived IgMs also showed polyreactivity
M
compared to the non antigens-coating microwells, they could bind ssDNA, dsDNA, LPS, insulin as well as different types of microbes (Fig. 3Aa). Our results revealed that
ed
epithelial cell-derived IgM showed a similar reaction degree to the spleen-derived IgM under conditions where the total IgM levels were lower in the liver than in the
ce pt
spleen (Fig. 3Ab). Interestingly, different types of epithelial tissue-derived IgM displayed their own dominant identification profile, such as liver, lung or spleen-IgM
Ac
dominantly recognize Staphylococcus aureus (S. aureu), pancreas-IgM dominantly recognize DNA, small intestine-IgM, without obvious tendency, recognize all of antigens detected in this study. To determine whether the anti-IgM antibody tested was detecting something other than IgM, we used purified mouse IgM (Commercialization) to block the anti-IgM antibody before it was used in ELISA assay. The result showed that when anti-IgM antibodies were block with mouse IgM, the binding abilities of cell lysates from liver or spleen to different antigens were
Page 8 of 35
significantly decreased (Figure 3c). To further prove that the monoclonal IgM has polyreactivity, the recombinant Ig μ with the VIGHV1-37*01/D
IGHD2-13*01/JIGHJ2*01
rearrangement pattern, which was found to be frequently expressed by liver epithelial
ip t
cells, was expressed by Escherichia coli (E. coli) (Fig. 3B,C). As shown in Fig. 3D, the recombinant Ig μ was polyreactive to ssDNA, dsDNA and insulin, but it was not
cr
reactive to LPS.
bacterial and bacterial DNA analogue challenge
us
The polyreactivity of epithelial cell-derived IgM could be enhanced after
an
Generally, natural IgM antibodies, which are secreted in the absence of antigenic
M
challenge, are important contributors to antimicrobial immunity. In this study, whether CpG could promote secretion of the nIgM from epithelial cells was essential to
ed
determine, although the total IgM level could be upregulated by CpG, as described above. As is well known, the natural IgM activity was significantly enhanced
ce pt
following CpG exposure, particularly against ssDNA, dsDNA and LPS (Fig. 4A). Importantly, the promoting effect of CpG on the nIgM polyreactivity in liver epithelial
Ac
cells was abolished when the myeloid differentiation primary response gene 88 (MyD88) gene (the cytosolic adapter protein for TLR9) was knocked out (Fig. 4B). Additionally, we further analyzed whether there was a sequence change in the Ig µ chain variable region after CpG challenge. The results showed that the usage of the preferential VDJ rearrangement pattern in the liver epithelial cells was not significantly changed.
Page 9 of 35
The natural antibody properties of epithelial cell-derived IgM were further characterized by assessing the reactivity against various bacterial pathogens. As expected, the epithelial cell-derived IgM could recognize different bacteria with low
ip t
affinity, including E. coli, Staphylococcus aureus, Salmonella typhi, Paratyphoid bacilli, Proteus vulgaris, and Shigella. Importantly, immunization with inactivated E.
cr
coli enhanced the reactivity against all 6 pathogens (Fig. 5A-F), and this result was
us
accompanied by increased IgM expression (Fig. 5G).
DISCUSSION
an
In this study, we revealed IgM expression in normal epithelial cells that frequently
M
encounter pathogen infections in adult mice, and we observed that TLR9 agonists could stimulate these epithelial cells to secrete IgM. More importantly, the epithelial
ed
cell-derived IgM showed natural antibody properties. Our findings highlight, for the first time, a novel concept that in addition to the IgM that is produced by B1 cells,
ce pt
many epithelial cell types are important sources of natural IgMs, in addition to that produced by B1 cells, which contribute to innate immune function.
Ac
The characteristics and functions of IgM have been well studied. Many properties of IgM make this immunoglobulin particularly well-suited for its role in microbial immunity. It is present in high concentrations in the blood (in the range of 1.5 mg/ml) and is the first antibody elicited in an immune response following immunization or infection. IgM has a relatively short half-life, approximately 28 h, in the serum of normal mice in the absence of antigen (27). IgM is expressed as a membrane-bound antibody by all naive B cells, but B cells also secrete it as a pentamer (28). The
Page 10 of 35
majority of IgM is mainly generated from germline-configured transcripts. This generation occurs prior to the onset of class switch recombination (CSR) and somatic hypermutation (SHM); therefore, IgM provides a first line of defence during
ip t
microbial infections (29). Natural IgMs constitute the majority of circulating IgMs, and most natural IgM antibodies are germline-encoded and bind with low affinity to a
cr
number of different antigens (including microbial antigens), thereby contributing to
us
effective pathogen clearance by innate defense mechanisms without necessitating activation of the adaptive humeral response. Additionally, ‘innate’ IgM accelerated
an
microbial pathogen clearance by innate immune mechanisms (30, 31, 32). Although
M
B-1 cells have been long associated with natural IgM, there is no study showing that non-B-1 cells significantly contribute to serum IgM levels in naïve mice.
ed
It was previously found that not only IgG and IgA but also IgM were expressed in some non-B cells, including breast cancer cells, colon cancer cells, and cervical
ce pt
cancer cells (1, 3, 6). Moreover, similar to B cell-derived IgM, the non-B cancer cell-derived IgM showed almost no somatic hyper mutations in their V region,
Ac
namely, the germline configured transcripts. However, the γ chain-derived Ig heavy chain revealed typical somatic hyper mutation in the same cells (6). Importantly, some evidence suggested that non-B cells, especially epithelial cells, might produce natural IgM (26, 33). The epithelium has immense importance in host defense and immune surveillance, as it is the primary cell layer that initially encounters the majority of microorganisms. Originally, it was thought that the epithelium serves as only a passive barrier against
Page 11 of 35
invading pathogens. Barrier function alone is usually adequate to restrain commensal microbes; however, it is often insufficient to protect against microbial pathogens. However, recently, it has become apparent that epithelial cells are capable of
ip t
triggering an immune response that is similar to that of myeloid lineage cells, thus playing a crucial role in the active recognition of microbes. The epithelium could play
cr
an important role in immune surveillance (34), as it can actively exclude exogenous
us
pathogens by exocytosis (35), and it can evoke a vigorous cytokine response (36, 37). Epithelial cells can secrete natural antimicrobial peptides, such as defensins, to
an
eliminate pathogens and to alert the innate and adaptive immune system to the
M
presence of foreign antigen (38, 39, 40). Epithelial cells can also directly recognize microbial PAMPs and can produce cytokines and chemokines as part of the innate
ed
immune system for initiating immune responses (36, 37, 39, 40). Nevertheless, it
secretion.
ce pt
remains unclear whether the epithelium is involved in humeral immunity via antibody
To explore the IgM expression levels in normal epithelial cells, we isolated mouse
Ac
liver epithelial cells and IECs of mice by FACS sorting to exclude the contamination of B cells. RT-PCR and FACS analysis confirmed the existence of IgM and its transcripts in these epithelial cells. Moreover, the Ig κ and Ig μ variable regions that were derived from these cells exhibited conserved usage among different individuals, which is different from that of spleen cells with regard to diversity. V IGKJ2*01
and V
IGKV9-120*01
J
IGKJ2*01
IGKV9-124*01
J
were frequently used in the Ig κ variable region that
was derived from the sorted liver epithelial cells. These rearrangements were also
Page 12 of 35
detected in lactating mammary epithelial cells and in germ cells (5, 7). The Ig κ and Ig μ variable region transcripts in the epithelial cells were germline-encoded, which is a characteristic of natural IgM.
ip t
The IgM expression in the observed epithelial cells and their natural IgM properties led us to ask whether the IgM in the epithelial cells has natural IgM activity and
cr
participates in innate immunity. Here, the results clearly showed that epithelial cells,
us
especially those frequently exposed to pathogen infections, such as liver epithelial cells and IECs, could express a large amount of natural IgM. Similar to the B1
an
cell-derived nIgM, these IgM antibodies had natural IgM properties, as they were
M
germline-encoded (6) and polyreactive. Given the large numbers of epithelial cells, we do not doubt that their secreted IgM constitutes a large part of the body’s natural
ed
IgM pathogen clearance activity. These results suggest a new mechanism by which epithelial cells are involved in innate immunity. Furthermore, our results strongly
ce pt
suggest that epithelial cells, rather than B1 cells, are the main source of the natural IgM pool in circulation or in the body fluids of normal individuals.
Ac
The pattern recognition receptors (PRR), which recognize highly conserved components of PAMPs, are crucial for the initiation of the innate immune response (41). Among these receptors, toll-like receptors (TLRs) are the key mediators of the innate immune response (42, 43). TLRs are highly conserved from Drosophila to humans, with similar structures and functions (44). TLR9-deficient mouse analysis revealed that TLR9 is a receptor for CpG DNA (45). Unmethylated CpG motifs in bacterial DNA confer its immunostimulatory activity. It is well known that B cells can
Page 13 of 35
be stimulated by TLR9 ligand, the unmethylated CpG motif, to secrete IgM (41, 46-49). In our study, IgM secretion was enhanced in mouse epithelial cells after stimulation with CpG but not with GpC or PBS. Moreover, enhanced affinity to
ip t
antigens, such as ssDNA, dsDNA, LPS, and insulin, was observed. Nevertheless, this natural IgM activity enhancement was eliminated with the knocked out of MyD88,
cr
which can regulate TLR signalling pathways as a TIR domain-containing adaptor.
us
Moreover, the use of the preferential VDJ rearrangement pattern in liver epithelial cells was not significantly changed. Additionally, in only one of the CpG treated mice,
an
the Ig μ variable region had somatic hypermutation. Therefore, the improved
M
epithelial cell-derived IgM may cause the increased polyreactivity. Additionally, the epithelial cell-derived IgM showed low affinity to different bacteria, and it could be
ed
enhanced by E. coli infection, which further expounded that epithelial cell-derived IgM is involved in innate immune responses.
ce pt
In summary, we found that epithelial cells could express and secrete natural IgM for pathogen infection defence, and TLR9 agonists, the main component of bacterial
Ac
DNA, could stimulate the secretion of nIgM to enhance the epithelial cell involvement in the innate immune response. However, the other functions of epithelial cell-derived IgM need to be further explored.
Acknowledgment We thanks Yu Zhang, Department of Immunology, Peking University Health Science Center for giving MyD88-/- mice; we also thanks Harvey Cantor for guidance and
Page 14 of 35
comments．This work was supported by grants from the National Nature Science Foundation of China (No. 81272237, No. 91229102, No. 81320108020).
The authors declare that they have no competing interests
ip t
MATERIALS AND METHODS Mice and bacterial strains
cr
Balb/c mice were purchased from Vital River Company and used at 6–8 weeks of age.
us
MyD88-/- mice were kindly provided by Dr. Yu Zhang (Peking University Health
an
Science Center). All mice were housed in a pathogen-free facility at the Peking University Health Science Center, and all animal studies were performed according to
M
the institutional and national guidelines for animal use and care. Escherichia coli (E. coli) strains DH10B (Biomed, Beijing,China), Staphylococcus
ed
aureus (S. aureus) strains Cowan I (Sigma, St. Louis, MO) and ATCC 25923,
ce pt
Staphylococcus epidermidis (S. epidermidis) clinical isolate strain (donated by Prof. Hui Wang, Peking University People’s Hospital, Beijing, China), Salmonella typhi, Paratyphoid bacilli, Proteus vulgaris, Shigella (donated by department of Pathogenic
Ac
biology, Peking University Health Science Center, Beijing, China) and Candida albicans SC5314 (donated by Prof. Ruoyu Li, Peking University First Hospital, Beijing, China) were used in this study. Isolation of mouse liver cells, IECs and B cells.30 μg mouse CpG ODN (5′-TCCATGACGTTCCTGACGTT-3′)
or
GpC
ODN
(5′-TCCATG
AGCTTCCTGAGTCT-3′) resuspended in 300 μl sterile PBS were injected intraperitoneally in Balb/c mice, each group having 3 mice. 5 days after injection, the
Page 15 of 35
mice were anesthetized, heart perfused and then sacrificed. The livers and small intestines were harvested, cut into pieces and digested with 500 U/ml collagenase Ⅳ (Sigma, St. Louis, Mo, USA) plus 150 U/ml DNase Ⅰ (Sigma, St. Louis, Mo, USA) and 60 U/ml collagenase Ⅺa (Sigma, St. Louis, Mo, USA) plus 0.02 mg/ml dispase
ip t
Ⅰ(Boehringer Mannheim, Indianapolis, Ind.), respectively. The resulting liver cells
cr
were stained with rabbit anti-mouse cytokeratin 18 (CK18, Santa Cruz, San Diego,
CA, USA) while the IECs were stained with rabbit anti-mouse CK8 (Santa Cruz, San
us
Diego, CA, USA), after which the cells were further labled with FITC-conjugated anti-rabbit IgG. The spleen cells were stained with PE-Cy™7 rat anti-mouse CD19
an
(BD Pharmingen, NJ, USA). The CK18+ liver cells, CK8+ IECs and CD19+ B cells
M
were sorted on the FACSAria™ Ⅱ Flow Cytometer (Becton Dickinson, San Diego, CA, USA). The obtained cells were further used for RT-PCR, qPCR, ELISPOT and
ed
FACS analysis.
ce pt
Immunohistochemistry. In general, tissue sections were deparaffinized, rehydrated through ethanol washes of graduated concentrations, placed in a 10-mM citrate buffer (pH 6.0), and then heated twice in a microwave oven for 5-min cycles. The slides
Ac
were then incubated with 0.3% hydrogen peroxide for 5 min, washed with PBS, and blocked in PBS plus 10% normal goat or rabbit serum for 10 min. After removing excess blocking buffer, we performed indirect immunohistochemical staining with the indicated antibodies, such as Biotin-conjugated goat anti-mouse IgM (mu-chain specific) antibody (VECTOR LAB, CA, USA). Slides were incubated for 60 min in a humidified chamber at 37C, washed thoroughly, and then incubated with the HRP-conjugated anti-goat IgG second antibodies (ZSGB-BIO, Beijing, China) at
Page 16 of 35
37C for 45 min. Slides were washed again, and bound antibodies were detected using DAB (Dako, Carpinteria, CA, USA). Control sections were stained with HRP-conjugated second antibodies alone.
ip t
RT-PCR, DNA sequence and real-time PCR. Total RNA was extracted from cell
cr
lines using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and then treated by TURBO DNase (Ambion, Austin, TX, USA) to eliminate the contamination of
us
genomic DNA. For the sorted primary cells, RNA was prepared using the RNeasy Plus Micro Kit (Qiagen, Hamburg, Germany). Reverse transcription (RT) was
an
performed with RevertAid First Strand cDNA Synthesis Kit (Fermentas, Glen Burnie,
M
MD, USA) according to the manufacturer’s instructions, and the resulting cDNA was subjected to polymerase chain reaction (PCR).
ed
To amplify the aimed genes, the primers in table 3 were used. The PCR products were separated on 1% agarose gel by electrophoresis and were visualized by ethidium
ce pt
bromide staining. For analyze VDJ orVJ usage and N-addition, the PCR products was inserted T vector, and sequence. The VDJ orVJ rearrangements was submitted to Ig blast in NCBI web set.
For real-time PCR, all the primers and reagents were purchased from Applied
Ac
Biosystems, and the reaction was run on the 7500HT Fast Real-Time PCR System (Applied Biosystem, FosterCity, CA, USA).
Flow cytometry. To analyze the expression of IgM, cells were harvested and washed twice with PBS. To detect membrane molecules, cells were blocked with 2% FBS in PBS at 4°C for 30 min and then stained with PE-conjugated goat anti-mouse IgM (Santa Cruz, CA, USA) at 4°C for 40 min. After two washes with PBS, we incubated
Page 17 of 35
the cells at 4°C for 30 min in 100 μl PBS containing FITC-conjugated goat anti-mouse IgG (Santa Cruz, CA, USA, 1:100). The cells were washed twice with PBS, and 10,000 cells were analyzed on a FACS Calibur using CellQuest software
ip t
(Becton Dickinson, San Diego, CA, USA). The background fluorescence was determined using cells incubated with the secondary antibody but without the primary
cr
antibody. To detect intracellular molecules, we harvested the cells and fixed them with
us
4% paraformaldehyde at room temperature for 30 min. After resuspending them and washing them with ice-cold PBS, we permeabilized the cells with 1 ×
an
permeabilization buffer (eBioscience, San Diego, CA, USA) twice, and each
M
centrifuging at 1,500 rpm for 5 min at 4°C. The supernatant was dropped, the appropriate primary antibodies and secondary antibodies were added, and the cells
ed
were analyzed on FACSCalibur, as described previously.
ce pt
Protein extraction and western blot analysis. To extract cytoplasmic proteins, cells were pelleted by centrifugation, lysed in RIPA lysis buffer (10 mM Tris-HCL, pH 7.2,
Ac
1% Triton X-100, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 0.15 M NaCl, protease inhibitor cocktail), and incubated on ice for 30 min. Lysates were then centrifuged at 16,000 rpm for 30 min at 4°C, and the supernatants were collected for western blot. To prepare tissue lysates, mouse tissues were minced with PBS in liquid nitrogen. After ultrasonication, the lysates were centrifuged at 16,000 rpm for 30 min at 4°C, and the supernatants were collected for further use.
Page 18 of 35
For western blot, the reduced (with ß-mercaptoethanol) or nonreduced (without ß-mercaptoethanol) protein samples were separated by SDS-PAGE and transferred onto nitrocellulose membranes (Amersham Pharmacia, UK). Membranes were
ip t
blocked in tris-buffered saline containing 0.1% Tween-20 (TBS-T) and 5% nonfat milk or 5% BSA (for detecting phospho-proteins) for 2 h at room temperature and
cr
were incubated overnight at 4°C with the appropriate primary antibodies:
us
Biotin-conjugated goat anti-mouse IgM (μ chain specific) antibody (VECTOR LAB, CA, USA), and anti-actin mAb (Sigma, St. Louis, MO, USA). Blots were washed
an
three times for 10 min each with TBS-T and incubated for 1 h at room temperature in
M
the dark with the appropriate IRDye 800-conjugated secondary antibodies or IRDye 700-conjugated secondary antibodies (LI-COR Bioscience Inc., Lincoln, NE, USA).
Lincoln, NE, USA).
ed
The signal was detected using the Odyssey Imaging System (LI-COR Bioscience Inc.,
ce pt
Prokaryotic protein expression and purification. The pGEX-4T-2 plasmid containing the sequence of VIGHV1-37*01DIGHD2-13*01JIGHJ2*01 segment and CH1-CH2
Ac
segment, whish contain 329 amino acids and its molecular weight should be 36.33KDa was constructed by EcoRⅠ and XhoⅠ. Considering the 26KDa GST, the molecular weight of fusion protein containing IgM heavy chain and GST was about 62KDa. The protein was expressed by Escherichia.coli induced by 0.1mM IPTG. The prokaryotic protein with GST-tag in lysate of E.coli was purified by GST affinity chromatography (GE healthcare). IgM ELISPOT and polyreactivity studies. To detect the secretion of IgM in mouse
Page 19 of 35
epithelial cells, ELISPOT was performed using the ELISpotPLUS for Mouse IgM Kit (MABTECH AB, Sweden). 1 × 105 sorted mouse liver cells, IECs, or B cells were used for the total IgM ELISPOT while 2 × 105 cells were used for the antigen-specific
analyze
the
polyreactivity
of
IgM,
antigen-specific
ip t
IgM ELISPOT, with 10 μg/ml of ssDNA, dsDNA, or LPS as the specific antigens. To ELISA and
indirect
cr
immunofluorescence were performed. 100 μg spleen or liver lysates prepared from
us
PBS-, GpC-, or CpG-injected mice were tested. For the ELISA analysis, microtiterplates were coated with 10 μg/ml of ssDNA, dsDNA, LPS, bacteria or 5
an
μg/ml recombinant insulin (Sigma, St. Louis, MO, USA). Biotin-conjugated goat
M
anti-mouse IgM and HRP-labled streptavidin (Santa Cruz, CA, USA) were used for the IgM detection, with TMB as the substrate. OD450 was measured using a
ed
microplate reader (Bio-Tek, Winooski, VT, USA).
Statistical analysis. All statistical calculations were performed with the statistical
ce pt
software program SAS ver. 8.1 (SAS Institute Inc., Cary, NC, USA). Differences between various groups were evaluated by the Student’s t test or chi-square test.
Ac
Differences were statistically significant when P was
S1357-2725(16)30017-6 http://dx.doi.org/doi:10.1016/j.biocel.2016.01.017 BC 4784
To appear in:
The International Journal of Biochemistry & Cell Biology
Received date: Revised date: Accepted date:
30-8-2015 1-12-2015 22-1-2016
Please cite this article as: Shao, W., Hu, F., Ma, J., Zhang, C., Liao, Q., Zhu, Z., Liu, E., and Qiu, X.,Epithelial cells are a source of natural IgM that contribute to innate immune responses, International Journal of Biochemistry and Cell Biology (2016), http://dx.doi.org/10.1016/j.biocel.2016.01.017 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Manuscript
Epithelial cells are a source of natural IgM that contribute to innate immune responses Wenwei Shao1*, Fanlei Hu1, 2*, Junfan Ma3, Chi Zhang1, Qinyuan Liao1, Zhu Zhu1,
1
ip t
Enyang Liu3, Xiaoyan Qiu1, 3† Department of Immunology, School of Basic Medical Sciences, Peking University
Department of Rheumatology and Immunology, Peking University People’s
us
2
cr
Health Science Center, Beijing, 100191, China
Hospital, Beijing, China
Key Laboratory of Medical Immunology, Ministry of Health, Beijing, 100191,
an
3
†
These authors contributed equally to this work.
Corresponding author: Xiaoyan Qiu, MD, PhD, Center for Human Disease
ed
*
M
China
Genomics, Peking University, 38 Xue-yuan Road, 100191, Beijing, China. Tel.:
Ac
ce pt
86-10-82805477; Fax: 86-10-82801149; E-mail: [email protected].
Page 1 of 35
ABSTRACT Currently, natural IgM antibodies are considered to be the constitutively secreted products of B-1 cells in mice and humans. In this study, we found that
ip t
mouse epithelial cells, including liver epithelial cells and small intestinal epithelial cells (IECs), could express IgM that also showed natural antibody
cr
activity. Moreover, similar to the B-1 cell-derived natural IgM that can be
us
upregulated by TLR9 agonists (mimicking bacterial infection), the expression of epithelial cell-derived natural IgM could also be significantly increased by TLR9
an
signalling. More importantly, the epithelial cell-derived IgM was polyreactive,
M
and it could recognize single-stranded DNA (ssDNA), double-stranded DNA (dsDNA), lipopolysaccharide (LPS), and insulin with low affinity; additionally,
ed
TLR9 agonists could enhance it in a MyD88-dependent manner. Furthermore, epithelial cell-derived IgM could bind various bacteria; therefore, it could be
ce pt
involved in anti-infection responses. Together, these results highlight the fact that epithelial cells are an important source of natural IgM, in addition to that
Ac
produced by B-1 cells, and IgM contributes to the innate immune responses in local tissues, further demonstrating that the epithelium is a first line of defense in the protection against invading microbes.
Key words: Epithelial cells • Natural IgM • TLR9 agonists • Innate immune response
Page 2 of 35
INTRODUCTION Immunoglobulin (Ig) was previously thought to be produced by only B-lineage cells. However, in the last decades, the results from a series of studies have indicated that
ip t
non-B cells can also produce Igs. As early as 20 years ago, studies by our group showed that many types of non-B cancer cells, especially epithelial cancer cells, could
cr
also express Igs, including IgG, IgA and IgM (1-7, 8, 9). Moreover, the epithelial
us
cancer cells that expressed IgG showed growth factor-like activity that could promote cancer progression (1, 2, 10). Subsequently, other researchers have also demonstrated
an
these unusual findings (11-19). Recently, growing evidence has been found to support
M
the finding that normal non-B cells, including epithelial cells, endothelial cells, neurons, germ cells, and even monocytes, can also express Igs, such as IgG, IgA and
ed
IgM (4, 5, 13, 20, 21). All of these studies indicated that the classical concept suggesting that B cells are the only source of Igs should be updated. However, the
ce pt
detailed physiological significance of non-B cell-derived Igs remains unclear. Natural antibodies are preformed antibodies that are present, even in naive germ-free
Ac
mice, in the absence of any exogenous antigenic exposure. Consistent with their specificities for microbial antigens, natural antibodies play an important non-redundant role in the first line of defence against bacterial and viral infections. Additionally, natural antibodies have also been shown to have specificities for self-antigens and have therefore been proposed to provide important homeostatic "house-keeping" functions. Thus far, B-1 cells that express CD5 have been considered as the major source of natural antibodies, including natural IgM, IgG and IgA (22, 23).
Page 3 of 35
Particularly, B-1a cells, which are enriched in the peritoneal cavity, have been considered as the only source of natural IgM. However, the B-1a cells in the peritoneal cavity have been proven to poorly contribute to serum IgM levels (24, 25),
ip t
suggesting that in addition to B-1 cell, there are other cell types involved in the production of natural IgM. Recently, many epithelial cell lineages, including breast,
cr
colon, and epidermal squamous cells, were found to spontaneously express Igs,
us
including IgG, IgA and IgM. Importantly, epithelial cell-derived Igs showed natural antibody activity. For example, the IgG and IgA that are spontaneously secreted by
an
epidermal squamous cells showed anti-bacterial activity (9). A cervical cancer cell
single-stranded
DNA
(ssDNA),
M
line (HeLa)-derived IgM displayed natural IgM activity; it could recognize double-stranded
DNA
(dsDNA)
and
ed
lipopolysaccharide (LPS) and was upregulated by a TLR9 agonist (26). By integrating these facts, we hypothesize that in addition to B-1 cells, epithelial cells are an
ce pt
important source of natural antibodies, and they might play an important role in innate immunity via participating in humeral immunity.
Ac
The IgM isotype represents one of the major Ig classes in the body. IgM is also the earliest type of membrane-associated Ig expressed during B cell ontogeny and secreted during Ag-specific immune responses. IgM can be divided into two types: circulating IgM that exists independent of known immune exposure, which are referred to as natural IgM (nIgM), and immune IgM, which is generated in response to defined antigenic stimuli. In healthy adults, circulating polyclonal IgM is generally present at 1 to 2 mg/ml in the blood. As the first line of defence against invading
Page 4 of 35
microbes or stress damage, polyclonal nIgM can directly recognize a wide range of pathogen-associated molecular patterns (PAMPs) on different microbial pathogens or damage-associated molecular patterns (DAMPs), such as ssDNA and dsDNA, which
ip t
are released by damaged or necrotic tissues. nIgM usually displays polyreactivity and triggers a rapid humeral immune response (2-7, 11).
cr
In this study, using liver and small intestine epithelial cells in adult Balb/c mice as
us
models, we further revealed IgM expression and expression of its functional transcripts in epithelial cells that were sorted by flow cytometry. Moreover, we
an
showed that IgM expression and secretion were significantly upregulated by CpG
M
ODN, a synthetic analogue of bacterial DNA. Importantly, we further identified the natural antibody properties of this IgM.
ed
RESULTS
IgM expression in the epithelial cells of Balb/c mice
ce pt
We first determined whether IgM is present in epithelial cells, including epithelial cells from the pancreas, liver, lung, kidney, stomach, uterus and small intestine of
Ac
mice. Immunohistochemistry studies showed that IgM was present in the cytoplasm of these epithelial cells (Fig. 1A). To further elucidate that IgM could be spontaneously expressed by this cell type, epithelial cells were isolated and sorted from the liver and small intestine of mice according to their cytokeratin (CK) expression profiles using flow cytometry (CK18 for liver epithelial cell, and CK8 for IECs) (Fig. 1B, D), and the IgM gene rearrangement and transcription levels were assessed. To exclude the possibility of B cell contamination, the sorted epithelial cells
Page 5 of 35
was evaluated for the absence of transcripts of CD19, CD20, and CD138. Then, specific primers were used to amplify the constant and variable regions of the Ig κ and μ chains, respectively. This analysis demonstrated the presence of rearranged Ig µ and
ip t
Ig transcripts in liver epithelial cells and IECs (Fig. 1C, E). IgM was also detected in sorted liver epithelial cells and IECs with goat anti-mouse IgM by Western blot
cr
analysis (Fig. 1F). FACS staining with PE-conjugated goat anti-mouse IgM further
us
confirmed the existence of the IgM protein (Fig. 1G).
Sequence analysis of the variable regions of the Ig µ and transcripts
an
The IgM gene transcripts and repertoires of the liver epithelial cells and IECs from six
M
Balb/c mice were detected. Spleen cell-derived IgM gene transcripts served as the positive control. The sequencing results demonstrated that 30 sequences of 32 VHDJH
ed
rearrangements and all of the 42 VκJκ rearrangements from the liver epithelial cells showed a functional and typical V(D)J rearrangement pattern. Additionally, nIgM
ce pt
from B-1 cells are often without N-region additions and are germline-encoded or with minimal somatic hypermutations. The sequence analysis revealed that although
Ac
almost all of the liver epithelial cell-derived Ig µ transcripts had N-regions, all of the Ig transcripts showed no N-region additions. Moreover, both of the Ig µ and Ig transcripts possessed germline or near germline sequences. Different from the rearrangement diversity that was observed in the spleen cells, the liver epithelial cell-derived Ig µ variable regions showed conserved usage in different individuals (Table 1). VIGHV1-37*01/D
IGHD2-13*01/JIGHJ2*01,
the predominate VDJ rearrangement
pattern in two of the five mice, was the preferential usage in the epithelial cells.
Page 6 of 35
Similarly, the rearrangement patterns of the Ig variable region in the liver epithelial cells revealed conserved usage in different individuals (Table 2). We also obtained the transcripts of the Ig variable region in the IECs, which showed similar
IGKV9-124*01J IGKJ2*01
ip t
characteristics to that in the liver epithelial cells (supplementary Table 1). V recombination occurred in both the liver epithelial cells and the
cr
IECs but not in the spleen cells.
us
A TLR9 agonist stimulated epithelial cells to secrete IgM in mice
The TLR9 signalling pathway is an important regulator for IgM secretion in B cells.
an
To determine whether IgM from epithelial cell has effects on bacterial invasion
M
defence, the epithelial cell IgM expression levels were detected when mice were stimulated with the TLR9 agonist, CpG, which mimics bacterial infection. TLR9
ed
expression in liver and small intestine epithelial cells was detected (data not shown) by IHC. Additionally, the TLR9 transcripts in these cells were further confirmed by
ce pt
RT-PCR (Fig. 2A). Subsequently, the TLR9 signalling pathway regulating IgM secretion in epithelial cells was activated carried out by the CpG treatment. As
Ac
expected, the mouse serum IgM levels were significantly elevated (Fig. 2B). Importantly, the result revealed that the IgM transcription after CpG stimulation (but not with a GpC control or PBS) was significantly upregulated in flow cytometer-sorted liver epithelial cells and IECs (Fig. 2 C, D, E, F). The Western blot results further revealed that the IgM levels in the liver and small intestine were increased after stimulation with CpG (Fig. 2G). Moreover, the CpG induction punctuated the accumulation of IgM in the extracellular matrix of the liver epithelial
Page 7 of 35
cells and in the small intestinal lumen (Fig. 2H). ELISPOT results further confirmed that the CpG treatment promoted IgM secretion of these epithelial cells (Fig. 2I). Epithelial cell-derived IgM displayed natural IgM properties
ip t
The fact that IgM was spontaneously expressed in epithelial cells guided us to explore its potential functions as a natural antibody. Natural antibodies are often polyreactive
cr
and can bind to multiple antigens with low affinity. Natural antibody activities of IgM
us
produced from several epithelial tissues of mice, including liver, pancreas, lung and small intestine, were analyzed. Using an ELISA assay, we found that, similar to the
an
B-1 cell-derived nIgM, the epithelial cell-derived IgMs also showed polyreactivity
M
compared to the non antigens-coating microwells, they could bind ssDNA, dsDNA, LPS, insulin as well as different types of microbes (Fig. 3Aa). Our results revealed that
ed
epithelial cell-derived IgM showed a similar reaction degree to the spleen-derived IgM under conditions where the total IgM levels were lower in the liver than in the
ce pt
spleen (Fig. 3Ab). Interestingly, different types of epithelial tissue-derived IgM displayed their own dominant identification profile, such as liver, lung or spleen-IgM
Ac
dominantly recognize Staphylococcus aureus (S. aureu), pancreas-IgM dominantly recognize DNA, small intestine-IgM, without obvious tendency, recognize all of antigens detected in this study. To determine whether the anti-IgM antibody tested was detecting something other than IgM, we used purified mouse IgM (Commercialization) to block the anti-IgM antibody before it was used in ELISA assay. The result showed that when anti-IgM antibodies were block with mouse IgM, the binding abilities of cell lysates from liver or spleen to different antigens were
Page 8 of 35
significantly decreased (Figure 3c). To further prove that the monoclonal IgM has polyreactivity, the recombinant Ig μ with the VIGHV1-37*01/D
IGHD2-13*01/JIGHJ2*01
rearrangement pattern, which was found to be frequently expressed by liver epithelial
ip t
cells, was expressed by Escherichia coli (E. coli) (Fig. 3B,C). As shown in Fig. 3D, the recombinant Ig μ was polyreactive to ssDNA, dsDNA and insulin, but it was not
cr
reactive to LPS.
bacterial and bacterial DNA analogue challenge
us
The polyreactivity of epithelial cell-derived IgM could be enhanced after
an
Generally, natural IgM antibodies, which are secreted in the absence of antigenic
M
challenge, are important contributors to antimicrobial immunity. In this study, whether CpG could promote secretion of the nIgM from epithelial cells was essential to
ed
determine, although the total IgM level could be upregulated by CpG, as described above. As is well known, the natural IgM activity was significantly enhanced
ce pt
following CpG exposure, particularly against ssDNA, dsDNA and LPS (Fig. 4A). Importantly, the promoting effect of CpG on the nIgM polyreactivity in liver epithelial
Ac
cells was abolished when the myeloid differentiation primary response gene 88 (MyD88) gene (the cytosolic adapter protein for TLR9) was knocked out (Fig. 4B). Additionally, we further analyzed whether there was a sequence change in the Ig µ chain variable region after CpG challenge. The results showed that the usage of the preferential VDJ rearrangement pattern in the liver epithelial cells was not significantly changed.
Page 9 of 35
The natural antibody properties of epithelial cell-derived IgM were further characterized by assessing the reactivity against various bacterial pathogens. As expected, the epithelial cell-derived IgM could recognize different bacteria with low
ip t
affinity, including E. coli, Staphylococcus aureus, Salmonella typhi, Paratyphoid bacilli, Proteus vulgaris, and Shigella. Importantly, immunization with inactivated E.
cr
coli enhanced the reactivity against all 6 pathogens (Fig. 5A-F), and this result was
us
accompanied by increased IgM expression (Fig. 5G).
DISCUSSION
an
In this study, we revealed IgM expression in normal epithelial cells that frequently
M
encounter pathogen infections in adult mice, and we observed that TLR9 agonists could stimulate these epithelial cells to secrete IgM. More importantly, the epithelial
ed
cell-derived IgM showed natural antibody properties. Our findings highlight, for the first time, a novel concept that in addition to the IgM that is produced by B1 cells,
ce pt
many epithelial cell types are important sources of natural IgMs, in addition to that produced by B1 cells, which contribute to innate immune function.
Ac
The characteristics and functions of IgM have been well studied. Many properties of IgM make this immunoglobulin particularly well-suited for its role in microbial immunity. It is present in high concentrations in the blood (in the range of 1.5 mg/ml) and is the first antibody elicited in an immune response following immunization or infection. IgM has a relatively short half-life, approximately 28 h, in the serum of normal mice in the absence of antigen (27). IgM is expressed as a membrane-bound antibody by all naive B cells, but B cells also secrete it as a pentamer (28). The
Page 10 of 35
majority of IgM is mainly generated from germline-configured transcripts. This generation occurs prior to the onset of class switch recombination (CSR) and somatic hypermutation (SHM); therefore, IgM provides a first line of defence during
ip t
microbial infections (29). Natural IgMs constitute the majority of circulating IgMs, and most natural IgM antibodies are germline-encoded and bind with low affinity to a
cr
number of different antigens (including microbial antigens), thereby contributing to
us
effective pathogen clearance by innate defense mechanisms without necessitating activation of the adaptive humeral response. Additionally, ‘innate’ IgM accelerated
an
microbial pathogen clearance by innate immune mechanisms (30, 31, 32). Although
M
B-1 cells have been long associated with natural IgM, there is no study showing that non-B-1 cells significantly contribute to serum IgM levels in naïve mice.
ed
It was previously found that not only IgG and IgA but also IgM were expressed in some non-B cells, including breast cancer cells, colon cancer cells, and cervical
ce pt
cancer cells (1, 3, 6). Moreover, similar to B cell-derived IgM, the non-B cancer cell-derived IgM showed almost no somatic hyper mutations in their V region,
Ac
namely, the germline configured transcripts. However, the γ chain-derived Ig heavy chain revealed typical somatic hyper mutation in the same cells (6). Importantly, some evidence suggested that non-B cells, especially epithelial cells, might produce natural IgM (26, 33). The epithelium has immense importance in host defense and immune surveillance, as it is the primary cell layer that initially encounters the majority of microorganisms. Originally, it was thought that the epithelium serves as only a passive barrier against
Page 11 of 35
invading pathogens. Barrier function alone is usually adequate to restrain commensal microbes; however, it is often insufficient to protect against microbial pathogens. However, recently, it has become apparent that epithelial cells are capable of
ip t
triggering an immune response that is similar to that of myeloid lineage cells, thus playing a crucial role in the active recognition of microbes. The epithelium could play
cr
an important role in immune surveillance (34), as it can actively exclude exogenous
us
pathogens by exocytosis (35), and it can evoke a vigorous cytokine response (36, 37). Epithelial cells can secrete natural antimicrobial peptides, such as defensins, to
an
eliminate pathogens and to alert the innate and adaptive immune system to the
M
presence of foreign antigen (38, 39, 40). Epithelial cells can also directly recognize microbial PAMPs and can produce cytokines and chemokines as part of the innate
ed
immune system for initiating immune responses (36, 37, 39, 40). Nevertheless, it
secretion.
ce pt
remains unclear whether the epithelium is involved in humeral immunity via antibody
To explore the IgM expression levels in normal epithelial cells, we isolated mouse
Ac
liver epithelial cells and IECs of mice by FACS sorting to exclude the contamination of B cells. RT-PCR and FACS analysis confirmed the existence of IgM and its transcripts in these epithelial cells. Moreover, the Ig κ and Ig μ variable regions that were derived from these cells exhibited conserved usage among different individuals, which is different from that of spleen cells with regard to diversity. V IGKJ2*01
and V
IGKV9-120*01
J
IGKJ2*01
IGKV9-124*01
J
were frequently used in the Ig κ variable region that
was derived from the sorted liver epithelial cells. These rearrangements were also
Page 12 of 35
detected in lactating mammary epithelial cells and in germ cells (5, 7). The Ig κ and Ig μ variable region transcripts in the epithelial cells were germline-encoded, which is a characteristic of natural IgM.
ip t
The IgM expression in the observed epithelial cells and their natural IgM properties led us to ask whether the IgM in the epithelial cells has natural IgM activity and
cr
participates in innate immunity. Here, the results clearly showed that epithelial cells,
us
especially those frequently exposed to pathogen infections, such as liver epithelial cells and IECs, could express a large amount of natural IgM. Similar to the B1
an
cell-derived nIgM, these IgM antibodies had natural IgM properties, as they were
M
germline-encoded (6) and polyreactive. Given the large numbers of epithelial cells, we do not doubt that their secreted IgM constitutes a large part of the body’s natural
ed
IgM pathogen clearance activity. These results suggest a new mechanism by which epithelial cells are involved in innate immunity. Furthermore, our results strongly
ce pt
suggest that epithelial cells, rather than B1 cells, are the main source of the natural IgM pool in circulation or in the body fluids of normal individuals.
Ac
The pattern recognition receptors (PRR), which recognize highly conserved components of PAMPs, are crucial for the initiation of the innate immune response (41). Among these receptors, toll-like receptors (TLRs) are the key mediators of the innate immune response (42, 43). TLRs are highly conserved from Drosophila to humans, with similar structures and functions (44). TLR9-deficient mouse analysis revealed that TLR9 is a receptor for CpG DNA (45). Unmethylated CpG motifs in bacterial DNA confer its immunostimulatory activity. It is well known that B cells can
Page 13 of 35
be stimulated by TLR9 ligand, the unmethylated CpG motif, to secrete IgM (41, 46-49). In our study, IgM secretion was enhanced in mouse epithelial cells after stimulation with CpG but not with GpC or PBS. Moreover, enhanced affinity to
ip t
antigens, such as ssDNA, dsDNA, LPS, and insulin, was observed. Nevertheless, this natural IgM activity enhancement was eliminated with the knocked out of MyD88,
cr
which can regulate TLR signalling pathways as a TIR domain-containing adaptor.
us
Moreover, the use of the preferential VDJ rearrangement pattern in liver epithelial cells was not significantly changed. Additionally, in only one of the CpG treated mice,
an
the Ig μ variable region had somatic hypermutation. Therefore, the improved
M
epithelial cell-derived IgM may cause the increased polyreactivity. Additionally, the epithelial cell-derived IgM showed low affinity to different bacteria, and it could be
ed
enhanced by E. coli infection, which further expounded that epithelial cell-derived IgM is involved in innate immune responses.
ce pt
In summary, we found that epithelial cells could express and secrete natural IgM for pathogen infection defence, and TLR9 agonists, the main component of bacterial
Ac
DNA, could stimulate the secretion of nIgM to enhance the epithelial cell involvement in the innate immune response. However, the other functions of epithelial cell-derived IgM need to be further explored.
Acknowledgment We thanks Yu Zhang, Department of Immunology, Peking University Health Science Center for giving MyD88-/- mice; we also thanks Harvey Cantor for guidance and
Page 14 of 35
comments．This work was supported by grants from the National Nature Science Foundation of China (No. 81272237, No. 91229102, No. 81320108020).
The authors declare that they have no competing interests
ip t
MATERIALS AND METHODS Mice and bacterial strains
cr
Balb/c mice were purchased from Vital River Company and used at 6–8 weeks of age.
us
MyD88-/- mice were kindly provided by Dr. Yu Zhang (Peking University Health
an
Science Center). All mice were housed in a pathogen-free facility at the Peking University Health Science Center, and all animal studies were performed according to
M
the institutional and national guidelines for animal use and care. Escherichia coli (E. coli) strains DH10B (Biomed, Beijing,China), Staphylococcus
ed
aureus (S. aureus) strains Cowan I (Sigma, St. Louis, MO) and ATCC 25923,
ce pt
Staphylococcus epidermidis (S. epidermidis) clinical isolate strain (donated by Prof. Hui Wang, Peking University People’s Hospital, Beijing, China), Salmonella typhi, Paratyphoid bacilli, Proteus vulgaris, Shigella (donated by department of Pathogenic
Ac
biology, Peking University Health Science Center, Beijing, China) and Candida albicans SC5314 (donated by Prof. Ruoyu Li, Peking University First Hospital, Beijing, China) were used in this study. Isolation of mouse liver cells, IECs and B cells.30 μg mouse CpG ODN (5′-TCCATGACGTTCCTGACGTT-3′)
or
GpC
ODN
(5′-TCCATG
AGCTTCCTGAGTCT-3′) resuspended in 300 μl sterile PBS were injected intraperitoneally in Balb/c mice, each group having 3 mice. 5 days after injection, the
Page 15 of 35
mice were anesthetized, heart perfused and then sacrificed. The livers and small intestines were harvested, cut into pieces and digested with 500 U/ml collagenase Ⅳ (Sigma, St. Louis, Mo, USA) plus 150 U/ml DNase Ⅰ (Sigma, St. Louis, Mo, USA) and 60 U/ml collagenase Ⅺa (Sigma, St. Louis, Mo, USA) plus 0.02 mg/ml dispase
ip t
Ⅰ(Boehringer Mannheim, Indianapolis, Ind.), respectively. The resulting liver cells
cr
were stained with rabbit anti-mouse cytokeratin 18 (CK18, Santa Cruz, San Diego,
CA, USA) while the IECs were stained with rabbit anti-mouse CK8 (Santa Cruz, San
us
Diego, CA, USA), after which the cells were further labled with FITC-conjugated anti-rabbit IgG. The spleen cells were stained with PE-Cy™7 rat anti-mouse CD19
an
(BD Pharmingen, NJ, USA). The CK18+ liver cells, CK8+ IECs and CD19+ B cells
M
were sorted on the FACSAria™ Ⅱ Flow Cytometer (Becton Dickinson, San Diego, CA, USA). The obtained cells were further used for RT-PCR, qPCR, ELISPOT and
ed
FACS analysis.
ce pt
Immunohistochemistry. In general, tissue sections were deparaffinized, rehydrated through ethanol washes of graduated concentrations, placed in a 10-mM citrate buffer (pH 6.0), and then heated twice in a microwave oven for 5-min cycles. The slides
Ac
were then incubated with 0.3% hydrogen peroxide for 5 min, washed with PBS, and blocked in PBS plus 10% normal goat or rabbit serum for 10 min. After removing excess blocking buffer, we performed indirect immunohistochemical staining with the indicated antibodies, such as Biotin-conjugated goat anti-mouse IgM (mu-chain specific) antibody (VECTOR LAB, CA, USA). Slides were incubated for 60 min in a humidified chamber at 37C, washed thoroughly, and then incubated with the HRP-conjugated anti-goat IgG second antibodies (ZSGB-BIO, Beijing, China) at
Page 16 of 35
37C for 45 min. Slides were washed again, and bound antibodies were detected using DAB (Dako, Carpinteria, CA, USA). Control sections were stained with HRP-conjugated second antibodies alone.
ip t
RT-PCR, DNA sequence and real-time PCR. Total RNA was extracted from cell
cr
lines using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and then treated by TURBO DNase (Ambion, Austin, TX, USA) to eliminate the contamination of
us
genomic DNA. For the sorted primary cells, RNA was prepared using the RNeasy Plus Micro Kit (Qiagen, Hamburg, Germany). Reverse transcription (RT) was
an
performed with RevertAid First Strand cDNA Synthesis Kit (Fermentas, Glen Burnie,
M
MD, USA) according to the manufacturer’s instructions, and the resulting cDNA was subjected to polymerase chain reaction (PCR).
ed
To amplify the aimed genes, the primers in table 3 were used. The PCR products were separated on 1% agarose gel by electrophoresis and were visualized by ethidium
ce pt
bromide staining. For analyze VDJ orVJ usage and N-addition, the PCR products was inserted T vector, and sequence. The VDJ orVJ rearrangements was submitted to Ig blast in NCBI web set.
For real-time PCR, all the primers and reagents were purchased from Applied
Ac
Biosystems, and the reaction was run on the 7500HT Fast Real-Time PCR System (Applied Biosystem, FosterCity, CA, USA).
Flow cytometry. To analyze the expression of IgM, cells were harvested and washed twice with PBS. To detect membrane molecules, cells were blocked with 2% FBS in PBS at 4°C for 30 min and then stained with PE-conjugated goat anti-mouse IgM (Santa Cruz, CA, USA) at 4°C for 40 min. After two washes with PBS, we incubated
Page 17 of 35
the cells at 4°C for 30 min in 100 μl PBS containing FITC-conjugated goat anti-mouse IgG (Santa Cruz, CA, USA, 1:100). The cells were washed twice with PBS, and 10,000 cells were analyzed on a FACS Calibur using CellQuest software
ip t
(Becton Dickinson, San Diego, CA, USA). The background fluorescence was determined using cells incubated with the secondary antibody but without the primary
cr
antibody. To detect intracellular molecules, we harvested the cells and fixed them with
us
4% paraformaldehyde at room temperature for 30 min. After resuspending them and washing them with ice-cold PBS, we permeabilized the cells with 1 ×
an
permeabilization buffer (eBioscience, San Diego, CA, USA) twice, and each
M
centrifuging at 1,500 rpm for 5 min at 4°C. The supernatant was dropped, the appropriate primary antibodies and secondary antibodies were added, and the cells
ed
were analyzed on FACSCalibur, as described previously.
ce pt
Protein extraction and western blot analysis. To extract cytoplasmic proteins, cells were pelleted by centrifugation, lysed in RIPA lysis buffer (10 mM Tris-HCL, pH 7.2,
Ac
1% Triton X-100, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 0.15 M NaCl, protease inhibitor cocktail), and incubated on ice for 30 min. Lysates were then centrifuged at 16,000 rpm for 30 min at 4°C, and the supernatants were collected for western blot. To prepare tissue lysates, mouse tissues were minced with PBS in liquid nitrogen. After ultrasonication, the lysates were centrifuged at 16,000 rpm for 30 min at 4°C, and the supernatants were collected for further use.
Page 18 of 35
For western blot, the reduced (with ß-mercaptoethanol) or nonreduced (without ß-mercaptoethanol) protein samples were separated by SDS-PAGE and transferred onto nitrocellulose membranes (Amersham Pharmacia, UK). Membranes were
ip t
blocked in tris-buffered saline containing 0.1% Tween-20 (TBS-T) and 5% nonfat milk or 5% BSA (for detecting phospho-proteins) for 2 h at room temperature and
cr
were incubated overnight at 4°C with the appropriate primary antibodies:
us
Biotin-conjugated goat anti-mouse IgM (μ chain specific) antibody (VECTOR LAB, CA, USA), and anti-actin mAb (Sigma, St. Louis, MO, USA). Blots were washed
an
three times for 10 min each with TBS-T and incubated for 1 h at room temperature in
M
the dark with the appropriate IRDye 800-conjugated secondary antibodies or IRDye 700-conjugated secondary antibodies (LI-COR Bioscience Inc., Lincoln, NE, USA).
Lincoln, NE, USA).
ed
The signal was detected using the Odyssey Imaging System (LI-COR Bioscience Inc.,
ce pt
Prokaryotic protein expression and purification. The pGEX-4T-2 plasmid containing the sequence of VIGHV1-37*01DIGHD2-13*01JIGHJ2*01 segment and CH1-CH2
Ac
segment, whish contain 329 amino acids and its molecular weight should be 36.33KDa was constructed by EcoRⅠ and XhoⅠ. Considering the 26KDa GST, the molecular weight of fusion protein containing IgM heavy chain and GST was about 62KDa. The protein was expressed by Escherichia.coli induced by 0.1mM IPTG. The prokaryotic protein with GST-tag in lysate of E.coli was purified by GST affinity chromatography (GE healthcare). IgM ELISPOT and polyreactivity studies. To detect the secretion of IgM in mouse
Page 19 of 35
epithelial cells, ELISPOT was performed using the ELISpotPLUS for Mouse IgM Kit (MABTECH AB, Sweden). 1 × 105 sorted mouse liver cells, IECs, or B cells were used for the total IgM ELISPOT while 2 × 105 cells were used for the antigen-specific
analyze
the
polyreactivity
of
IgM,
antigen-specific
ip t
IgM ELISPOT, with 10 μg/ml of ssDNA, dsDNA, or LPS as the specific antigens. To ELISA and
indirect
cr
immunofluorescence were performed. 100 μg spleen or liver lysates prepared from
us
PBS-, GpC-, or CpG-injected mice were tested. For the ELISA analysis, microtiterplates were coated with 10 μg/ml of ssDNA, dsDNA, LPS, bacteria or 5
an
μg/ml recombinant insulin (Sigma, St. Louis, MO, USA). Biotin-conjugated goat
M
anti-mouse IgM and HRP-labled streptavidin (Santa Cruz, CA, USA) were used for the IgM detection, with TMB as the substrate. OD450 was measured using a
ed
microplate reader (Bio-Tek, Winooski, VT, USA).
Statistical analysis. All statistical calculations were performed with the statistical
ce pt
software program SAS ver. 8.1 (SAS Institute Inc., Cary, NC, USA). Differences between various groups were evaluated by the Student’s t test or chi-square test.
Ac
Differences were statistically significant when P was
