Identification of Clec4b as a novel regulator of bystander activation of auto-reactive T cells and autoimmune disease

The control of chronic inflammation is dependent on the possibility of limiting bystander activation of autoreactive and potentially pathogenic T cells. We have identified a non-sense loss of function single nucleotide polymorphism in the C-type lectin receptor, Clec4b, and have shown that it controls chronic autoimmune arthritis in rat models of rheumatoid arthritis. Clec4b is specifically expressed in CD4+ myeloid cells, mainly classical dendritic cells (DCs), and is defined by the markers CD4+/MHCIIhi/CD11b/c+. We found that Clec4b limited the activation of arthritogenic CD4+αβT cells and the absence of Clec4b allowed development of arthritis already 5 days after adjuvant injection. Clec4b sufficient CD4+ myeloid dendritic cells successfully limited the arthritogenic T cell expansion immediately after activation both in vitro and in vivo. We conclude that Clec4b expressed on CD4+ myeloid dendritic cells regulate the expansion of auto-reactive and potentially pathogenic T cells during an immune response, demonstrating an early checkpoint control mechanism to avoid autoimmunity leading to chronic inflammation.


Author summary
To identify early disease regulatory mechanisms in autoimmune diseases such as rheumatoid arthritis (RA) is challenging not only because of the genetic and environmental complexity but also because of the critical autoimmune time-period that precedes the clinical diagnosis. Therefore, we set out to study the complex disease pathways in a more restricted setting. Through genetic segregation of rat crosses, followed by the selection of recombinants to produce minimal congenic strains, we have identified a single nucleotide polymorphism regulating the expression of Clec4b2 that in turn controls the development of arthritis. The Clec4b gene is normally expressed in a population of antigen-presenting cells that can limit enhanced activation of bystander autoreactive T cells during an immune-priming response. This previously unknown type of immune regulation reveals the existence of a mechanism protecting against autoimmune dieases by the avoidance of

Introduction
A tissue-specific autoimmune disease process starts decades before the clinical onset of autoimmune diseases, such as rheumatoid arthritis (RA) [1]. Most likely the first trigger involves the activation of autoreactive T cells, which are normally regulatory or anergic, into a more aggressive state. The activation requires strong costiumulation, which during an immune response is mediated by adjuvants carried by infectious organisms or possibly damaged endogenous cells, or environmental hazards such as tobacco smoke [2]. These challenges trigger the innate immune system, leading to the activation of autoreactive T cells. Innate immune cells interpret infectious intruders or danger signals via a range of pattern-recognizing receptors (PRRs) on their cell surfaces. When the innate cells sense enhanced risk in the environment, these cells are able to activate other cells, such as T cells. If the activation of adaptive responses displays joint specificity, the situation could initiate clinical arthritis. Animal models of arthritis mimic these disease stages [3]. They are initiated by adjuvant immunization followed by an autoimmune response to a tissue-specific protein.
In the case of collagen-induced arthritis (CIA), it is the type II collagen (CII) that is involved and in the case of arthritis induced by various type of adjuvants, such as pristane-induced arthritis (PIA), or mineral-oil induced arthritis (OIA), a bystander response is raised to a pattern of unknown endogenous auto-antigens [4]. Clinical arthritis starts to develop 2 weeks after the injection, as a result of an inflammatory attack on peripheral cartilaginous joints, involving the autoimmune response, that later can develops into a chronic inflammatory disease.
To determine the basic mechanisms leading to an autoimmune disease we searched for the genetic polymorphisms that allow the development of disease in certain inbred strains. For our investigation, we selected a cross between the DA rat, which is highly susceptible to autoimmune diseases, and the disease resistant E3 rat strain. The rats were injected intra-dermally with pristane, a simple alkene adjuvant oil which triggers a disease that fulfils the classification criteria for RA [5]. Through genetic linkage mapping, we identified 20 arthritis-associated loci in the DA rat [6]. One of the major loci was localized to chromosome 4 and was denoted Pia7. The same locus was identified using several different arthritis models, including CIA, PIA and OIA [7][8][9]. With the use of recombinant congenic strains, the disease association could be mapped to a gene family encoding C-type lectin receptors, denoted hereafter as APLEC (antigen presenting cell expressing lectin like receptor gene complex) [10,11]. APLEC contains a cluster of 8 genes, 4 of which have a carbohydrate recognition domain highly similar to the full-length Clec4a/Dcir and are thus denoted Clec4a1-3. The complex also covers 4 non-Dcir like C-type lectins: Clec4b2b/Dcar, Clec4d/MCL, Clec4e/Mincle and pClecn/Dectin2 (a nonfunctional pseudo-gene in both the DA and E3 strains). It has been proposed that both Clec4a and Clec4b are important in antigen presentation and uptake [12,13] as well as in T cell responses [14] and can be assumed to be of importance early in an immune response. Furthermore, there is a genetic association to human DCIR in RA [15]. In contrast, Clec4d and Clec4e are involved in later inflammatory responses [16]. We have now located a single nucleotide polymorphism (SNP) in the Pia7 locus, regulating the Clec4b gene. Only E3 derived cells express the Clec4b/Dcar receptor and it is expressed on CD4 + Dendritic cells. The loss of Clec4b/Dcar1 expression leads to dysregulation of CD4 + myeloid antigen presenting cells, which lose the ability to control bystander activation of autoreactive T cells. This previously unknown physiological regulator of autoreactive T cells, are of critical importance to the development of autoimmune disease.

A SNP within the Clec4b gene explain the arthritis-association
The DA.E3-Pia7 congenic strain is less susceptible to arthritis compared to the DA strain [11]. We chose to fine map the underlying polymorphisms within the Pia7 locus for association with arthritis using OIA, which gives a milder arthritis compared with PIA. In fact, the same locus as Pia7 has also been mapped in OIA and designated Oia2 [7]. We aimed to produce recombinant congenic strains to identify the experimental arthritis regulating polymorphism in Pia7. Whole-genome sequencing of both the DA and the E3 genomes was advantageous as it enabled us to identify at least 1 microsatellite marker/50 Kb within the APLEC region [6]. More than 5000 pups were screened for recombination events using a panel of 37 microsatellite markers in the original 1.5 Mb DA.E3-Pia7 interval. Two recombinant congenic strains were produced. The first identified sub-congenic strains, DA.E3-Clec4a contained arthritis protective E3 alleles only in the 4 Clec4a isoforms of the APLEC region where the recombination event occurred immediately after the Clec4a gene but before Clec4b2 between MSA-plec710 and MSAplec740. The DA.E3-Clec4bde sub-congenic strain was the product of two separate recombination events, containing E3 alleles in Clec4a, Clec4b2, Clec4d, Clec4e, and the Clec4n/Dectin2, which is a pseudogene in both DA and E3 (Fig 1A). Limited on the centromeric side by a recombination between MSAplec573 and MSAplec633 and distally by a recombination event occurring right after Clec4e between the markers D4rat90 and MSAplec981.
OIA assessment in the DA.E3-Clec4a strain showed no protection within this region, contrary to the full DA.E3-Pia7 congenic strain (Fig 1C). The sub-congenic strain DA.E3-Clec4bde on the other hand showed almost complete protection against OIA, as well as to PIA and CIA (Fig 1C-1E). Sequencing the DA.E3-Clec4bde fragment identified two coding polymorphisms: a nonsense mutation in the DA strain that causes a truncated protein at exon two of Clec4b2; and a non-synonymous SNP in the membrane domain of Clec4e (Fig 1B). The mouse has two isoforms of Clec4b; Clec4b1 and Clec4b2. Since the rat only has the Clec4b2 homolog we refer to this gene as Clec4b. The Clec4d gene was excluded as a candidate since it does not contain any coding or regulatory polymorphisms and does not show any differences in gene expression between strains ( Fig 1F). A comparison of the genome sequence with other inbred rat strains showed that the arthritis protective PVG and E3 share the same Clec4bde haplotype [6], making them distinct from the DA Clec4bde haplotype. The arthritis protective F344 strain, on the other hand, have a Clec4bde haplotype that is identical to DA. When trying to identify strains that deviate in their Clec4bde haplotype, only 1 strain was different. The BN strain shares alleles with DA only in the Clec4e locus (see Fig 1B), whereas the remaining sequence of the BN Clec4bde region is shared with E3, the arthritis protective haplotype. In a literature search to compare F2 crosses between DA and PVG, and between DA and BN strains, the arthritis quantitative trait locus at APLEC was reproduced in both crosses despite the arthritis susceptible version of Clec4e in BN. An APLEC arthritis QTL could not be reproduced in crosses between DA and the arthritis protective F344 strain, principally because F344 has the DA version of the full APLEC [9]. The inherent ability to reproduce linkage in an F2 cross between DA and BN provided sufficient evidence to us to conclude that arthritis regulation most probably stems from variations in Clec4b rather than Clec4e. We identified a causative nonsense SNP in the Clec4b DA allele that causes a truncation of the DA allelic protein that is nonfunctional but with expressed transcripts. Thus, the DA strain is a natural functional knock-out for Clec4b, which further substantiates the gene as the arthritis regulating gene in the APLEC region. The two DA strains with different Clec4b alleles are hereafter denoted DA and DA.Clec4b E3 , abbreviated as Clec4b E3 and occasionally � E3. To investigate if other noncoding variations from the fragment could be regulatory, expression of all APLEC encoded genes were measured in DA and Clec4b E3 in naïve splenocytes. Only Clec4b showed differential expression ( Fig 1F). All raw data to figures are available in the S1 Data file.
Clec4b is expressed in steady-state/immature CD4 + myeloid-derived spleen cells In order to identify possible mechanisms by which Clec4b could regulate experimental arthritis we first aimed to identify which cell types express the Clec4b gene. We analysed gene expression in blood, lymph nodes, thymus, bone marrow and spleen (Fig 2A). The highest Clec4b expression was found in the spleen, but lower expression levels could also be observed in blood and lymph nodes. Bone marrow showed lower expression compared to spleen whereas Clec4b expression is restricted to immature CD4+ myeloid cells in the spleen. a) Relative expression of Clec4b in five naïve immune-related tissue; spleen, blood, thymus, bone-marrow and inguinal lymph nodes (n = 5 for spleen and thymus, remaining tissues n = 3). b) Clec4b expression of naïve spleen cells. The samples were pooled from three different spleens. Cells were first negatively selected for TCR, and then positively selected for CD4 + and CD8 + cells, respectively. After CD4/CD8 sorting, the remaining cells were selected for CD11b/c. Neutrophils were enriched from blood using Ficoll-Hypaque centrifugation. c) Non-pooled spleen samples were negatively selected for T cells and positively for CD4. The CD4-fraction is the flow-through cell (n = 5). d) Relative Clec4b expression in three naïve Clec4b E3 or one DA sample, all sorted with a Mo-flow FACS sorter. The FACS plot depicts the gating strategy for the three assessed population. Before CD4 and His24 (CD45R), the population determining gating, the cells were negatively sorted by gating for TCR and CD45RA, eliminating T cells and B cells. The cells were His24 -CD4 + (CD4+ DC), His24 -CD4 -(CD4 -DC) or His24 + CD4 + (pDCs) (n = 3 for Clec4b E3 and n = 1 for DA). e) Relative Clec4b expression in naïve and oil-primed splenocytes selected for TCR-CD4+ and for day 0, day 1 day 2, day 3, and day 5 after oil injection. n = 5 for all groups. Significance stars depict Mann. Whitney Test � = p<0.05, �� = p<0.01, ��� = p<0.001 and SEM.
Next, we aimed to identify which splenic cell population expresses Clec4b. First, we pooled three naïve Clec4b E3 and three DA spleens and selected T cells and B cells using anti-pan-T and B cell labelled magnetic beads. The T and B cell negative flow-through was elected in two new magnetic labelled fractions, one for CD4 + and another for CD8 + . The CD4flow-through was lastly also selected for CD11b/c + cells. Clec4b expression was low in all cell types tested except for CD4 + T and B cell negative cells (Fig 2B). Then unpooled individual magnetic bead selected splenocyte samples were assessed, Clec4b expression was significantly enriched in cells from Clec4b E3 compared to DA, and more than a tenfold increase in CD4 + TCRcompared to the CD4 -TCRcells was observed ( Fig 2C). More than a sixtyfold difference in Clec4b expression was confirmed between CD4 + and CD4non-T/B cells when sorted with a flow sorter. This showed Clec4b expression exclusively in CD4+αβTCR -CD45RA -(a B cell-marker) cells, whereas CD4 -TCR -CD45RAhad close to undetectable levels of Clec4b ( Fig 2D). Interestingly, plasmacytoid dendritic cells (pDCs) are known to be CD4 + in the rat, and to determine if pDCs also express Clec4b we used a previously established rat pDC sorting protocol [17] followed by gene-expression assay. Lower levels of Clec4b were observed in pDCs, albeit without any strain difference ( Fig 2D). A recent study by Daws et al [18] supports both of these findings, suggesting that rat Clec4b is highly expressed in CD4 + myeloid cells but not in CD4 + pDCs.
The fact that OIA is a rapid arthritis disease model, with an onset of clinical arthritis within 10-12 days of injection, suggests that disease regulating genes are already operating within the first few days of injection. In order to examine changes in the kinetic expression of Clec4b during early arthritis-induction, CD4 + TCRbead selected splenocytes were isolated from both naïve animals and at different time points after mineral oil injection. The highest levels of Clec4b expression were observed in naïve splenocytes. A daily decrease in gene-expression was shown until day 5 after injection, when Clec4b levels were close to undetectable ( Fig 2E). This supports the argument that regulation of Clec4b expression is sensitive to rapid environmental changes, either transmitted by instant general immune responses to the oil injection. Alternatively, it indicates that the receptor is sensitive to the oil itself. The rapid downregulation of Clec4b after oil injection suggests that the C-type lectin receptor is expressed in naïve or steady-state DCs as a possible steady-state off-switch and that this signal could be important for accurate immune-activation.

The CD4 + Clec4b + cell population consists of two distinct subtypes, classical and granular/plasmacytoid dendritic cells
The Clec4b gene is expressed in CD4 + cells that are TCR -. CD4 is frequently expressed on pDCs in both rats and humans, [17,19]. It has been suggested that CD4 is expressed on a subpopulation of conventional dendritic cells, cDC2 [20] but this population is poorly characterized in the rat. Rat CD4 + DCs have been described as being better regulators of the MHCII + Th1 adaptive responses than CD4 -DCs [21], therefore, we wanted to investigate the phenotype and function of the CD4 + TCR -DC population. The CD4 + Clec4b + population was selected from naïve splenocytes with magnetic labelled pan T cell microbeads and, in the second step, with beads isolated TCR -CD4 + cells. The selected cells were then labelled with fluorescent antibodies and studied in a flow cytometer.
The yield from 100 x10 6 splenocytes was 3.0 x 10 6 ± 0.3 CD4 + TCRcells and 26.1x 10 6 ± 3.6 CD4 -TCR -. More than 80% of the CD4 + selected fraction expressed CD4 and had less than 2% T cells (Fig 3A and 3B). Within the CD4 + selected cell fraction, CD11b/c and the granulocyte , first selected to be TCR-, and after a second sorting separated into either CD4 + or CD4fractions using a CD4 antibody coated beads. NK cells are CD161hi, Neutrophils are His48+ and CD11b/c+, pDC are CD4+ and (CD45R) His24 + , B cells are CD45RA + (OX33), CD11b/c cells are OX42 + and His48 cells are granulocytes. b-d) Antibody labeled cell frequencies for CD4 + and CD4fractions in the total mononuclear cell population. b) Shows CD4 and TCR (clone R73) expression. c) Expression of CD11b/c and His48. d) Illustrate MHCII + expression (clone OX17) separated by cellular size by the forward scatter, and the three populations identified are also marker His48 were expressed in noticeably differently patterns in different sub-populations ( Fig 3C). When the expression of major histocompatibility complex type II molecules (MHCII) was analysed in combination with forward scatter (FSC), three distinct populations were detectable, each representing a third of the total CD4 + pan Tpopulation ( Fig 3D). One population expressed high levels of MHCII, a second population expressed intermediate levels of MHCII, and a third population was found negative for MHCII. The MHCII bright population showed classical dendritic cell characteristics, such as high CD11b/c and CD172 (Sirpα) expression and negative for the granulocyte marker His48 (Fig 3E). The MHCII dim population, on the other hand, was granulocyte-like with very high expression of His48. The population also expressed His24, one of the markers for pDCs, here defined by CD45RA -/TCR -/His24 + / CD4 + expression [17]. The MHCIIpopulation was more monocyte-like, expressing CD68 and CD11b/c at levels comparable to the MHCIIpopulation in CD4 -TCR -, S1 Fig. iNOS expression was elevated in both MHC expressing populations (Fig 3E), from which one may infer that reactive nitrogen species are involved in shared regulatory functions. In conclusion, the Clec4b expressing CD4 + cells were shown to include two unique populations. The first is a dendritic cell-like population corresponding to the CD4 + classical DC subset that has been previously described in the rat spleen [22], and which are similar to human cDC2. A second granulocyte-like population, which was undetected in the CD4 -/Clec4bfraction and showed lower MHCII expression, corresponded to the pDC population. Since the pDC population was shown to express very low levels of Clec4b (see Fig 2D), we conclude that the CD11b/c + DC population should be the main Clec4b expressing CD4 + population in the spleen. The production of a monoclonal antibody for rat Clec4b is described in recently published data from Daws et al. [18]. They also detailed how Clec4b is expressed mainly on CD4 + DC/monocytes [18]. We found that Clec4b was only expressed in Clec4b E3 CD4 + DCs but not in B or T cells ( Fig 3F). Comparing cellular expression of the Clec4b receptor showed significant differential expression between DA and DA congenic rats, expressing Clec4b E3 (Fig 3G).

Clec4b deficient DA rats showed enhanced expansion of CD4 + DCs at day 3 after oil injection.
With the intention of identifying how the non-sense mutations in the Clec4b gene affect the cellular immune response during the early phase of experimental arthritis, splenocytes were harvested from the first days after oil-injection and then studied by flow cytometry. Assessing the cellular distribution in naïve spleens did not show any differences in the composition of cell sub-set populations between DA and Clec4b E3 rats, S2 Fig. However, three days after intradermal injection of arthritogenic oil we observed an increase of CD4 + myeloid cells in DA spleens, as measured ex vivo (Fig 4A). Furthermore, the MHCII expressing subset of CD4 + DCs had expanded after activation and the difference between the two genotypes was more marked, discernible in the FACS plots ( Fig 4B). After day three, MHCII expression was twofold in DA compared to Clec4b E3 , and DA populations also showed enriched expansion of the adhesion molecule CD11b (Fig 4C). On day three no differences were seen for either T or B analyzed for each of the 8 samples and the cell frequencies are illustrated in a box-plot graph. e) Illustration of the cellular distribution based on antibody labeled fluorescence intensities for the three CD4 + subsets, classified by differential MHCII expression and cell size. The antibody expression frequencies described in the graphs are for the specific MHCII subsets. CD161 + population is actually CD161 intermediary and therefore is not describing NK cells which should be CD161hi. pDCs are CD4 + HIS24 + . No strain comparisons for the CD4 + subsets showed any significant differences. f) Histograms describing anti-clec4b antibody staining in flow cytometer for either DA and Clec4b E3 splenocytes. Three gates are shown in the T cells, B cells, and the CD4 + TCR-population. g) Mean fluorescence intensity for the Clec4b antibody illustrating expression per cell on the CD4 + TCRpopulation for both DA and Clec4b E3 .
https://doi.org/10.1371/journal.pgen.1008788.g003 cell subsets in the spleen, S3 Fig. Populations of antigen-presenting cells were generally more activated in DA rats, so that there were more CD4 + MHCII + CD80 + and CD4 + MHCII + CD103 + cells ( Fig 4D). However, neither pDC nor monocyte populations expanded differentially between DA or Clec4b E3 during the first days of activation. Kinetic dissection of the early cellular responses to injection of an arthritogenic oil further supports dendritic cells as the cells initially involved in the Clec4b mediated responses and hence the arthritis development control. Aiming to delineate the expression differences in the CD4 + DCs, after oil injection, a set of DC-activation associated genes was compared by quantitative PCR between naïve cells and cells harvested three days after oil injection. CD4 + DCs at day three had increased gene expression of IL12p40, Ifna, CD4 + DC activation marker Ifr4, and the pro-inflammatory regulator Inos suggesting an enhanced activated phenotype at day three. A comparison of DA to Clec4b E3 expression showed significantly higher expression of both Inos, and Irf4 in DA. Furthermore, there was already a significant difference in Stat1 expression between DA and Clec4b E3 in the naïve CD4 + myeloid cell population (Fig 4E).
In summary, the DA CD4 + DC population expanded at day 3 compared to rats with Clec4b E3 , displaying a distinct pro-inflammatory expression profile that promoted up-regulation and expansion of MHCII and adhesion markers such as CD11b. Thus, Clec4b deficiency lowers the threshold for classical CD4 + DC activation and MHCII expression.

Clec4b regulates activation of pro-inflammatory T cells before day five of OIA pathogenesis
Adjuvant induced arthritis models in the rat, such as PIA and OIA, are mediated by arthritogenic auto-reactive T cells activated a few days after adjuvant injection (see Fig 1C and 1D). Arthritogenic T cells can be isolated in draining lymph nodes as early as day 7 whereas the disease onset occurs around day 12 after adjuvant injection [23]. To investigate the Clec4b-mediated regulation of auto-reactive T cells during OIA development and to identify the earliest time point by which T cells can transfer disease, we transferred bead selected T cells (98% purity) from rats primed with mineral oil IFA into naïve recipient rats. Sufficient T cell activation for mild joint swelling was seen already at day 5 after oil injection (Fig 5A), whereas T cells from day 6 induced severe arthritis. Importantly, spleen T cells from DA but not DA. Clec4b E3 induced arthritis in naïve DA recipient rats. However, the presence of Clec4b in the recipient was irrelevant. As Clec4b is expressed in DC and not on T cells, these data demonstrates that Clec4b suppress T cell activation before day 5 during the immune response. Moreover, the absence of T cell regulation in Clec4b sufficient recipient rats shows that this early regulatory effect was absent after the T cell activation has advanced to autoreactive. T cells from draining lymph nodes transfer arthritis similarly to T cells from spleen (Fig 5B).
Since the T cells isolated at day 5 after oil injection were already arthritogenic, we sought to assess the in vivo primed T cells from day 0 by flow cytometry. As T cells from both sources transfer arthritis T cells [23], subsets and activation markers were analyzed from both joint draining lymph nodes and splenocytes. Surprisingly, no expansion of CD4 + CD25 + T cells was observed within the first 6 days after arthritis induction ( Fig 5C). However, a dramatic decrease of CD62L + CD4 + T cells among all live cells was seen in the spleen, from 15% to almost undetectable numbers at day two. This dramatic reduction of naïve T cells did not differ between the genotypes (Fig 5C) but provides a clue that this population is downregulated immediately only hours after oil injection.

Co-cultures of CD4 + DCs and T cells identifies differential limiting effects on naïve T cell already 24hrs after stimulation
The CD4 + DC subset is able to efficiently present antigen to T cells, and induce T cell activation [21,24]. Since no clear population differences between DA and Clec4b E3 rats could be identified for T cells in vivo within the first six days after disease induction, T cell activation was evaluated during in vitro stimulation where proliferative responses can also be studied. Cocultures were made to assess the CD4 + DC regulatory function on T cell proliferation by labelling bead-selected T cells with CFSE. Moreover CD4 + DC have been shown to be superior in stimulation of allogenic T cell responses in the rat compared to CD4 -DCs. We harvested allogenic T cells from rats with the MHC haplotype RT1 u , which differ from RT1 a in the DA rat, from both naïve rats and day three after IFA injection. Syngenic DA T cells were also harvested from day three after IFA, from both Cle4b E3 and DA rats. The different T cells were co-cultured with bead-selected CD4 + DCs from either DA or Clec4b E3 in all combinations. Naïve allogenic T cells were co-cultured with naïve CD4 + DCs, and the day three derived T cells, either allogenic or syngenic, were all co-cultured with day three in vivo derived CD4 + DCs. Since arthritogenic T cells in the rat have been characterized as activated CD4 + T cells, proliferation was assessed in CD25 + CD4 + T cells [23]. Proliferation was near 100% for both naïve and OIA day three harvested allogenic T cells however no difference in DC activation could be seen from Clec4b. This implies that Clec4b regulation on arthritogenic T cell development operates independent of MHCII. The CD4 + DC mediated T regulation occurs, most likely, by limiting bystander activation via cytokine secretion. Interestingly, syngeneic Clec4b E3 derived T cells showed dramatically lower proliferation compared to DA. Since no effect could be seen for the two CD4 + DCs populations in the co-cultures, the restricted proliferation seen for Clec4b E3 derived T cells must be due to regulatory differences during the first days in vivo rather than in the in vitro culture (Fig 6A and 6B).
In order to find an appropriate stimulatory agent triggering down-regulation of Clec4b expression, known pro-inflammatory compounds (ConA, LPS, mannan, thiodimycholate and CpGs) were incubated with freshly isolated naïve spleen cells. Remarkably, there was downregulation of Clec4b expression already after 24h of ConA stimulation (Fig 6C). However, there was no difference in viability of DCs after Con A between the two strains. Overnight ConA cultures activated naïve CD4 + DCs cells to become strong inducers of T cell proliferation compared with CD4 -DCs cells (Fig 6D). This ability to induce proliferation of CD4 + CD25 + T cells was unique to DA derived CD4 + DCs (Fig 6D).
Endeavoring to assess the restraining ability on T cell activation/proliferation from Clec4b E3 derived DCs, two sets of DC:T cell co-cultures were prepared. Naïve Con A overnight activated CD4 + DCs from DA and Clec4b E3 were combined with naïve T cells from either DA or Clec4b E3 . Overnight ConA stimulation of CD4 + DCs enabled the DCs from Clec4b E3 derived rats to limit proliferation of CD25 + CD4 + T cells from either DA and Clec4b E3 equally efficiently after 72 hours of culture (Fig 6E).
In summary, Clec4b E3 DCs from day 3 after oil injection failed to limit T cell activation in co-cultures with day 3 activated T cells, and T cell activation status correlated with the genotype of its origin (Fig 6B). Naïve Clec4b E3 DCs, on the other hand, that had been ConA stimulated before co-culture with naïve T cells were able to inhibit activation and proliferation of T cells independent of the T cells Clec4b genotype (Fig 6E). These data reveal that the arthritis regulating gene Clec4b expressed in CD4 + DCs can efficiently limit T cell activation in response to adjuvant/PRR stimulation, thus preventing the expansion of recently activated auto-reactive and potentially pathogenic T cells.

Fig 5. Arthritis transmitting T cells are induced already 5 days after oil injection. a) Transfer arthritis development in recipients after
injection with 15 x 10 6 spleen derived T cells from donor rats previously injected with mineral oil either 5 or 6 days beforehand. Donors and recipients are noted as d. and r. from either genotype; DA, or Clec4b E3 ( � E3). n = 5 for all groups. b) Transfer arthritis development in DA recipients after injection with 15 x 10 6 T cells from inguinal lymph nodes taken from donor rats day 6 after oil injection (both groups n = 5). c) Day to day distribution of T cell subsets as frequencies of total cell (n per time point = 5). Statistical analysis was performed using Mann-Whitney Test � = p<0.05 SEM. d) Day to day distribution of activation associated T cell markers as frequencies of total cells (n per time point = 5). https://doi.org/10.1371/journal.pgen.1008788.g005

Clec4b in recipient naïve DCs regulates T cell proliferation on day four activated T cells in vivo after challenge by innate stimulus and downregulation of IL17 production
Given the fact that only DA.Clec4b E3 and not DA derived T cells are prevented from the development into arthritogenic T cells after an intradermal injection of mineral oil, we propose that Clec4b should be able to allow DA donor T cell limiting responses in vivo after a second adjuvant injection in naïve recipients from the arthritis transfer model, as described in Fig 5A and  5B. Since T cells have already become arthritogenic 6 days after arthritogenic oil injection, we investigated the activation and proliferation status of T cells transferred earlier than this time point. CFSE + labeled T cells harvested from day 3 after oil primed DA rats were transferred to naïve DA.Clec4b E3 or DA rats. After transfering the primed T cells, the recipient was also challenged by an intradermal oil-injection to trigger the CD4 + DCs. After 3 days in the recipients, the spleens were harvested and more than 30 million splenocytes were analyzed per sample by flow cytometer. 3 distinct population of T cells could be visualized in recipient spleens: an unlabeled population (CFSE -); a proliferative population (CFSE low ); and a non-proliferating population (CFSE high ) (Fig 7A). Whilst the non-proliferating CFSE high population corresponds to the main part of the donor T cell population, a considerable part of the donor CD4 + T cells (i.e. the CFSE low population) were proliferating. Intriguingly, all cells in the CFSE low subset were CD4 + T cells, which corresponds well with previous findings from the Transfer arthritis model that has described the arthritogenic T cells as mainly CD4 + [23]. Since these donor T cells most likely have been exposed to a form of CD4 + DC bystander activation of abundantly polyclonal antigen specificity, there will have been no definite starting point for the T cell proliferation. Thus, the CFSE density could not be visualized into distinct cell divisions. Therefore, we used a second proliferation marker (Ki67) to assess proliferation within the CFSE categories. Further analysis of donor T cells in the Clec4b E3 recipients revealed a significant regulation of the proliferating CD4 + T cells in terms of both cell expansion and activation (both lower KI67 and MHCII expression) (Fig 7B and 7C). Conversely, no difference in activation and proliferation markers was detected in non-dividing (CFSE high ) and endogenous T cells (CFSE -) between both congenic rats. Furthermore, the effects of polyclonal stimulation of transferred cells indicate a role of Clec4b in modulating IL-17 producing T cells on the basis that a reduction in IL-17 + dividing T cells frequency was found in Clec4b E3 compared to DA rats (Fig 7D and 7E). However, no changes in the frequency of IFNγ + T cells was detected. Notably, CFSE high population failed to produce both IL-17 and IFNγ while endogenous T cells (CFSE-) produced IFNγ only. These findings corroborate the role of Clec4b in limiting T cell activation and proliferation in vivo. They also infer that Clec4b plays a part in controlling Th17 cells that together account for a substantial proportion of the DA susceptibility to adjuvant triggers that lead to arthritis development in the rat. TCR-spleen cells, after 24 h culture with either Con A, TDM, mannan, CpG, LPS, complete media only (n = 5 per group) Mann. Whitney Test �� = p<0.01 SEM. d) Comparisons of T cell subset in co-cultures mixing naïve DA T cells, cultured with Con A stimulated selected DCs, either CD4 + DCs or CD4-DCs from DA or Clec4b E3 . Frequencies of total cells are displayed for the co-culture at different DC:T cell ratios explained in 3 graphs; total CD4 + cells; total CD25 + CD4 + T cells; and proliferating (CFSE low) CD25 + CD4 + T cells. (n = 10 for all groups) Mann Whitney Test � = p<0.05, �� = p<0.01, ��� = p<0.001 and SEM. e) Describes two four-way grouped co-culture experiments comparing the proliferative regulation from either DA or � E3 (Clec4b E3 ) CD4 + DCs on bead selected CFSE labeled T cells from both rat strains and mixed in combinations in co-cultures for 72 h. n = 10 for all groups. 3 Graphs illustrate proliferative T cell subsets: total proliferating T cells, total CD4 + CD25 + T cells and the frequencies of proliferating CD4 + CD25 + T cells in all cells. Graphs describes regulation of proliferation on recently harvested naïve T cells by in vitro Con A activated CD4 + DCs in different T cell to DC ratios/concentrations. Mann. Whitney Test � = p<0.05, �� = p<0.01, ��� = p<0.001 and SEM. https://doi.org/10.1371/journal.pgen.1008788.g006

Fig 7.
Clec4b regulate T cell proliferation also in vivo and the activation correlate with down-regulation of IL17 production. DA rats were injected intradermally with oil and 4 days later the spleens were used as a source of oil primed T cells. These were CSFE labeled and injected intravenously into naïve DA and Clec4b E3 recipients that had been injected intradermaly with mineral oil the same day. a) Proliferation FACS plot describing CFSE intensity in CD4 versus CD8 T cells harvested from spleens of the recipients day 3 after oil injection. CFSE negative T cells are the recipient endogenous T cells. The CFSE high are the non-proliferating T cells from the donors and the CFSE low are the donor T cells that have been proliferating in the recipient. b) Histogram over the distribution of the three groups from a and how they differ in the expression of both Ki67 and MHCII on their surface. c) Graphs describing how the two donor groups, non-dividing and dividing T cells, differ in the expression of MHCII and Ki67 between DA and Clec4b E3 recipients. d,e) Illustration on how the expression of the two cytokines IFNg and IL17 is distributed between the three groups of differentially proliferating T cells between the Clec4b sufficient and deficient recipient rats. d) Illustrative FACS plot. e) Statistics analysis and graph. n = 4 for all groups, Mann Whitney Test � = p<0.05, �� = p<0.01, ��� = p<0.001 and SEM.

Discussion
The identification of a single nucleotide polymorphism controlling the expression of Clec4b in the rat highlights a new role of Clec4b in regulation of autoreactive T cell activation. The suggested identification of the Clec4b SNP to be causative was based different type of evidence. Firstly, it was found within a 200kb chromosomal interval defining the QTL Pia7. This genomic region harbors three protein coding genes, whereof two had non-synonymous SNPs. Of these, only the SNP in Clec4b showed functional evidence as it led to differential gene expression in naïve rats. Secondly, data from several genetic mapping studies aiming to identify arthritis susceptibility regulating loci in the rat, shows that only the DA version of Clec4b segregates with disease susceptibility, and not strains lacking the Clec4b SNP but maintaining the polymorphism outside the Clec4b locus. Thirdly, only the DA Clec4b.E3 congenic, and not the DA rat, expresses the Clec4b protein. More specifically, Clec4b is highly expressed on CD4 + DCs, a population known to be regulating MHCII interaction and T cell activation. Interestingly after an intradermal oil injection, the DA strain displays a rapid onset of autoreactive T cells whereas Clec4b.E3 rats have normal control of T cell activation after oil injection. Clec4b is a C-type lectin receptor, expressed in classical CD4 + DCs, and we now find that it is essential for the control of bystander activation of autoreactive T cells during immune priming. Deficient expression of Clec4b allows the escape and expansion of autoreactive T cells leading to arthritis.
T cells from a rat strain deficient in Clec4b expression are able to transfer arthritis to naïve rats already 5 days after being activated by adjuvants. Two-photon imaging of DC T cell interactions in vivo has revealed that, after adjuvant stimulation, T cells quickly attach to the activated DCs via interaction with peptide-loaded MHCII that then leads to T cell viability decisions within the first 12 hours [25]. Our data suggest that Clec4b is a regulator of these viability decisions and thereby helps to prevent T cells from responding to self-peptides that are already present on classical DCs and vulnerable to activation. It is suggested that immature/ steady state DCs have a tolerogenic function during steady-state [26]. During pattern recognition, receptor-mediated activation of DCs will occur at a critical time point when these selfpresenting DCs need to differentiate or die in order to prevent development of autoimmunity [27]. We suggest that Clec4b function is a control switch for the tolerogenic DCs to either transform or die during the short interval when the DCs have received secondary inflammatory signals, yet are still capable of activating T cells directed to an immunogen that are, in most cases, derived form a pathogenic intruder. We have identified two triggers of Clec4b, mineral oil and Con A, which will activate DCs lacking Clec4b expression, leading to uncontrolled T cell proliferation. The presence of the lectin receptor Clec4b/Dcar on the DC surface cause a more controlled activation of DC with Clec4b E3 , compared with DC with Clec4b DA , inducing less expression of proinflammatory genes such as Stat1, Irf4 and iNos, that could limit bystander T cell activation.
Importantly, mature T cells that possess a variable degree of self-reactivity and have escaped negative selection-through low avidity or by the absence of the relevant tissue antigens during their selection-can potentially cause autoimmune disease [28,29]. During the first priming immune response there are multiple mechanisms leading to activation of T cells with high affinity to peptides derived from the invading pathogen or immunogen. In the same time period, however, other cells with variable degree of self-reactivity need to be suppressed. We have now identified a way in which such autoreactive T cells can be suppressed by dendritic cells expressing Clec4b during a limited period during the immune priming. If Clec4b is not functioning properly, autoreactive T cells could be activated that could be pathogenic and cause an autoimmune arthritis that mimics rheumatoid arthritis.
The CD4 + DCs population in the rat spleen have previously been described to be highly efficient inducers of MHCII restricted T cell activation and to express enhanced levels of co-stimulatory receptors [22]. These also express PRRs on their surface [21]. A possible mouse counterpart to the rat CD4 + DC is CD11b + CD8a -DC, which have been described as a mediator of peripheral tolerance via MHCII [30]. Targeting antigen bound antibodies to a steadystate associated C-type lectin receptor DEC-205 without antigen can induce tolerance in mice [31,32]. In the mouse, there are two isoforms of Clec4b and of these Cle4b2 could have a different function as it is expressed on CD8 + DC and involved with cross-presentation to MHCI of peptides circulating in the early endosome [13]. Instead, the less homologous isotype Clec4b1 (DCAR2) could be functionally more similar to rat Clec4b as it is co-expressed with Clec4a4 (DCIR2) [33] on mouse CD8 -CD11b + DCs, the mouse splenic DC subset analogous to the rat CD4 + DCs. Clec4a4 was recently identified as a mediator of inflammation and MHCII:T cell immunity because both its function and the DC specificity show similarities with the identified function of rat Clec4b [14,34]. However the proportion of CD4 + and CD8 + dendritic cell populations differ between rats and mice where rat DCs better resembles human DC subsets [35]. However, rat Clec4b most likely does not have a completely overlapping set of functions with human CLEC4C, since the latter is a human pDC bio-marker [36,37] and Clec4b fails to show significant expression in rat pDCs.
Clec4b expressing CD4 + DCs plays an important role in limiting the expansion of autoreactive T cells during an immune response and, in particular, when the immune response is triggered in the absence of strong antigens. This new role of antigen-presenting DCs is of fundamental importance in understanding autoimmune diseases but may also have important implications for current explanations of the lack of autoreactive T cell responses in growing tumors.

Selective breeding of congenic animals
DA/ZtmRhd (DA.RT1 Av1 ) and E3/ZtmRhd (E3) were originally obtained from Zentralinstitut fur Versuchstierkunde (Hannover, Germany) but maintained as inbred strains in our laboratory for more than 20 generations. The original 2.3Mb DA.E3-Pia4 congenic strain was produced and back-crossed to DA for at least 16 generations [11]. Two recombinant congenic strains were produced by marker-assisted selective breeding from back-crossed DA.E3-Pia4 heterozygotic parents in an F2 inter-cross. During the process of producing the recombinant strains the congenic strains were back-crossed to DA three more times. The DA.E3-Clec4a recombinant was obtained after 567 meiosis and occurred between MSaplec710 and MSa-plec740 (which is downstream of Clec4a before Clec4b2. DA.E3-Clec4bde) was produced from intercrossed DA.E3-Pia4 in two steps. The first recombination occurred between MSaplec573 and MSaplec633 upstream of Allogenic rats were congenic DA.RT1 u . Animals were maintained at the Scheele Laboratory, in the Karolinska Institutet, in a pathogen-free environment according to the Federation of European Animal Laboratory Science Association guidelines (FELASA). All procedures and animal experiments have been approved by the local ethics committee (Stockholms djurförsöksetiska nämd). The animals were maintained in a 14h light and 10h dark cycles and held individually in cages containing wood shavings and aspen. Moreover, they had free access to water and were fed with the standard rodent chow.

Arthritis experiment
Arthritis was elicited by a single intradermal injection at the dorsal side of the tail base. For PIA induction, 150 μl pristane (2,6,10,14-tetramethylpentadecane, 95%; Acros Organics, Morris Plains, NJ, USA) was injected. For induction of OIA, 200 μl mineral oil (incomplete Freund's adjuvant /IFA/, BD, Franklin Lakes, NJ, USA) was injected. For induction of CIA, rat collagen type II (CII), dissolved in 0.1M acetic acid and emulsified in IFA, was injected. All used rats were >8 weeks old. All experiments were performed with littermate rats that were age-matched, distributed within the cages and blind evaluated by the investigator. The rats were regularly inspected, including monitoring the limbs for arthritis development by a previously described macroscopic scoring systems. Briefly, 1 point was given for each individual swollen and erythematic toe and up to 5 points for an inflamed ankle (15 points in total per paw) [38,39]. The scoring was carried out every second or third day for 20-30 days after disease induction. No points were given to deformed paws that did not exhibit signs of inflammation.
For transfer OIA, we harvested spleens or inguinal lymph nodes. Concisely, at day 6 or day 5 post injection organs were made in to single cell suspension. T cells were positively selected using magnetic CD6 Macs beads. Cells were then culture for 56 hours in complete media and stimulated with 10 μg/ml αTCR (R73) and 5 μg/ml αCD28 (JJ319) and later injected into recipient naïve rats. Cells from DA or Clec4b E3 were injected into DA or Clec4b E3 recipients and arthritis development was monitored using a 0-60 macroscopic scoring system. All experiments follow established guidelines involving littermate controls, mixing of groups in cages and blinded scoring evaluation.

Cell preparation and immuno-sorting
Splenocytes were harvested from naïve spleens 1-5 days after oil injection. Single cell suspensions were prepared as follows. Inguinal joint draining LNs and spleens were harvested from euthanized rats and digested for 30 min at 37˚C with 1 ml digestion buffer (1 mg/ml collagenase IV, Sigma-Aldrich, St. Louis, MO, USA) in DMEM media (Sigma-Aldrich), Spleens were passed through 40-μm cell strainers and washed twice in sterile PBS. Red blood cells were lysed by adding 2 ml of ACK buffer (155 mM NH 4 Cl, 10 mM KHCO 3 and 0.1 mM EDTA) to the pellet. After incubation for 3 mins at 4-8˚C, lysis reaction was stopped by adding cold PBS, centrifuged (8 min, 300g) and washed in ice-cold PBS.
Immuno-precipitation of samples in single cell suspensions was performed according to the manufacturer's protocol (Miltenyi biotec. Gladbach, Germany). T cells were selected using OX52 (CD6) Macs beads. T cell negative fractions were harvested from the flow-through. The T cell-purity was never less than 97%, the non-T cell flow-through had less than 3% contaminant T cells. The non-T cells flow-through was used for sorting the CD4+ non-T cell fraction using Macs CD4+ beads (coated with OX38). CD8+ or CD11b/c+ non-T cells were selected by using streptavidin-coated Macs beads coupled to biotin conjugated anti-CD8 antibodies (clone OX8, BD Franklin Lakes, NJ, USA) or CD11b/c+ (clone OX42, Biolegend San Diego, CA, USA). B cells were selected from single cell suspensions using CD45RA coated Macs beads (Clone OX33). Neutrophils were selected from EDTA-treated whole blood. First from this selection were PBMCs isolated by 6% dextran and centrifugation, followed by neutrophils isolated using Ficoll Hypaque, according to Current Protocols in Immunology 7.23.1.

Co-culture assay
Splenocytes from either naïve-or oil-injected rats were prepared in single cell suspension and selected with antibody coated beads, either in a negative or positive selection for the following surface receptors: CD4 + CD6 -(CD4 + DCs), CD4 -CD6 -(CD4 -DCs) or CD6 + (total T cells) as previously described. Cells were cultured in complete DMEM media (10% FCS, 1M Hepes (Sigma, St Louis, MO, USA), 50 uM 100U/ml penicillin, 100 ug/ml streptomycin, 50 uM βmercaptoethanol (Sigma St Louis, MO, USA)). In-vitro stimulation of naïve CD4 + DCs was prepared in 24-hour cultures of complete media in addition to one of the following: 3ug/ml Concanavalin A (Sigma, St Louis, MO, USA); 5 ug/ml Trehalose Dimycolate (TDM)(Sigma, St Louis, MO, USA); 10 ug/ml Mannan (Sigma, St Louis, MO, USA); 50 ug/ml CpG-A 1668 (Invivogen, San Diego, CA, USA) or 3 ug/ml LPS (Sigma, St Louis, MO, USA). For proliferation assays magnetic-bead selected T cells were labeled with CFSE 2 uM before 72 hrs culture in complete media with either in vivo activated CD4 + DCs taken from day one, day two or day three of after oil injection or CD4 + DCs activated in vitro for 24 hours with Con A.

In vivo proliferation assay and cytokine ELISA
T cells were harvested from four DA rats at day four after oil injection. T cells were harvested from whole spleen and lymph nodes and selected using magnetic CD6 Macs beads and subsequently labeled with CFSE. The labeled T cells where then immediately injected to naïve rats from either the DA or Clec4b E3 strain. T cells from one rat from day four oil injected donors, were split between one DA and one Clec4b E3 recipient. The naïve recipients where then challenged with an intradermal injection of oil and, three days later, spleens and draining lymph nodes were collected. The cells were labeled with antibodies to CD3, MHCII and CD25 in order to measure expression together with CFSE from the donor T cells. More than 100 million cells where assessed per recipient animal using a flow cytometer. Other cells were cultured with PMA, ionomycin and brefeldin A for 6 hours and further stained intracellularly with anti-IL17, IFNg and IL10 antibodies and analyzed with the flow cytometer.

Ethics statement
All animal experiments were approved by the local ethics committee (Stockholms djurförsöksetiska nämd) and all experiments were carried out according to method approved (Ethical Approval Number N35/16). Animal anesthesia was performed using an inhalant gas mixture of isoflurane and air. All euthanasia was carried out using CO 2 gas.

S1 Fig. Illustration of cell population subsets within both the CD4+ as well as the CD4fraction of antigen presenting cells depending on their expression of MHCII.
Three spleens harvested from day 3 after oil injection were selected first for T cells were the negative fraction was collected and selected a second time for CD4 expression. The twice selected cells were either CD4 + or CD4 -, each represented as 3 samples. The samples were then labeled with fluorescence conjugated antibodies to identify subsets of cells. Since there appeared to be 3 clear subsets of CD4+ DCs depending on the MHCII expression the cells were first gated as either MHCII negative, MHCII high/bright and MHC dim. The value on the x axis is the present of total number of cells.