Background
Age-related reductions in immune functions in vaccination, infection, and cancer have been better characterized [
1]; however, little is known regarding the age-associated changes in immune function that result in autoimmunity. This is a pivotal field that warrants a strong research focus as there is a clear association between age and increased incidence of many autoimmune diseases including rheumatoid arthritis (RA), myositis, and Sjögren’s syndrome [
2]. RA onset can occur at any age, but it usually develops in middle-aged adults between 40 and 60 years old [
3]. Recently, gut microbiota have been demonstrated to have a profound influence on host health and disease [
4‐
7]. Reports have found that microbiota in older people are different from those in younger adults [
8,
9]. We and others have demonstrated that gut microbiota can act as an environmental cue to induce autoimmune arthritis in both humans and mice [
10‐
13]. However, a causative effect of microbiota in age-related disease development among older individuals remains to be determined.
Here, we aimed to investigate the age-associated and microbiota-associated impacts on autoimmune disease. Because RA is an autoantibody-mediated autoimmune disease [
14,
15], we are interested in two crucial subsets of effector T cells that provide help to B cells for autoantibody production. T follicular helper (Tfh) cells are a crucial subset of CD4
+ T cells that helps B cells produce high-affinity and high-titer antibodies [
16‐
18], and an excessive Tfh cell response can lead to many autoimmune conditions including RA [
19]. T helper 17 (Th17) cells, a T effector cell type involved in many autoimmune diseases, promote both autoantibody production and inflammation [
20]. We used a TCR transgenic (Tg) autoimmune arthritis model, K/BxN mice [
21], to study the effect of age and microbiota on autoimmune Tfh and Th17 cells. K/BxN mice are an autoimmune arthritis model in which transgenic KRN T cells recognize glucose-6-phosphate isomerase (GPI), the self-antigen (Ag) presented by major histocompatibility complex (MHC) class II I-A
g7 molecules. As in humans with RA, autoantibodies are crucial for disease pathogenesis in K/BxN mice [
21]. By examining TCR Tg T cells in young and middle-aged K/BxN mice, we can compare T cells of the same TCR specificity with the only difference being the age of the mouse. This has been shown to be beneficial in aging T cell studies as it eliminates change in the T cell repertoire, an important variable that complicates the interpretation of aging T cell function [
22].
In contrast to the abundant gut-luminal commensals, mucosa-associated commensal species such as segmented filamentous bacteria (SFB) represent a minor but important part of the commensal community, as they can powerfully modulate host immunity [
23‐
27]. Our previous studies have shown that SFB-induced Tfh and Th17 cells contribute significantly to autoantibody production in young K/BxN mice, and a lack of either T effector cell type strongly ameliorates autoantibody production and autoimmune arthritis development [
11,
12]. Here we reveal that despite both Tfh and Th17 cells having been reported to participate in the pathogenesis of autoimmune arthritis in most of the studies using young adult mice in experimental settings, there is a clearly age-associated accumulation of Tfh but not Th17 cells in the middle-aged group compared to their young counterparts. Our results suggest that most of the accumulated middle-aged Tfh cells are of the effector phenotype. Our results further indicate that exposure to commensal bacteria SFB causes the young host to develop an overactive immune system, with a strong elevation in their Tfh cell response reminiscent of the middle-aged condition.
Methods
Mice
K/BxN mice were generated by crossing KRN TCR transgenic mice on the C57BL/6 (B6) background with NOD mice (F1 mice of KRN/B6 x NOD). Ankle thickness was measured with a caliper (J15 Blet micrometer) as described previously [
11]. All mice were housed at the SPF animal facility at the University of Arizona. The young mice were used at 6–8 weeks of age, while the middle-aged K/BxN mice were used at 10–15 months of age using the age guideline of the Jackson Laboratory (Jax) described as:
Middle age refers to a phase during which senescent changes can be detected in some, but not all, biomarkers of aging. For the middle-aged group, mice should be at least 10 months old. Senescence processes that begin in younger adults (for example, collagen cross linking and accumulation of activated/memory T cells) often can be detected by then. The upper age limit for the middle-aged group is typically 14–15 months, because at this age, most biomarkers still have not changed to their full extent, and some have not yet started changing.
All experiments were conducted according to the guidelines of the Institutional Animal Care and Use Committee at the University of Arizona under the protocol reference number 11-278.
Preparation of single-cell suspension from the lung and small intestine-lamina propria (SI-LP)
Lungs were perfused with 10 ml PBS to remove blood, and were finely minced. Minced lung was placed into 10 ml of digestion buffer containing 1 mg/ml each of Collagenase D (Roche) and MgCl
2 and 0.15 mg/ml DNase I (Sigma) in DMEM (HyClone). Lungs were digested for 20–25 min at 37 °C at 200 rpm then passed through a 100-μm cell strainer. A plunger from a 5-ml syringe was then used to grind remaining tissue pieces through the cell strainer. SI-LP cells were isolated as described, with some modification [
11]. Briefly, Peyer’s patches (PPs) were removed from the small intestine. The small intestine was opened longitudinally and excess mucus was removed by scraping gently with forceps along the length of the intestine. The intestine was then thoroughly washed in 5 mM ice-cold ethylenediaminetetraacetic acid (EDTA) in PBS and cut into 1-cm pieces, which were incubated in 40 ml of 5 mM EDTA and 0.145 mg/ml of DL-dithiothreitol (DTT) in DMEM for 60 min at 37 °C horizontally at a rotation speed of 100 rpm. After incubation, the epithelial cell layer, containing the intraepithelial lymphocytes, was removed by shaking, and then cleaned by pressing intestinal pieces over a 100-μm nylon layer on top of a paper towel. Intestinal pieces were then transferred to an eppendorf tube with 600 μl of digestion solution containing 1 mg/ml Collagenase D (Roche), 0.15 mg/ml DNase I (Sigma), and 200 ng/ml liberase Cl (Roche). The pieces of intestine were finely minced and transferred to a 50-ml conical tube containing 5 ml of digestion solution. Digestion was performed by incubating the pieces at 37 °C for 15 min with rotation at 200 rpm. After digestion, 10 ml EDTA/PBS was added and the solution was passed first through a 100-μm cell strainer, then through a 40-μm cell strainer. Cells were centrifuged and washed again with EDTA/PBS before being resuspended in 10% FBS DMEM for stimulation.
Antibodies and flow cytometry
For surface staining, fluorophore-conjugated monoclonal antibodies (mAbs) specific for CD4 (RM4-5), CD19 (6D5), CD45 (30-F11), PD-1 (RMP1-30), CD11b (M1/70), CD44 (IM7), CD62L (MEL-14), and TCRβ (H57-597) were obtained from BioLegend. Abs recognizing Fas (Jo2), CXCR5 (2G8), and TCR Vβ14 (14-2) were from BD Pharmingen. Anti TCR Vβ6 (RR4-7) was from eBioscience. FITC-conjugated peanut agglutinin (PNA) was from Vector Laboratories. For intracellular cytokine staining, cells were incubated for 4 h with BD GolgiPlug (1:1000 dilution), 50 ng/ml phorbol 12-myristate 13-acetate, and 1 μM ionomycin in DMEM (HyClone) supplemented with 10% FCS, 1% nonessential amino acids, penicillin, streptomycin, and glutamine at 37 °C. Intracellular cytokine staining was performed with Cytofix/Cytoperm (BD Pharmingen). Abs recognizing interleukin-17A (IL-17A, TC11-18H10.1) were obtained from BioLegend and Abs recognizing IL-23 (fc23cpg) were obtained from eBioscience. For intra-nuclear staining, buffers from a Foxp3 Staining Buffer Set (eBioscience) were used to stain Abs recognizing Ki-67 (B56, eBioscience) and Bcl-6 (K112-91, BD Pharmingen). Cells were run on an LSRII (BD Biosciences), and analyses were performed using FlowJo (TreeStar) software.
Microbiota reconstitution and quantification
Our SPF mouse colony and SFB colonization were maintained as described previously [
12]. Briefly, our SPF mouse colony was originally derived from Jax and was verified as SFB-negative (SFB– hereafter). SFB were initially introduced to our mouse colony by gavaging mice with feces containing SFB, from Taconic B6 mice. Later, SFB were passed by gavaging mice with feces containing SFB, collected from the SFB+ mice housed in our colony. SFB– mice were weaned at 21 days old and rested for 1 day. Then, mice were orally gavaged with feces containing SFB, collected in-house for 3 consecutive days starting at 23 days old. The middle-aged K/BxN mice aged 10–15 months were gavaged at the same time as the young mice. The SFB– mice were the un-gavaged littermate controls. Unless otherwise mentioned, the SFB colonization status was examined on day 10 after SFB gavage by SFB-specific 16S rRNA quantitative PCR as previously described [
11].
ELISA
Anti-GPI Ab titers were measured as described [
11]. Briefly, ELISA plates were coated with recombinant mouse GPI at 5 μg/ml, and diluted mouse serum was added. Subsequently, plates were washed and alkaline-phosphatase (AP)-conjugated anti-mouse IgG Ab was added. After the final wash, AP substrate was added and titers were quantified as optical density values using an ELISA reader. The Ab titers were expressed as arbitrary units, which were calculated from serial dilutions of sample serum and defined as the reciprocal of the highest dilution that gave a background optical density (OD) value set as 0.15.
In vitro Th17 polarization
Splenic naïve CD4+ T cells (2.5 × 104) from SFB young or middle-aged K/BxN mice were enriched by fluorescence-activated cell sorting (FACS Aria) and in vitro cultured in 96-well plates for 4 days in Th17 polarization conditions: anti-CD3ε (plate-coated, 2 μg/ml), anti-CD28 (2 μg/ml), anti-IL-2 (10 μg/ml, JES6-1A12, BioXCell), IL-6 (50 ng/ml, PeproTech), transforming growth factor (TGF)β1 (1 ng/ml, PeproTech), 6-formylindolo [3,2-b] carbazole (FICZ), (300 nM, Enzo Life Sciences).
Immunohistochemical analysis
Lung sections from SFB– and SFB+ K/BxN mice were perfused and fixed with 10% (vol/vol) buffered formalin and stained with hematoxylin and eosin (H&E) for histological evaluation of lymphocyte aggregation. Image analysis was performed using ImageJ software (NIH, Bethesda, MD, USA).
Statistical analysis
Differences were considered significant with p < 0.05 analyzed by Student’s t test (two-tailed, unpaired) or two-way analysis of variance (ANOVA) (Prism 6, Graph-Pad Software), with significance level denoted as: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. We also used Prism to calculate Spearman correlation.
Discussion
Despite progress in understanding loss of immune function in Tfh cells and GC B cells with age, much less is known about age-associated changes in these cell types in an autoimmune setting. We have found that in contrast to Th17 cells, Tfh cells are heavily accumulated in middle-aged autoimmune arthritic mice. Most of the Tfh cells in middle-aged K/BxN mice are of the effector phenotype (CD62L
loCD44
hi). However, we believe these CD62L
loCD44
hi Tfh cells are mostly effector memory cells and not effectors. Our rationale is that in contrast to their young counterparts, when middle-aged non-Tfh cells from PPs were exposed to SFB, a potent Tfh stimulus, they were not able to upregulate their Bcl-6 and undergo SFB-induced Tfh differentiation. Therefore, we propose that these middle-aged Tfh cells are generated at a young age while Tfh differentiation is active. They later survive as memory Tfh cells, which can readily become effector Tfh cells upon encountering self-antigen. In support of this hypothesis, we found that middle-aged Tfh cells in K/BxN mice are rather quiescent, demonstrated by their lower proliferation phenotype, which potentially helps their long-term survival [
36].
The data presented here, together with previous reports, suggest that the Tfh cells accumulated over time are major contributors to autoantibody production and arthritis development in middle-aged K/BxN mice. To begin with, the autoantibody level is an important disease index in the K/BxN model. We and others have reported that arthritis can be induced simply by passively transferring anti-GPI autoantibodies or serum containing anti-GPI autoantibodies into wild-type mice [
37,
38]. Our results also indicate that, compared to young mice, middle-aged mice had higher autoantibody titers, which corresponded with their increased ankle thickness (Fig.
1c). Secondly, our previous data showed that Tfh cells are required for autoantibody production in K/BxN mice by demonstrating that mice receiving CXCR5-deficient KRN T cells, which have impaired Tfh function, developed far fewer autoantibodies and less severe disease compared to those that received CXCR5-sufficient KRN T cells [
12]. Third, a soaring Tfh cell population accompanied by a robust GC B cell response was found in middle-aged mice compared to the young group (Fig.
4a and b), and it has been well-established that active Tfh cells are required to maintain GC B cell populations [
16,
17]. Finally, Bcl-6 is expressed at similar levels in Tfh cells in both the young and middle-aged groups (Fig.
5b and c). Because Bcl-6 is the master regulator of Tfh cells and controls the expression and function of many Tfh cell molecules [
16,
17,
39], together with the robust GC B cell response in the middle-aged group, these data suggest that middle-aged Tfh cells are functionally competent. Collectively, these data suggest that the accumulated Tfh cells in middle-aged K/BxN mice are key contributors to the robust GC B cell response, which in turn leads to increased autoantibody production and arthritis development in middle-aged mice. It is worth mentioning that research focusing on aging carries unique challenges, mainly attributable to the fact that the time required to age mice makes designing gain-of-function or loss-of-function experiments with older mice much less feasible than studies involving only young subjects. However, this is an urgent topic that requires intense future focus, as age is a major risk factor for countless diseases, including many autoimmune diseases, and the world’s elderly population has been increasing at an unprecedented rate.
Our results thus could help explain the long-standing observation that with age, there is a reduction in Ab response to infection or vaccination, whereas there is an increase in autoantibody production in autoimmune patients, a phenomenon that might appear to be controversial at first sight [
40]. This can be explained by an elegant study showing that despite aging mice displaying an increased Tfh population, the Tfh response to viral infections in aging mice is impaired, which is indicated by a severe reduction in the germinal center B cell response [
41]. This is in sharp contrast to our results, which demonstrate that the increased Tfh population helps B cells to develop a robust GC B cell response in an autoimmune condition. Our results provide a potential mechanism to explain these seemingly paradoxical findings. We propose that during the aging process, there are reduced T effector cell activities toward new antigens. Indeed, our results show declining differentiation of Th17 and Tfh cells in response to a newly introduced antigen, SFB, in middle-aged mice. However, autoimmune diseases such as RA often have a latent phase [
42], during which autoimmune Tfh cells respond to self-antigen from youth while Tfh differentiation is still active, and are accumulated with age. Because of their preferential accumulation and survival advantage over other non-Tfh cells during the aging process (Fig.
6a and b), Tfh cells eventually emerge as a significant cell type involved in the pathogenesis of RA in middle age. Recently, a population of blood circulating Tfh (cTfh) cells has been described as memory Tfh cells [
43,
44]. In support of our hypothesis that memory Tfh cells could contribute to autoimmunity in middle-age, a higher frequency of circulating Tfh cells have been detected in patients with RA, and there is a positive correlation of increased cTfh cells with disease activity and autoantibody production [
45]. T follicular regulatory cells (Tfr), serve as important counter-inhibitors that downregulate the Tfh cell response. It has been proposed that an increased ratio of inhibitory Tfr cells in aged mice contributes to defective Ab production in aging upon antigen immunization [
46]. However, in an autoimmune setting in middle-aged mice, we did not observe representation of Tfr cells over and above Tfh cells that could have better countered the accumulated Tfh population (unpublished observations).
It has been demonstrated that there is an increase in IL-17
+ CD4
+ T cells in aging mice and one of the reports further demonstrated that memory Th17 cells are an important source of IL-17 production in aging mice [
47,
48]. Our findings suggest that there is an overall reduction in Th17 cells in middle-aged K/BxN mice, which also displayed decreased differentiation of IL-17-producing cells from naive CD4
+ T cells compared to the young group. The discrepancy between our results and previous mouse data could be due to the difference between middle-aged mice used in this study compared to aged mice in the former study. The other, more likely reason is the difference between using mice with an autoimmune background (K/BxN) compared to wild-type mice (B6 and CBA) used in the previous study. On that note, it is worth mentioning that similar to our findings, there were two studies in humans reporting a reduction in IL-17
+CD4
+ T cells in the elderly compared to the young [
49,
50]. As in humans, CD4
+ T cells from autoimmune mice such as K/BxN mice are more antigen-experienced compared to naïve wild-type mice such as B6 and CBA. It will be interesting to determine whether the antigen experience of T cells contributes to the differences in IL-17 response between our findings and previous reports.
In humans, initial evidence suggests an important role for IL-17 in the pathogenesis of several inflammatory diseases, including RA [
20,
51]. However it was later discovered that the increased expression of IL-17 is not restricted to synovial tissue in patients with RA; it is also observed in patients with psoriatic arthritis and inflammatory osteoarthritis. The heterogeneous expression pattern of IL-17 in patients with RA has been proposed to be responsible for the non-responsiveness to anti-IL-17 clinical therapy in RA [
52]. The K/BxN model results further suggest that the problematic maintenance of Th17 cells in middle-aged compared to young individuals may further contribute to the ineffectiveness of anti-IL-17 therapy in patients with RA, who are mostly middle-aged and older individuals. This is supported by a report showing a significant but small number of IL-17
+ CD4
+ T cells are detected (0.9-1.2%) in the synovial fluid and peripheral blood of patients with RA [
53]. In this case, anti-IL-17 treatment may function better in younger or pre-clinical subgroups of patients.
Aging-associated alterations in gut microbiota composition have been well-documented [
8,
9]. They are suggested to be caused by an age-related decline in immune function. When we colonized K/BxN mice with SFB, we observed that middle-aged K/BxN mice displayed a trend toward an increase in susceptibility to SFB colonization compared to their young counterparts. Normal gut IgA is a key immune player in reducing SFB colonization [
54]. It is possible that the defect in SFB-induced Tfh differentiation in PPs indicated by the lack of SFB-mediated Bcl-6 induction in middle-aged mice could lead to reduced production of SFB-specific IgA, which results in an expansion of SFB in the gut. However, it appears that T cells in the middle-aged are less capable of responding to SFB, which can be seen at a molecular level, compared to young T cells. For example, in young K/BxN mice, we have previously reported that SFB promote non-Tfh cells differentiating into Tfh cells in PPs by upregulating Bcl-6 expression in non-Tfh cells in PPs [
12]. On the other hand, SFB colonization does not upregulate Bcl-6 expression in non-Tfh cells in PPs in middle-aged K/BxN mice.
A strong Tfh response can still be observed in SFB– middle-aged K/BxN mice, suggesting that commensals other than SFB may substitute for SFB in promoting the Tfh response during the aging process, which results in a strong accumulation of Tfh cells despite the lack of SFB. In contrast, SFB is more specifically required for Th17 than Tfh responses, which is supported by data from Littman’s group [
27]. Middle-aged K/BxN mice display reduced but still significant SFB-mediated Th17 induction in the lung (but not in the spleen or SI-LP), further supporting SFB-specific Th17 induction. We have previously reported that Tfh and Th17 cells can both contribute to autoantibody and arthritis development in young K/BxN mice [
12]. Thus, our data from middle-aged K/BxN mice suggest that the enormous amount of Tfh cells in the middle-aged group is sufficient to induce a robust GC B cell response that promotes severe arthritis with less help from Th17 cells. The reduced but still significant induction of Th17 cells (at least in the lung) in SFB colonized mice may partly contribute to the higher anti-GPI titer in SFB+ middle-aged K/BxN mice. Although this did not contribute to visible pathological differences in the mice, which could be due to the limitations of detection of disease symptoms and severity in mice, the differential auto-Ab titer may still contribute to different degrees of disease manifestation among patients. Using TCR Tg K/BxN mice in our study allowed us to significantly control the TCR repertoire over the aging process. However, it is a well-recognized phenomenon that incomplete allelic exclusion of TCRs can lead to dual TCR expression on T cells, not only in TCR Tg mice, but also in wild-type mice and humans [
55,
56]. Furthermore, dual TCR expression has been reported to increase with aging [
22]. Thus, it will be very interesting to determine whether microbiota contribute to dual TCR expression during the aging process, and how this may impact autoimmune development.
Our data also suggest that certain microbiota, such as SFB, may render young hosts prone to an overactive immune phenotype, as evidenced by the young K/BxN mice colonized with SFB displaying stronger disease and immune phenotypes, such as increased Tfh and GC responses, that are reminiscent of, though still much lower than, those in middle-aged mice. Interestingly, patients with RA have a phenotype of premature immune aging (immunosenescence) including telomere shortening and contraction of the T cell repertoire [
57]. Further investigation is required to determine whether immunosenescence in RA could be contributed to by the human equivalent of microbiota species similar to mouse SFB, with strong host immunomodulation effects.