Gut bacteria induce IgA expression in pituitary hormone-secreting cells during aging

Summary Pituitary hormone decline is a hallmark of aging. However, the precise gene regulation mechanism during pituitary aging is unclear. Here, we characterized the cell population alteration and global transcriptional change during pituitary aging through single-cell RNA sequencing (scRNA-seq). We found that mRNA-encoding components of protein translational machinery declined the most in the pituitary during aging. Remarkably, Immunoglobulin A (IgA) was found to be expressed in hormone-secreting cells, and the IgA expression level increased dramatically in aged pituitary. Moreover, the pituitary IgA expression was regulated by gut microbiota. The non-hematopoietic origin of the IgA+ cells in the pituitary was further confirmed through bone marrow transplantation. Somatotropes were identified as the most prominent IgA-producing cells through lineage tracing. Thus, pituitary hormone-secreting cells can generate IgA in an age-dependent manner, and such a process is influenced by gut bacteria.


INTRODUCTION
2][3] The anterior pituitary is the major hormone secretion gland that consists of five types of endocrine cells: growth hormone (GH) secretion somatotropes, adrenocorticotropin (ACTH) secretion corticotropes, thyroid-stimulating hormone (TSH) secretion thyrotropes, sex hormone secretion gonadotropes, and prolactin (PRL) secretion lactotropes. 4,5Besides hormone-secreting cells, other supporting cells, including mesenchymal cells, endothelial cells, and macrophages are important components of the normal pituitary. 6,7ituitary hormone alteration, such as GH decline, is a hallmark of mammalian aging. 8,9Although many studies address the cause and effect of pituitary hormone changes during aging, especially the GH-IGF-1 (insulin-like growth factor 1) axis, a global transcriptional picture showing the different types of cell alteration and gene expression changes at individual cell resolution during the pituitary aging process is still missing.
IgA is a unique type of antibody secreted to the mucosal surface, and its dimerization and secretion are facilitated by covalent attachment to the J chain. 10,113][14] In addition, IgA provides exquisite control of the microbiota within the intestine by affecting bacterial colonization and gene expression.Concurrently, microbiome signals also stimulate IgA production. 15,16Besides functions on epithelial surfaces, IgA is also found to regulate neuro-inflammation and defend against infections in the meninges. 17,18IgA has been known to be produced by plasma cells, the terminally differentiated B cells.
In this study, we provided a comprehensive scope of cell populations and gene expression changes during pituitary aging with individual cell resolution.The results showed that translational dysfunction, but not transcriptional decline, might be the cause of hormonal alteration during pituitary aging.Surprisingly, this study also revealed IgA generation in the hormone-secreting cells during aging of the pituitary, and such a process is influenced by the gut microbiome.

RESULT Messenger RNAs-encoding protein translational machinery declined in pituitary hormone-secreting cells during aging
To study cell population alteration and better characterize gene regulation during pituitary aging, we performed single-cell RNA sequencing (scRNA-seq) of cells isolated from the pituitary of young (2-3 months old) and old (18-24 months old) C57/B6 mice.0][21] Four young or old pituitaries were mixed together and then the live cells were sorted and collected by fluorescence-activated cell sorting (FACS).The scRNA-seq was performed on the sorted cells using the 10x Genomics Chromium platform.The analysis of 16,214 (9,407 from young and 6,807 from old) high-quality single-cell transcriptomes with uniform manifold approximation and projection (UMAP) revealed 7 distinct cell types (Figures 1A and S1A).Characterization of markers in these clusters identified them as several types of hormone-secreting cells (Hsc-1 to 4), endothelial cells (Endo), mesenchymal cells (Mes), and macrophages (Mac) (Figures S1B and S1C).When comparing the cell populations in young and old pituitary, we observed a shift in the largest population of hormone-secreting cells Hsc-1, while the other cell populations remained similar (Figures S2A  and S2B).
To define the specific change of Hsc-1 during aging, we compared the gene expression profiles of Hsc-1 in young and old pituitaries.Pathway analysis of differentially expressed genes showed that the genes functioning in ribosome-mediated cytoplasmic translation showed the most drastic decline in expression during aging (Figure S2C).Among the top 10 downregulated pathways in aged Hsc-1, 8 were ribosomal subunits, indicating that protein translation decline in hormone secretion cells was the most significant change during pituitary aging.This phenomenon is consistent with the previous observation that pituitary hormones, products of the protein translational machinery, decrease in aged mice. 8,9Among the top 10 upregulated pathways, 4 were transcription related, which might constitute a compensatory response to the translational decline in Hsc-1 cell populations in the aged pituitary.

Increase of B cell-independent IgA expression in aged pituitary
While investigating the gene expression change of Hsc-1 cells in aged pituitary, we noticed that Igha (IgA heavy chain), Igkc (kappa light chain), and Jchain, the three genes contributing to the assembly of dimeric IgA, were among the top upregulated genes (Figure S2D).Since Hsc-1 clustered as pituitary hormone-secreting cells, this result was surprising.Consistent with Igha, Igkc, and Jchain increase in Hsc-1, the number of cells expressing these three genes increased dramatically in the aged pituitary.(Figure 1B).The expression of B cell marker genes Cd19 and Ms4a1 (Cd20) was mostly absent in the pituitary regardless of age, suggesting the non-B cell origin of the IgA+ cells in the old pituitary (Figures 1C and 1D).Further investigation of single-cell transcriptomic data showed that Igkc, Igha, and Jchain were indeed expressed in some of the hormone-secreting cells (Figures 1E-1G).
To confirm that it was an age-dependent generation of IgA, we did transcription profiling of young and old pituitary.Circulating immune cells in blood were removed from the mice by perfusion, and mRNA was extracted from the whole pituitary for next-generation sequencing.The Igkc and Igha mRNA levels were significantly increased in the old pituitary, while expressions of genes encoding other antibody subtypes were largely unchanged (Figure 2A).Additionally, qRT-PCR and in situ hybridization results further confirmed the upregulation of Igkc and Igha mRNA in old pituitary (Figures 2B and S3A).At the protein level, immunoblotting of perfused pituitary cell lysates showed a pronounced increase of immunoglobulin k chain and a chain, products of Igkc and Igha, in old pituitary (Figure 2C).The immunohistochemistry (IHC) staining using an antibody against IgA heavy chain also showed positive signals in the pituitaries of old mice (Figure 2D).To better quantify the IgA+ cells, pituitaries from young or old mice were dissociated to single cells and stained with Phycoerythrin (PE)-conjugated anti-IgA antibody after fixation.The samples were then subjected to FACS analysis to calculate the percentage of IgA+ cells in the whole pituitary.The percentage of IgA+ in old pituitary increased significantly, although the overall percentage was low (Figures 2E and S3B).
In addition to the pituitary, we also checked the mRNA expression of Igha in other organs.Compared to young animals, the Igha mRNA level also increased serval fold in the liver and kidney of old mice, but the changes were much lower than that in the pituitary (>30-fold).The Igha mRNA levels deceased in old spleen and were unchanged in the intestine, lung, and brain (Figure S3C).These results indicated that IgA elevation in the pituitary was not a result of global immune system change during aging.

Clonal heterogeneity of the IgA+ cells in old pituitary
Antibody specificity is determined by its variable region (V-region).The V-region-encoding genes are primarily assembled from the germline V, D, and J gene segments by V(D)J recombination.To check the clonal characteristics of IgA+ cells in old pituitary, we analyzed the RNA-seq data and found the repertoire of V segments utilized in immunoglobulin heavy chain and light chain is comparable between the young and old pituitaries, indicating a preselected nature of the pituitary IgA (Table S1).To further analyze the IgA repertoire in the old pituitary, the RNA was isolated from perfused pituitary and transcribed into cDNA with primers specific for the constant region of Igha (Ca), and BCR sequences were amplified by PCR with primers annealing to FR1 (framework region 1) and the Ca domain.The amplicons were sequenced, and CDR3 diversity was analyzed.Among the top 50 CDR3 of each sample, the frequency of the top 5 was at an equal level, and there was no one dominant CDR3 sequence (Figure 2F).We also used paired primers specifically targeting different V chains and Ca to amplify BCR with different V chains.Consistently, both the distal and proximal V chains in the genome are expressed higher in the old pituitary (Figure 2G).Those data collectively suggested the clonal heterogeneity characteristics of IgA+ cells in aged pituitary.

Pituitary IgA level is positively correlated with gut-bacteria diversity
In addition to age-related IgA elevation, we also found that pituitary IgA levels varied among different mice, with one particular group expressing the highest level at 3 months old was thus labeled as IgA-high mouse (IgA-high).The mRNA levels of Igkc and Igha in the pituitary of IgA-high mice were higher than in other mice (Figure 3A).Immunoglobulin k chain and a chain proteins measured by Western blotting also increased markedly in the 3-month-old IgA-high mice (Figure 3B).To quantify the IgA+ cells in the IgA-high group, pituitary cells from the 3-month-old normal and IgA-high mice were stained with PE-conjugated anti-IgA antibody and analyzed by FACS.As shown in Figures 3C and S4A, the IgA+ cell population increased about 3-fold in IgA-high mice.Similar to the age dependency of normal mice, the gene expression level of Igha and Igkc in the IgA-high mice also increased as the mice grew older (Figure 3D).
Since IgA-high mice were not genetically different from normal mice and IgA expression was reported to be affected by gut microbiota, 15,[22][23][24] we characterized the gut-bacterial microbiome through 16S rRNA sequencing in young IgA-high mice compared with that of normal mice in both young and old age.The ileocecal microbiome was chosen as it was known to encompass much of the diversity within the gastrointestinal tract. 25The microbiome community structure of normal young and old mice or IgA-high young mice was assessed using principal coordinate analysis (PCoA) based on unweighted and weighted UniFrac distances.As shown in Figures S4B and S4C, the gut microbiome of normal young mice clustered together, while that of the normal old mice and IgA-high young mice overlapped.This result indicated the similarity of gut microbiota of normal old mice and IgA-high young mice, which differed from that of normal young mice.Further analysis showed bacterial diversity was lower in normal young mice compared with either normal old or IgA-high young mice (Figure 3E).To test whether gut-bacteria diversity controlled the expression of IgA in the pituitary, we orally treated IgA-high young mice with antibiotics for 2 months to reduce the gut-bacteria diversity (Figure S4D).As shown in Figures 4A and 4B, antibiotics treatment inhibited mRNA and protein expression of Igkc and Igha in the pituitary of those mice.FACS analysis of IgA+ cells in the pituitary of antibiotics treated group showed that IgA+ cells decreased to one-fifth of the untreated group (Figure 4C).These results confirmed that the pituitary IgA level was regulated by the diversity of gut bacteria.
To better characterize the relationship between gut bacteria and IgA expression in the pituitary, we transferred microbiomes from the normal young and old as well as IgA-high young mice to the young germ-free (GF) mice.The Igkc and Igha expression in GF mice pituitary were negligible compared with the specific pathogen-free (SPF) mice (Figures 4D and 4E).However, after we transferred the gut microbiome from different mice to the GF mice, the Igkc and Igha expressions were restored.Furthermore, the GF mice receiving microbiome from the normal old or IgA-high young mice expressed higher levels of IgA in their pituitaries than those that received microbiome from normal young mice (Figure 4F).Collectively, our results indicated that the diversity of gut bacteria controlled the IgA expression in the pituitary.

The hormone-secreting cells that express IgA
Our scRNA-Seq data indicated that the IgA+ cells in the pituitary were from hormone-secreting cells but not conventional B cells.To validate this finding and rule out the possibility that those IgA+ cells were the result of B cell contamination, we conducted a bone marrow transfer experiment using IgA knockout mice with Igha constant region exon2 deleted as doner (Figure S5A).The recipients were IgA-high young mice irradiated with a lethal dose of gamma rays and transferred with the bone marrow cells from wild-type or IgA knockout mice.Four weeks after transfer, the mRNA level of Igha in the recipient mice spleens and pituitaries was measured by qRT-PCR.The spleen Igha mRNA level in Igha knockout bone marrow recipients was much lower than that received wild-type bone marrow cells (Figure S5B).However, no difference in Igha mRNA levels was observed between mice that received wild-type or Igha knockout bone marrow cells, and the levels were comparable to that of non-irradiated mice (Figure 5A), indicating that IgA was generated in non-hematopoietic cells.
We also sorted Epcam+ hormone-secreting cells of normal young and old mice pituitaries to further examine the Igkc and Igha expression.The Epcam+ cells in the old pituitary expressed significantly higher levels of Igkc and Igha than the cells from the young pituitary (Figure S5C).These data further supported the notion that IgA in the old pituitary was generated by hormone-secreting cells rather than circulating B cells.
Since the finding that IgA was produced in the hormone-secreting cells from the pituitary was highly surprising, we tried to further verify this finding by generating Igha-cre mice.After crossing Igha-Cre mice with the Ai6 reporter mice, all cells that accomplished V(D)J recombination and class-switching to IgA would drive Cre expression and delete the stop codon before ZsGreen gene to induce its expression.Therefore, the cells that had expressed IgA would be labeled by ZsGreen.To obtain ZsGreen signal in the mouse pituitary, Igha-cre; Ai6 male mice were crossed with IgA-high female mice, and the male mice were removed one week later.This way, the maternal microbiota from IgA-high mice would be transferred to the offspring naturally and make the Igha-cre; Ai6 mice with higher IgA expression. 26,27In the intestine of those mice, ZsGreen signal was specifically detected in payer's patch and lamina propria, places where IgA plasma cells reside (Figure S5D).In the spleen, most ZsGreen positive cells were also positive for B cell marker Cd19 (Figure 5C), indicating successful Igha-dependent generation of ZsGreen signal.The ZsGreen signal was detected in the pituitary of those mice as well (Figure 5B).However, the ZsGreen-expressing cells in the pituitary were CD19 negative, a feature that was different from that in the spleen (Figure 5C).ZsGreen-positive cells from the pituitaries of Igha-Cre; Ai6 mice were then sorted and subjected to scRNA-seq.We obtained 1,339 single-cell transcriptomes and divided them into 9 distinct cell types by UMAP analysis (Figure 5D).Marker genes identified them as different types of hormone-secreting cells (cluster 0, 1, 2, 3, 6, and 8), macrophages (cluster 4), mesenchymal cells (cluster 5), and a unique group that expressed the highest level of IgA (cluster 7) (Figure S5E).B cell markers Cd19 and Ms4a1 were not detected in most Igha-Cre-traced pituitary cells (Figure 5E).Different from cluster 7 in which Igha and Igkc were highly expressed, other cell clusters express IgA at low levels (Figure S5F).Further analysis of the Igha-Cre traced hormone-secreting cells revealed most cells (cluster 0, 1, 2, and 3) were Pou1f1 positive (Figure 5F).The top 3 clusters (cluster 0-2) with the largest number of cells were somatotropes (cluster 0, 30%), thyrotropes (cluster 1, 24%), and lactotropes (cluster 2, 18%), respectively (Figure 5G).Consistently, IgA expression could be observed in Pit1-Cre traced hormone-producing cells from the pituitary of 20M old mice (Figure S5G).These data collectively supported that different hormone-secreting cells in pituitary expressed IgA and somatotropes were the most prominent IgA-producing cells.
Using those Igha-Cre; Ai6 mice, we further validated the relationship between gut-bacteria diversity and IgA+ cells in the pituitary.The Igha-Cre; Ai6 mice were treated with antibiotics to reduce gut-bacteria diversity, and ZsGreen-positive cells in the pituitary were quantified.Consistently, ZsGreen-positive cells in pituitary decreased dramatically after antibiotics treatment (Figure 5H).We also tested the contribution of bone marrow cells to ZsGreen-positive cells in the pituitary.Igha-Cre; Ai6 reporter mice were irradiated with a lethal dose of gamma ray and transplanted with bone marrow cells from Ai6 or Igha-Cre; Ai6 mice.The ZsGreen-positive cells in pituitaries were quantified one month later.The FACS data showed that the ZsGreen-positive cells in recipient mice were at equal levels regardless of the genotype of the donor (Figure 5I).Thus, IgA in the pituitary was again confirmed to originate from non-hematopoietic cells.

IgA in pituitary prevents virus infection
Lastly, we tried to figure out the function of IgA in the pituitary.Previous report shows IgA plays an important role in trapping microbes and protecting the body from pathogens challenge. 14,28,29Given the pituitary is closely related to the circulation system, it may easily be attacked by pathogens in blood and IgA may play some protective roles.To test this hypothesis, we infected the 3-4 months old control and IgA-high mice with Adeno-associated viruses (AAV) expressing EGFP by tail vein injection.Two weeks after virus infection, EGFP expression could be diminished function of immune system in old mice would promote virus infection, we did not compare the AAV infection in young and old pituitary.The functional test suggested IgA in the pituitary prevents it from virus attack, and it might be a compensation for immune aging.

DISCUSSION
The current study revealed that decreasing mRNAs-encoding protein translational machinery was one of the most significant changes in aged pituitary hormone-secreting cells.This phenomenon could explain the pituitary hormone decline during aging.Consistent with our data, a recent study reported that increased ribosome pausing contributes to proteostasis impairment during aging. 30urprisingly, scRNA study of young and old mouse pituitary revealed IgA expression in the hormone-secreting cells in aged pituitary.This finding is remarkable since plasma cells are the only known source of IgA antibodies. 31,32][35] We therefore tried several experimental approaches aiming to eliminate the possibility of B cell contamination in the old pituitaries: (1) identification of Igha, Igkc, and Jchain expression by single-cell sequencing in a B cell-free condition; (2) bone marrow transplant from IgA knockout mice; (3) scRNA sequencing of Igha-cre; Ai6 traced IgA-producing cells from pituitaries further validated their identities as hormone-secreting cells and revealed somatotropes were the most prominent IgA-producing cells.
The finding of gut-bacteria diversity regulated IgA production in the pituitary was consistent with previous findings that gut bacteria influence IgA plasma cells in the meninges. 18However, the meninges-residing IgA plasma cells were proposed to be educated in the gut and circulated to the meninges after maturation, which was different from our observations that circulating B cells did not affect IgA expression in aged pituitary.It remains unclear how the signals transmitted from gut-bacteria affect IgA gene expression in the pituitary.One possibility may be a specific or combination of metabolites generated by gut bacteria that could be sensed by pituitary cells.How those metabolites are sensed and initiate the activation of IgA gene recombination and expression will be an interesting future research direction.Alternative hypotheses, such as communication of gut bacteria on the nerve of intestines somehow affecting the pituitary, also exist.Nevertheless, the generation of IgA signal in pituitary could be used as a reporting system to study gut-pituitary communication in the future.
Since IgA expressed in pituitary is not at a high level, we are curious about the reason why it is essential, especially during aging.The virusdefending effect of IgA in the pituitary may give us some indication in explaining this puzzle.The central nervous system (CNS) is protected by the blood-brain barrier (BBB). 36,37However, the pituitary gland is directly connected to the hypothalamus through a stalk of blood vessels and nerves (the pituitary stalk), where there is a BBB leakage.Therefore, pathogenic viruses and bacteria in the circulation system may attack the CNS through this hypothalamus-pituitary interaction site.IgA expressed in the pituitary may prevent the pathogens from infecting the CNS system through this hypothalamus-pituitary connection.The protecting role of the IgA in the pituitary may become more critical in old organisms, in which gut bacteria diversity increases while humoral responses decline.Igha gene through homologous recombination.The gRNAs used for Igha-Cre mice generation were 5'-gggccacaggctggcacctg-3' and 5'-ggtgaggaaggtcacagtggt-3'.For Pit1-Cre (Pou1f1-Cre) mice, P2A tag and Cre was added to the C terminal of Pou1f1 gene through homologous recombination.The gRNAs used for Pit1-Cre mice generation were 5'-ggcgtgtgatagccatgtg-3' and 5'-ggtgttgtgagggactttcc-3'.The genotyping primer for Igha KO mice were Igha KO-F: 5'-GAAGGTGTTCACTGGAAGGTAG-3'; and Igha KO-R: 5'-GGATGTGAAGTGC TCTGGTAAA-3', the wildtype allele PCR product is 905 bp and KO product is 220 bp.The genotyping primer for Igha-Cre and Pit1-Cre mice were: Cre-F: 5'-TTGCCAGGATCAGGGTTAAAG-3'; Cre-R: 5'-CTTGCATGATCTCCGGTATTGA-3'; Igha-WT-F: 5'-GTCAAGAGCCAC TTTCCTATGT-3'; Igha-WT-R: 5'-CCTGGAGTTGGGTTATGTCTTC-3'.The PCR product of Cre is 405 bp and PCR product of Igha-WT is 387 bp.Igha-cre mouse is Cre positive and there is no band of Igha-WT band in Igha-Cre homozygous mouse.

METHOD DETAILS Pituitary gland dispersion
0][21] Briefly, pentobarbital (50mg/kg, i.p.) anesthetized mice were perfused via the aorta with cold PBS and dissected pituitary was placed into an enzyme mix containing 0.5% w/v collagenase type 2 (Worthington, LS004176), 0.25% trypsin (Thermo Fisher Scientific, 25200056), in Hanks balanced salt solution (Gibco, 14175095) for 2-3 hours at 37 C.The tissue and cells were gently triturated and pelleted at 900g for 3 minutes, and resuspended for further single cell sorting.For the young and old pituitary sample, all live cells from young (2-3M) or old (18-24M) pituitary negative for DAPI (Thermo Fisher Scientific, D1306) staining were sorted for further scRNA-seq.For the Epcam+ cells sample, old (18-24M) pituitary cells were stained with DAPI and FITC conjugated Epcam antibody (Invitrogen, 11-5791-82), then DAPI negative and Epcam positive cells were sorted for further processing.For the Igha-cre; Ai6 sample, zsGreen positive and DAPI negative pituitary cells were sorted for sequencing.

Single cell library preparation
Single-cell RNA library construction and sequencing Single-cell cDNA libraries were prepared using the Chromium Single Cell 3 0 Library and Gel Bead Kit v2 according to the manufacturer's instructions.In brief, cell suspensions in a chip were loaded on a Chromium Controller (10x Genomics) to generate single-cell GEMs (gel beads in emulsion).scRNA-seq libraries were then prepared using the Chromium Single Cell 3 0 Gel Bead and Library Kit (P/N 120236, 120237, 120262; 10x Genomics).Qualitative analysis of each DNA library was performed using an Agilent 2100 Bioanalyzer.The concentration of each DNA library was measured using a Qubit fluorometer (Invitrogen).Libraries were sequenced on an Illumina NextSeq 500 (23150 paired-end reads).
Single-cell RNA sequencing (scRNA-seq) analysis For young and old pituitary, the 10x Cell Ranger (V3.0.2, 10x Genomics) software was used to demultiplex the sequencing data, to map reads to the mus musculus genome assembly mm10 (version 3.0.0,10x Genomics), to generate feature-barcode matrices and quality control metrics.Then all the feature-barcode matrices were imported into SoupX (v1.6.2) 38 to removes ambient RNA contamination and then were imported into Scanpy (v1.7.1) 39 to generate AnnData objects.Cells with detected genes < 500 or >4,000 were filtered.Cells with UMI counts > 40,000 or mitochondrial UMI ratio > 10% were also excluded.Cell doublets were removed from the objects using Scrublet (version 0.2.1) 40 with the doublet score threshold=0.32.Genes expressed in less than 7 cells were filtered.The objects were concatenated into a single object and genes expression were normalized to 10,000 reads per cell and logarithmized.Then highly variable genes were selected with max_mean=5 and min_disp=0.6.Effects of total counts per cell and the percentage of mitochondrial genes expressed were regressed out using pp.regress_out() function and scaled.Dimensionality reduction was performed by running principal component analysis and then a nearest-neighbor graph was calculated using 'pp.neighbors()',followed by clustering using leiden algorithm 'tl.leiden()' with a resolution adopting from prior knowledge on pituitary cells population composition.Then uniform manifold approximation and projection (UMAP) dimension reduction was performed for visualizing high-dimensional data.Scanpy rank_genes_groups() and Wilcoxon rank-sum test were used to compute the statistical significance of differential genes expression between any groups and identify marker genes of each cluster.For Igha-cre; Ai6 sample, ambient RNA contamination removal with SoupX was not applied, and cells with detected genes < 200 or >8,000 were filtered.Cells with UMI counts > 80,000 or mitochondrial UMI ratio > 10% were also excluded.The top 3,000 highly variable genes were selected.

Gene set enrichment analysis
Gene Set Enrichment Analysis (GSEA) 41 was conducted using GSEA software (Version 4.1.0)to identify significantly regulated pathways.The significantly differentially expressed genes (p-value <0.1) were ranked by their log-transformed fold change value.The pre-ranked gene list and MousePath_All_gmt-Format.gmt (http://ge-lab.org/gskb/)were used for running the tool ''Run GSEA Preranked'' in the classic mode.

RNA-sequencing and data processing
The isolated pituitary from young (2-3M) or old (18-24M) mouse was directly lysate in lysis buffer and RNA was extracted using RNAprep pure Cell/Bacteria Kit (TIANGEN, DP430) following the manufacturer's instructions.Concentration of RNA was measured and the sample passed quality checks was used for RNA sequencing library preparation at NIBS Center for Nucleic Acid Sequencing following manufacturer's instruction (Illumina).HiSeq 2500 was used for the library sequencing.Fastq files for the sequencing data was assessed by FastQC (v0.11.8) to ensure the good quality and then were mapped with STAR aligner (version 2.6.1a) to the GRCm38 mouse reference genome with default parameters.The mapped sequencing reads were assigned to Ensembl gene annotation with featureCounts function of Rsubread package (version 2.0.1) with parameters specific to paired end reads.Gene expression was quantified by the reads that fall into exon and normalized to FPKM (fragments per kilobase of exon model per million mapped reads) by edgeR (version 3.28.1).Differential expression analysis was also executed by edgeR.Specifically, the counts were normalized using TMM (trimmed mean of M-values) normalization method, then the common dispersion and tagwise dispersions were estimated, and genewise quasi-likelihood F-tests using GLMs in edgeR were performed to calculate p-values for fold changes.

16S rRNA gene sequencing
To harvest viable mucosal-associated and luminal microbial communities from the cecum of indicated mice, animals were sacrificed and abdomen was opened under sterile conditions.The cecum was opened and manually extruded along the cephalocaudal axis to collect the fecal material into a clean tube for 16S rRNA extraction.16S rRNA gene sequencing methods were adapted from the methods developed for the Earth Microbiome Project (Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample; Ultrahigh-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms).Briefly, bacterial genomic DNA was extracted using DNA Isolation Kit (TIANGEN, DP328-02).The 16S rDNA V3-V4 region was amplified by PCR and sequenced in the MiSeq platform (Illumina) using the 2x250 bp paired-end protocol yielding pair-end reads.The primers used fornamplification (515F-806R) contain adapters for MiSeq sequencing and single-end barcodes allowing pooling and direct sequencing of PCR products on the Illumina MiSeq platform. 42

16S rRNA gene compositional analysis
Adapters and primers were removed from the raw sequencing data using trimmoatic to get clean data.The paired-end sequence reads were imported into a QIIME2 (version 2020.6)artifact, and then length trimming, denoising, chimera and PhiX removal were implemented with dada2 plugin with default parameters.The resulting feature table were summarized for further exploration and the resulting feature sequences were used to generate a tree using q2-phylogeny plugin for phylogenetic diversity analyses.The alpha diversity analyses were performed using q2-diversity plugin to compute several alpha and beta diversity metrics, and generate principal coordinates analysis (PCoA) plots using Emperor for each of the beta diversity metrics.Finally, Naı ¨ve Bayes classifier using Silva v132 reference taxonomy files were trained to classify the feature sequences using q2-feature-classifier plugin.

CDR3 Sequence analysis
Sequences were assigned to individual samples according to their MID.Sequences lacking one or both primer sequences or shorter than 320 bp were excluded.Sequences were further analyzed with MiXCR (version 3.0.13),a universal tool for fast and accurate analysis of T-and B-cell receptor repertoire sequencing data.The whole analysis was performed using analyze amplicon command.Specifically, paired-end reads were merged and aligned to reference V/D/J and C genes, then clone-types were built and assembled from the alignments and exported to a human readable text file.The top 50 CDR3 were selected for generating the pie chart according to their percentages.

BCR amplification using nested PCR
RNA was extracted from young or old pituitary the same as described in RNA-sequencing.cDNA synthesis was performed with Rever Tra Ace (TOYOBO, TRT-101) based on a mix of three gene-specific primers for the IgA locus (5'-ATCAGGCAGCCGATTATCAC-3', 5'-TCTCCTTCTG GGCACTCG-3', and 5'-TGAATGATGCGCCACTGT-3'). Then cDNA was used as template for PCR with primers listed in Extended Data Table S2.
In situ RNA hybridization (RNAscope) Pituitary was isolated from perfused mouse and fixed with 4% Paraformaldehyde (PFA) (Solarbio, P1110) overnight at room temperature.Paraffin embedded tissues were sectioned at 5mm and processed to RNAscope.RNAscope was performed using the Multiplex Fluorescence v.2 kit (Advanced Cell Diagnostics, 323100) according to the manufacturer's protocol.Probes for mouse Igkc (414291) and Igha (414281-C2) were commercially available from the manufacturer, and secondary Opal 520 reagents (FP1487001KT, Akoya Biosciences) were diluted 1:500 in TSA buffer.

Immunofluorescence staining
For pituitary IgA staining, paraffin sections (5mm) were prepared as described in in situ RNA hybridization.The sections were de-paraffinized in isopropanol and graded alcohols, followed by antigen retrieval and blocked with SuperBlockä (PBS) Blocking Buffer (Thermo Fisher Scientific, 37515) containing 0.2% Triton X-100 for 30 min and incubated with goat anti-mouse IgA-PE (SouthernBiotech, 1040-09, 1:200) antibody overnight at 4 C.After washing with PBST 3 times, sections were mounted with DAPI Fluoromount-G (SouthernBiotech, 0100-20).Section samples were imaged on a Nikon A1-R confocal microscope.For the IgA+ cell quantitation, each pituitary was taken at least 3 pictures and the IgA+ cells in each picture were quantitated and the average number was used as IgA+ cells per view of each mouse.For intestinal and pituitary section staining, tissues from Igha-cre; Ai6 or Pou1f1-cre; Ai6 mice were fixed in 4% PFA in PBS at room temperature for one hour and embedded in OCT compound, followed by frozen and cryosection (20mm).Cryosections were fixed for 10 min in 4% PFA in PBS at room temperature and mounted with DAPI Fluoromount-G after washing or staining with goat anti-mouse IgA-PE.Images were taken on Olympus VS120.
For AAV-EGFP infected pituitary analysis, fixed pituitary was embedded in OCT compound, frozen, cryosectioned (20mm) and fixed for 10 min in 4% PFA in PBS at room temperature.The OCT was washed out in PBS and sections were mounted with DAPI Fluoromount-G.Samples were imaged on a Nikon A1-R confocal microscope.

FACS
For Epcam positive and ZsGreen positive cells analysis and isolation, single cell suspension preparation was described in pituitary gland dispersion.The cells were stained with indicated antibody together with DAPI (1mg/mL) and analyzed or sorted by gating DAPI negative and FITC positive population.
For EGFP positive cells quantitation in AAV infected mice, pituitary single cell suspension was prepared and stained with DAPI before analysis.For IgA+ cell quantitation, pituitary was dissociated and single cell suspension was prepared firstly.Then the cells were fixed for 30 min in 4% PFA in PBS at room temperature.After washing with PBS twice, SuperBlockä (PBS) Blocking Buffer containing 0.2% Triton X-100 was added for permeabilization and blocking at room temperature.One hour later, the cells were incubated overnight with goat anti-mouse IgA-PE (1:200) at 4 C.After washing with PBST 3 times, cells were centrifuged and resuspended in 250mL PBS and passed through 40mM cell strainer (Corning, 352340).During the staining, cell suspension was centrifuged at 5000rpm for 3min to remove supernatant each time before changing incubation buffer.
For CD19 staining of spleen and pituitary cells from Igha-cre; Ai6 mouse, spleen live cell suspension was obtained by directly grinding the spleen and pituitary cell suspension was prepared as described in pituitary gland dispersion.Cells were passed through 40mM cell strainer and stained with APC anti-mouse CD19 Antibody (BioLegend, 115512, 1:500) and DAPI (1mg/mL).
Cell analysis and isolation were performed on a BD FACSAria sorter equipped with FACSDiva software (BD Bioscience).Data were analyzed with FACSDiva software.

Antibiotics treatment
IgA-high mice or Igha-cre; Ai6 mice were divided into two groups (control and treatment) randomly.Antibiotics cocktail containing ampicillin (Sigma, A9393, 5g/L), vancomycin (Sigma, V2002, 2.5g/L), metronidazole (Sigma, M1547, 5g/L), neomycin (Sigma, N6386, 5g/L) and Amphotericin B (Sigma, A4888, 50mg/L) were used for the treatment group.The treatment group mice were firstly daily oral gavage with 500uL antibiotics cocktail for two weeks when they were one month old, and then oral gavage with 500uL antibiotics once every three days until die.0.01% hydrogen chloride (HCl) was added to the drinking water of treatment group after the first two-week antibiotics gavage.Then the mice were sacrificed at 3-4 months old for pituitary analysis.

Figure 1 .
Figure 1.Single-cell RNA-seq (scRNA-seq) analysis reveals IgA+ cells increase in old mouse pituitary (A) ScRNA-seq of the pituitary from young (2-3 months old) and old (18-24 months old) female mice using the 10x Genomics Chromium platform.Uniform manifold approximation and projection (UMAP) clustering of 16,214 single-cell transcriptomes (9,407 from young and 6,807 from old) colored by cell-type clusters.Hsc represents for hormone-secreting cells, Endo for endothelial cells, Mes for mesenchymal cells, and Mac for macrophages.(B) Violin plot showing expression of immunoglobulin kappa constant (Igkc), immunoglobulin heavy constant alpha (Igha), and Jchain as log-normalized counts in young and old pituitary cells.Each dot represents the expression level in one single cell.Igkc***p = 4.72x10 À28 , Igha***p = 6.23x10À8 , Jchain*p = 0.03.p values are false discovery rate (FDR)-corrected, Wilcoxon differential expression test.(C and D) Scatterplots showing the expression of Cd19 (C) and Ms4a1 (D) in mice pituitary cells projecting on the UMAP plot.Gray to red coloring indicate low to high-expression levels.(E-G) Scatterplots showing the expression of Igkc, Igha, and Jchain in young and old mice pituitary projecting on the UMAP plot.Gray to red coloring indicate low to high expression levels.

Figure 2 .
Figure 2. Increase of clonal heterogeneous IgA+ cells in old pituitary (A) Heatmap showing the normalized expression of immunoglobulin light chains and immunoglobulin heavy chains in young and old pituitary.n = 4 samples per group.Gray to red indicate low to high-expression levels.(B and C) Quantitative polymerase chain reaction with reverse transcription (qRT-PCR) (B) and Immunoblotting (C) showing mRNA and protein expression of Igkc (k chain) and Igha (a chain) in young and old pituitary.n = 4 mice for young and n = 5 mice for old in B. Representative data of three independent experiments for immunoblotting is shown in C. (D) Immunohistochemistry image (left) and quantification (right) of IgA+ cells in the pituitary of young (n = 9) and old (n = 11) mice.Paraffin sections of pituitary from young and old mice were stained with 4 0 ,6-diamidino-2-phenylindole (DAPI) and IgA-PE antibody as described in STAR Methods.Blue is DAPI signal and red is IgA signal.Data were compiled from three independent experiments.Scale bar, 20mM.(E) Quantitation of the IgA+ cells in young and old pituitary.Pituitary cells from young and old mice are stained with IgA-PE antibody and analyzed by Fluorescence-activated cell sorting (FACS) as described in STAR Methods.n = 9 mice per group.(F) Pie charts depicting the relative abundance of 50 most frequent complementarity-determining region 3 (CDR3) amino acid sequences of immunoglobulin a heavy chain in aged pituitary from three mice.Data shown as mean G s.d., unpaired two-tailed Student's t test.*p < 0.05, **p < 0.01, ***p < 0.001.(G) PCR result showing selected v chains of IgA heavy chain expression in young and old pituitary.Primers specific for Igha are used for reverse transcription of mRNA from young and old mice pituitary and selected v chains are amplified by PCR.The darkest DNA marker band is 1000bp.Image is representative data of three independent experiments.

Figure 3 .
Figure 3. IgA-high mice are with higher gut-bacteria diversity (A and B) qRT-PCR (A) and immunoblotting (B) showing mRNA and protein level of Igkc (k chain) and Igha (a chain) in normal and IgA-high mouse pituitary.n = 5 mice for Con and n = 6 mice for IgA-high in A. Representative data of three independent experiments is shown in B. (C) FACS analysis and quantitation of IgA+ cells in pituitary of normal and IgA-high mice.n = 6 mice per group.(D) Igkc and Igha mRNA level in pituitary of IgA-high mice of different ages.M as Month, n = 6 mice for 1M, 2 M, 4M and n = 5 for 8M.(E) 16S rRNA gene profiling data comparing the gut microbiome of young, old and IgA-high mice.Alpha diversity indicated by Shannon index is shown.n = 9 mice in young and old, n = 10 mice in IgA-high.Data shown as mean G s.d., unpaired two-tailed Student's t test for a and c, one-way ANOVA test for d and Kruskal Wallis test for e. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 4 .Figure 5 .
Figure 4. Microbiota determines the amount of IgA+ cells in pituitary (A and B) mRNA (A) and protein (B) level of Igkc (k chain) and Igha (a chain) in IgA-high mouse pituitary without (Con) or with (Antibiotics) antibiotics treatment.n = 6 mice for Con and n = 5 mice for Antibiotics in A. Representative data of three independent experiments is shown in B. (C) FACS analysis and quantitation of IgA+ cells in pituitary of control and antibiotics treated mice.n = 9 mice per group.(D and E) qRT-PCR (D) and Immunoblotting (E) result showing Igkc (k chain) and Igha (a chain) expression in pituitary of specific pathogen-free (SPF) and germfree (GF) mice.n = 5 mice per group in D. Representative data of three independent experiments is shown in E. (F) qRT-PCR result showing mRNA level of Igkc and Igha in GF mice and GF mice transferred with microbiota from young, old and IgA-high mice as indicated.n = 6 mice per group.Data shown as mean G s.d., unpaired two-tailed Student's t test.*p < 0.05, **p < 0.01, ***p < 0.001.