Phagosome-regulated mTOR signalling during sarcoidosis granuloma biogenesis

Discussion

Despite morphological similarities of the granuloma-like cell aggregates formed by human PBMCs in response to M. tuberculosis antigens in vitro, the molecular profile of sarcoidosis granulomas differs from LTBI granulomas based upon gene expression analysis and is reflective of fundamentally different macrophage features. In particular, sarcoidosis granulomas consist of macrophages featuring an alternatively activated CD163 phenotype [14], exhibiting abnormal sustained phagolysosome activation, predicted to promote enhanced antigen processing and presentation. Our findings are corroborated by prior reports showing enhanced macrophage antigen presentation by sarcoidosis macrophages [15, 16]. Our data provide evidence that phagolysosomes promote signal transduction via mTORc1 (figure 3), which has been shown to be essential for early sarcoidosis granuloma formation. Our findings expand upon recent reports incriminating aberrant mTORc1 signalling in the pathogenesis of a sarcoidosis-like, systemic granulomatous disorder in mice and in humans with progressive sarcoidosis disease [8].

The findings in the in vitro human granuloma model seem to contradict prior studies showing similar gene profiles of PBMCs in sarcoidosis and TB [17, 18]. However, unstimulated PBMCs do not form granulomas, and we have previously reported that unstimulated PBMCs of sarcoidosis and disease-free controls have similar molecular signatures [12]. We have also shown that extracellular cytokine release from unstimulated PBMCs is minimal and does not differ when comparing unstimulated PBMCs from sarcoidosis to those of LTBI or controls [9]. However, there are interesting differences in extracellular cytokine release following stimulation with M. tuberculosis antigens, as per figure 6, and conditions promoting fusion of endosomes with lysosomes to yield acidified phagolysosomes are shown herein to enhance mTORc1 signalling, which aligns with recent investigations by Linke et al. [8].

The molecular mechanisms by which phagolysosomes promote mTORc1 signalling are well established and are implicated by our genomic analysis of sarcoidosis granulomas. Whereas mTORc1 is also expressed in T-cells, our data indicate that mTORc1 signalling in macrophages is of primary relevance during early granuloma formation given that granuloma-like cell aggregates are primarily comprised of macrophages, and disruption of phagolysosome formation or function (acidification) attenuates granuloma formation as effectively as inhibition of mTORc1. In macrophages, uptake of microbial-associated antigens by endosomes to form phagosomes is followed by fusion with lysosomes, mediated by Rab GTPases, including Rab5 and Rab7 [19]. As indicated in figure 2 and supplementary table S1, Rab5 and Rab7 genes are highly expressed in sarcoidosis granulomas compared to LTBI granulomas forming in response to M. tuberculosis antigens. The next step in phagosome maturation is progressive acidification mediated by vATPase enzymes derived from lysosomes [20]. Sarcoidosis granulomas exhibited higher expression of multiple vATPase transcripts (figure 2, supplementary table S1). Finally, a specific vATPase enzyme encoded by ATP6V1A, also highly expressed in sarcoidosis granulomas (six-fold higher compared to LTBI; p<0.05) forms a complex with mTORc1 to promote mTORc1 signalling via S6 kinase (S6K) [10]. Once activated, mTORc1/S6K promotes alternative macrophage polarisation characterised by increased expression of CD163 [21], such as documented in these experiments (CD163 expression is 117-fold higher in sarcoidosis compared to LTBI granulomas; p<0.001) (supplementary table S1). Reduced expression of mTORc1 is a known feature of primary human macrophages from healthy donors in response to M. tuberculosis and has implications for the control of infection [22]. Thus, enhanced mTORc1 signalling in sarcoidosis suggests that the sarcoidosis immune phenotype may confer protection against M. tuberculosis infection. STAT3, activated downstream of mTORc1 [23], is incriminated in the pathogenesis of sarcoidosis granulomas [24] and was shown to be rapidly activated in response to M. tuberculosis antigens and suppressed in response to the mTORc1 inhibitor, rapamycin, in this in vitro model (figure 3). Together, our data support the role of mTORc1/S6/STAT3 pathway activation during the genesis of sarcoidosis granulomatous cell aggregates.

Sarcoidosis granuloma macrophages exhibit abnormally enhanced and sustained antigen processing. The normal progression from early microbe recognition through the late phagosome maturation stages typically requires ∼10–30 min, whereas phagolysosomes may be sustained to promote microbial killing [25]. As such, it is noteworthy that the genes regulating all stages of microbe recognition, phagosome-lysosome-mediated intracellular killing and antigen presentation, remained highly expressed in the sarcoidosis PBMC granulomatous cell aggregates 7 days after a single exposure to immunogenic M. tuberculosis antigens in the form of PPD-coated beads. These findings are consistent with those of Spiteri et al. [26] who reported increased macrophage phagocytic activity in bronchoalveolar lavage (BAL) macrophages obtained from patients with active pulmonary sarcoidosis, as well as studies by Lem et al. [27] showing enhanced antigen presentation capacity of sarcoidosis BAL macrophages.

Given that the study patients had no evidence of active or latent TB infection, these experiments support the notion that sarcoidosis patients are predisposed to an exaggerated antimicrobial response to TB antigens. These findings lead us to speculate that sarcoidosis patients are genetically predisposed, when confronted with a critical microbial challenge, to respond with a potent bactericidal granulomatous response aimed at intracellular pathogens at the expense of unremitting inflammation. This line of reasoning would explain well-documented cases of sarcoidosis following the resolution of infections caused by intracellular pathogens, most notably M. tuberculosis, Propionibacterium acnes [28] and Histoplasma capsulatum [29], and explains why immune suppression therapy typically improves sarcoidosis manifestations and rarely results in disseminated infection. Future studies are planned to determine whether the exaggerated macrophage-mediated immune response of sarcoidosis patients is M. tuberculosis-specific or if other pathogens (e.g. P. acnes, H. capsulatum) also promote granuloma formation.

The uptake of antigens, presumably mediated by scavenger receptors such as CD163 or CD206, is linked to granuloma formation through the formation of phagolysosomes. Interruption of any of the steps required for phagolysosome formation and acidification was shown to disrupt mTORc1 signalling and suppress granulomatous cell aggregate formation (figures 4 and 5) and inflammatory pathways in sarcoidosis (figure 6). CLQ and NH₄Cl prevent lysosome fusion with endosomes [30, 31]; BAFA1 inhibits vATPase-mediated acidification [32]; and RAP binds to a regulatory component of mTORc1 to suppress its activation [33]. The paradoxical finding of near-complete inhibition of granuloma formation in the setting of unchanged or enhanced TNF-α release by the vATPase inhibitor is probably related to the induction of NF-κB signalling, as reported previously [14]. However, vATPase inhibitors strongly suppressed IFN-γ production, which is vital for granuloma formation based upon prior studies of M. tuberculosis-induced granuloma formation [34, 35]. Finally, the lack of granuloma inhibition by NH₄Cl is difficult to explain, because it should suppress phagolysosome-mediated signalling by preventing fusion of lysosomes, thereby increasing pH, as per vATPase inhibitors. In this regard, it is interesting to note that NH₄Cl suppresses lysosome–phagosome fusion, but has been shown to concurrently induce phagosome–endosome fusion in the setting of mycobacterial infection [36], suggesting that the fusion of phagosome with endosome is sufficient to promote granuloma formation. Regardless of the mechanisms by which NH₄Cl influences phagosomes, compared to CLQ, NH₄Cl treatment less effectively suppressed IFN-γ, TNF-α and IL-1β, which may further explain the discrepant results [37–39].

Although the roles of mTORc1 signalling and various cytokine and chemokines during the pathogenesis of sarcoidosis appear to be complex, our data support the promotion of early granuloma formation by M. tuberculosis antigens via phagolysosome-mediated mTORc1/S6/STAT3 signalling. However, a number of important questions follow from these findings. For instance, are other signalling pathways working in concert with mTORc1 during sarcoidosis granuloma formation; are sarcoidosis macrophage populations homogeneous in terms of mTORc1 activation and related overactive phagosomal functions; and can microbial antigens other than those derived from M. tuberculosis also promote sarcoidosis granuloma formation? Furthermore, how does the cytokine milieu influence granuloma formation? In concert with IFN-γ, it is likely that TNF-α and other cytokines, such as IL-13 [12] and IL-17 [40] can amplify granuloma formation. Whereas IL-10 and transforming growth factor-β are reported to suppress granuloma formation in models of M. tuberculosis infection [41], the role of these cytokines in sarcoidosis remains unclear. The full consequences of dysregulated mTORc1 signalling as they relate to antigen processing require further investigation. For example, enhanced mTORc1 and STAT3 signalling have implications for the effective clearance of antigens via inhibition of autophagy [42], which may contribute to sustained granuloma formation, as proposed by Pacheco et al. [43].