MR1-restricted T cells (MR1Ts) are an abundant class of CD8+ T cells with a high prevalence in the peripheral blood as well as certain organs such as the lungs [1-3]. They detect a variety of pathogens, including Mycobacterium tuberculosis (Mtb), and produce proinflammatory cytokines in response to these microbes [1,2]. Unlike MHC-Ia molecules, which display peptides, MR1 is an MHC-Ib molecule that displays small microbial metabolites derived from riboflavin biosynthesis and other metabolic pathways [4-6]. Our previous work established the reliance of MR1-dependent presentation of Mtb on endosomal trafficking [7]. We also described differences between MR1 presentation of Mtb, an intracellular microbe, and exogenously added antigens [8]. Critically, the endosomal trafficking proteins identified in MR1-dependent antigen presentation played no role in HLA-B antigen presentation [7,8].
The human airway epithelial cell line BEAS-2B is very effective at presenting Mtb antigens to MR1Ts despite a lower infection efficiency compared to human dendritic cells (DCs) [9]. In DCs, Mtb interference with phagolysosome maturation results in a phagosome that is positive for the early endosome marker Rab5 [10]. In contrast, in BEAS-2B, Mtb resides in a late endosomal compartment defined by Rab7 and LAMP1 [9]. When MR1 tagged with GFP is overexpressed in these cells, the MR1 vesicles also associate with Rab7 and LAMP1 [7]. Despite the commonality of endosomal markers between the MR1 and Mtb compartment in BEAS-2B, it is unclear whether MR1 is physically present in the Mtb compartment.
Calcium signaling is necessary for the proper functioning of the immune system. The intracellular calcium concentration is tightly controlled and elevation of the calcium concentration triggers cellular mechanisms such as cytokine release and proliferation [11]. Elevation of the intracellular calcium concentration can be diffuse or local. Late endosomes are rich in intracellular calcium and local calcium release regulates their trafficking [12,13]. This calcium release is governed by endosomal calcium channels, which include the mucolipins and two-pore channels (TPC) [11,13]. Inhibition of TPCs leads to impaired endosomal trafficking, which has implications for the control of intracellular infections. For example, in a mouse model of Ebola virus infection, pharmacologic inhibition of TPCs decreased infectivity and improved survival by preventing the trafficking of the virus in the infected cell [14]. Given the importance of endosomal trafficking in MR1-dependent presentation of Mtb, we hypothesized that endosomal calcium signaling would play a role in MR1-dependent presentation of Mtb.
First, we investigated whether calcium channel blockade with the L-type calcium channel blocker tetrandrine affects MR1-dependent presentation of Mtb. To measure antigen presentation, we used specific human T cell clones that produce IFN-γ in response to Mtb infected antigen presenting cells (APCs). We infected BEAS-2B with H37Rv Mtb (MOI 8). After 6 hours, cells were treated with 10uM tetrandrine or DMSO. After overnight treatment, the cells were incubated with MR1Ts or HLA-B45-restricted human T cell clones in an IFN-γ ELISpot assay. We found that tetrandrine decreased both MR1- and HLA-B45-restricted antigen presentation (Fig. 1a). The small molecule 6-FP is an MR1 antagonist derived from the photodegradation of folic acid [4]. Although unable to activate MR1Ts, 6-FP is a potent MR1 ligand that is loaded in the endoplasmic reticulum and induces MR1 translocation to the cell surface [15,16]. Using BEAS-2B stably transduced with a doxycycline inducible MR1-GFP construct (TET-MR1GFP), we found that tetrandrine decreased MR1 surface stabilization after treatment with 6-FP compared to control (Fig 1b). We also found that tetrandrine caused enlargement of MR1GFP vesicles (Fig 1c).
The effects of tetrandrine on MR1- and HLA-B45-restricted antigen presentation suggested that the spectrum of activity of tetrandrine was too broad, therefore we sought to test the role of specific endosomal calcium channels in antigen presentation. We used RT-qPCR to determine the relative amounts of TPC1 and TPC2 in BEAS-2B and primary human DCs. We found that TPCs were expressed at levels similar to or greater than MR1 in both cell types and DCs had a higher expression of MR1, TPC1 and TPC2 compared to BEAS-2B (Fig. 2a). TPCs release endosomal calcium in response to the ligand nicotinic acid adenine dinucleotide phosphate (NAADP), a process inhibited by the small molecule trans-Ned-19 (N19) [17,18]. To test the effect of N19 on antigen presentation, we treated BEAS-2B with 25uM N19 or DMSO 6 hours after Mtb infection and performed IFN-γ ELISpot assays. We found that N19 decreased MR1 presentation of Mtb but had no effect on HLA-B45 presentation (Fig. 2b). Next, we used exogenous antigens to test whether the effect of N19 was limited to Mtb. For MR1, we used filtered Mycobacterium smegmatis supernatant (Msmeg supernatant) and for the HLA-B45-restricted T cells, we used peptide CFP102-9. We found that antigen presentation of exogenously added antigens was unaffected by N19 (Fig. 2c). Microscopy of BEAS-2B transduced with TET-MR1GFP showed no perturbation of the vesicles by N19 (Fig. 2d). These data indicate that TPC calcium release is important for MR1-mediated presentation of Mtb but not for HLA-B45 presentation of Mtb or any exogenously delivered antigens.
To determine if N19 impacted MR1 surface stabilization, we performed flow cytometry on TET-MR1GFP cells treated with 6-FP. We found that N19 had no effect on MR1 surface stabilization after addition of 6-FP (Fig. 3a). These data indicate that N19 does not interfere with MR1 loading in the ER or the ability of the 6-FP loaded MR1 to reach the cell surface. To determine whether N19 caused a change in MR1 transcripts in Mtb infected cells, we performed RT-qPCR on Mtb infected cells that were treated with N19 versus control (Fig. 3b). We found no difference in MR1 transcripts between the two conditions, indicating that the effect of N19 on Mtb infected cells was not due to changes in MR1 mRNA levels.
The most potent MR1 ligand is 5-OP-RU, a product of 5-A-RU and methylglyoxal [19]. Although 6-FP does not activate MR1Ts, pretreatment of APCs with 6-FP boosts T cell responses when exogenous antigens, including 5-OP-RU, are used [8]. Because 5-A-RU is unstable, a 5-A-RU prodrug was developed that requires enzymatic cleavage in acidic endosomes to form 5-A-RU [20]. To test whether N19 affected 6-FP mediated boosting of the 5-A-RU prodrug, we treated BEAS-2B with N19 or DMSO and 6-FP or 0.01M NaOH. The next day, the cells were used in an IFN-γ ELISpot assay and additional N19 was added into the ELISpot wells so that N19 would be present with the 5-A-RU prodrug and the MR1Ts. We found that N19 had no effect on 5-A-RU prodrug presentation or the boosting effect from 6-FP pretreatment (Fig. 3c).
The exclusive effect of N19 on MR1 presentation of Mtb raised the possibility that N19 could affect Mtb uptake or Mtb viability. Either mechanism would lead to a decrease in the quantity of Mtb antigens available to MR1, and it was shown that N19 can affect small and large particle uptake in short assays of up to 90 minutes [21]. While N19 did not affect the processing or presentation of Mtb derived CFP10 (Fig 2c, left), we sought to directly test whether N19 affected Mtb uptake or Mtb viability. We performed growth assays on Mtb infected BEAS-2B treated with N19 and found a mean difference of 8.23 Mtb colony forming units (CFU) per 100 cells, which was not statistically significant. (Fig. 3d). These data are consistent with the Mtb/HLA-B45 assays, which showed no effect from N19.
Although BEAS-2B present Mtb to MR1Ts, they are not professional APCs and the mechanisms underlying myeloid presentation of Mtb likely differ from those of airway epithelial cells. To determine whether the effect of N19 was generalizable to professional APCs, we treated primary human DCs with 100uM N19 or DMSO 6 hours after Mtb infection and performed IFN-γ ELISpot assays. We found that N19 affected MR1-dependent presentation of Mtb in DCs (Fig. 4a). There was no effect on MR1-dependent presentation of exogenous antigens (Fig. 4a). Since DCs have HLA-II, which has a well-established endosomal trafficking component, we characterized the effect of N19 on HLA-II presentation of Mtb using an HLA-II-restricted T cell clone that detects CFP10 [10]. An HLA-B8-restricted T cell clone was used as a control. We found that N19 had no effect on HLA-II or HLA-B8 presentation in DCs (Fig. 4b,c). These data show that in DCs, N19 specifically affects MR1 presentation of Mtb without impacting HLA-II, HLA-B8 or exogenous antigen presentation.
In humans, there are two TPC proteins, TPC1 and TPC2 [12,13]. To confirm that the effect of N19 was due to inhibition of TPCs, we first used the TPC2 inhibitor YM201636 [22]. Inhibition of TPC2 in BEAS-2B had no effect on MR1 or HLA-B45 antigen presentation of Mtb or exogenous antigens (Fig 5a,b). Since a TPC1 inhibitor is not available, we performed siRNA knockdown of TPC1 in BEAS-2B. Knockdown of TPC1 reduced TPC1 transcripts by 76% in 48hrs (Fig 5c). Functionally, TPC1 knockdown resulted in a significant reduction in MR1-mediated presentation of Mtb and a small effect on MR1 presentation of exogenous antigens (Fig. 5d,e). However, there was no significant effect on HLA-B45 presentation of Mtb or exogenous antigen (Fig. 5d,e).
Since TPC-mediated calcium release facilitates endosomal trafficking, and TPC1 is associated with recycling and late endosomes [23.24], we hypothesized that calcium-sensitive endosomal trafficking proteins play a role in MR1 presentation of Mtb. Synaptotagmin 7 (Syt7) is a calcium sensitive endosomal trafficking protein implicated in LAMP1 delivery to phagosomes and translocation of MHC-II to the plasma membrane [25,26]. To test the role of Syt7 in MR1-dependent antigen presentation, we knocked down Syt7 using siRNA, resulting in an 86% reduction in Syt7 transcripts (Fig 5f). Functionally, Syt7 knockdown affected MR1-dependent presentation of Mtb without a significant effect on MR1 presentation of exogenous antigens (Fig. 5g). Taken with the TPC1 knockdown experiments, these data show that TPC1 and Syt7 knockdown have MR1-Mtb specific effects consistent with N19 treatment.
McWilliam and colleagues established the importance of the ER MR1 antigen presentation [16]. In this model, a preformed pool of MR1 is retained in the ER. After acquiring a ligand that forms a Schiff base with lysine 43 of MR1, MR1 associates with β2M and translocates to the cell surface [16]. This has been demonstrated with exogenously added ligands such as Acetyl-6-FP and 5-OP-RU. In the setting of Salmonella enterica serovar Typhimurium infection, MR1-dependent antigen presentation was substantially reduced by Brefeldin A, which prevents protein egress from the ER [16]. Follow-up studies using a fluorescent MR1 ligand confirmed that the loading occurred in the ER [27]. In contrast, our studies focus on the mechanism of MR1 presentation of antigens derived from Mtb. We previously found that certain endosomal trafficking proteins played a role in MR1 antigen presentation and elucidated differences in antigen presentation pathways between intracellular Mtb infection and exogenously added mycobacterial antigens [7,8]. Specifically, we showed that pretreatment with the inhibitory ligand 6-FP boosted presentation of exogenously added antigens but had no such effect with Mtb infection [8]. Furthermore, we found that Syntaxin 4 knockdown inhibited exogenous antigen presentation by MR1 but had no effect on Mtb presentation [8]. These results implied that the MR1 antigen presentation pathway differed for Mtb compared to exogenous antigens.
In this study, we expand on our earlier work and definitively show that the mechanism of MR1 presentation of intracellular Mtb infection differs from exogenous antigen presentation in human APCs. First, we show that treatment with the L-type calcium channel blocker tetrandrine decreased both MR1- and HLA-B45-dependent presentation of Mtb. Next, we investigated endosomal calcium channels given our prior work on the importance of endosomal trafficking in MR1. Since late endosomes are rich in calcium and express TPCs [12], we targeted TPCs using N19, which inhibits NAADP-mediated calcium release from TPCs [18]. We found a significant and specific decrease in Mtb antigen presentation to MR1Ts in both human airway epithelial cells and primary human DCs. There was no effect on exogenous antigen presentation, further supporting that the mechanisms of MR1 presentation of Mtb differs from that of exogenously added antigens. It is unlikely that the mechanism of N19 inhibition of MR1-Mtb presentation stems from blocking of acidification of the endosome because N19 had no effect on the 5-A-RU prodrug, which requires an acidic compartment for cleavage into 5-A-RU [20].
Previous work indicated that TPCs are important for the phagocytic function of murine bone marrow-derived macrophages [21]. In that work, uptake of silica beads, Mycobacterium smegmatis and an attenuated form of Mycobacterium bovis (BCG) were used to measure phagocytic function. However, uptake was measured at a maximum of 90 minutes. Given those results, we performed orthogonal assays to determine whether our findings were due to impaired Mtb uptake or unexpected toxicity from N19. First, we used an HLA-B45 restricted T cell clone that detects CFP102-9 from Mtb. Results showed no significant effect of N19 on HLA-B45 antigen presentation, indicating a similar amount of internalized Mtb antigen between the groups. Additionally, we performed growth assays between the treatment and control groups and found no significant difference in the number of viable Mtb between conditions. These data argue that the results with N19 did not reflect impaired uptake or mycobacterial viability.
The results obtained with N19 were confirmed with TPC1 knockdown. We then evaluated the role of the calcium sensitive endosomal trafficking protein Syt7 and found that it is also key to MR1-dependent presentation of Mtb. Mechanistically, there are different plausible roles for TPC1 and Syt7 in MR1-dependent antigen presentation. TPC1 is expressed in different organs, including in the lungs, and colocalizes with recycling endosomes, endolysosomes and late endosomes [23,24]. Therefore, TPC1 could be responsible for local endosomal calcium release that activates Syt7. However, in HeLa cells and primary human fibroblasts, TPC1 is required for the formation of late endosome-ER contact sites [28]; both N19 and TPC1 knockdown reduced the number of contacts between late endosomes and the ER [28]. Therefore, it is possible that TPC1 mediated endosome-ER contact sites could deliver MR1 antigens to the ER. Regarding the mechanism of Syt7 in MR1-dependent antigen presentation, at least two possibilities exist. First, Syt7 functions in the early recruitment of LAMP1 to the phagosome in murine bone marrow derived macrophages [25]. In human DCs and BEAS-2B, the percentage of Mtb positive for LAMP1 was 90% and 58%-80% (depending on the time point), respectively [9,29]. Thus, it is possible that Syt7 knockdown interferes with LAMP1 delivery to the Mtb compartment and that this is necessary for MR1 presentation of Mtb. A second possibility is that Syt7 shuttles loaded MR1 to the cell surface. This is mechanism is based on Syt7 knockout mice, where DCs had decreased in MHC-II surface expression [26].
In conclusion, we have identified a novel mechanism for antigen presentation of Mtb that utilizes endosomal calcium signaling. This calcium-sensitive pathway is specific to MR1 and had no role in HLA-B or HLA-II antigen presentation. At present, specific Mtb antigens for MR1 are not known, making it difficult to track the acquisition and delivery of antigens intracellularly. Despite these limitations, our data show a pathway, independent of exogenous antigen delivery, in which endosomal calcium signaling and calcium sensitive trafficking proteins are essential for MR1-dependent presentation of Mtb.