Sarcoidosis is a systemic granulomatous disease of unknown cause affecting the lung with a prevalence close to 90% (1, 2). Granulomas, the key lesion in sarcoidosis, are well circumscribed and coalescent clusters of macrophage-derived epithelioid and giant cells, associated with lymphocytes, subsequently wrapped by lamellar fibrosis (3). The prevailing hypothesis explaining granuloma formation is that exposure to unknown antigens triggers a disproportionate inflammatory response in genetically predisposed individuals (4). Sarcoidosis is particularly challenging in regards of identifying factors involved in initiation of inflammation, granuloma perpetuation and evolution towards fibrosis as observed mostly in pulmonary, but also in hepatic, renal or cardiac involvement (3, 5–7). These events are determinant for sarcoidosis outcomes which vary from benign to life-threatening especially due to pulmonary fibrosis and/or pulmonary hypertension (8).

Macrophages are pivotal cells in the sequence of events associated with sarcoidosis pathogenesis, granulomas having been clearly demonstrated constituted of macrophages-derived cells (9–13). Different types of macrophage populations (alveolar and interstitial) derivating from distinct precursors of embryonic/fetal and myeloid origin, reside in the lung (14, 15). Notably, upon injury, the myeloid precursors generate monocyte-derived (MD) macrophages, though to play a major role in fibrosis (16) and sarcoidosis (11, 17–20).

In the lung of patients, these cells are an important source of TNFα, a major cytokine in the pathogenesis of sarcoidosis (11). Circulating mononuclear cells from sarcoidosis patients are also able to reconstitute granulomas in vitro, and present a specific transcriptional profile compared to other granulomatosis such as tuberculosis (17, 18). In addition, analysis of granulomas from sarcoidosis recurrence after lung transplantation demonstrated that constitutive cells derived from recipient blood cells (19, 20), indicating that peripheral blood cells are major contributors in granuloma formation.

Hypoxia, defined as an imbalance between impaired tissue oxygen supply and cell demand, is a micro-environmental condition known to modulate innate and adaptive immunity (21) by promoting pro-inflammatory response (22) and to contribute to the pathophysiology of fibrotic diseases (23). Hypoxia impact on macrophages is particularly relevant to study in sarcoidosis as a previous study showed that the absence of in-depth vascularization makes sarcoidosis granulomas hypoxic (24) as for tuberculous granulomas (25). Moreover, from inflammatory tissues to granulomas, MD-macrophages may be exposed to low oxygen tension environments mainly due to O₂ consumption related to hypermetabolism of inflammatory cells (26). The cellular response to hypoxia is mostly controlled by the hypoxia-inducible factor (HIF) transcription factor and its target genes harboring the hypoxia-response element (27). Several studies in sarcoidosis reported the expression of HIF-target genes within lung and lymph node granulomas, while HIF-1α expression was inconsistently found (28, 29). Hypoxia and the HIF-1 signaling impact on the processes controlling granuloma evolution are still unclear and their effects on sarcoidosis immune cells have never been addressed.

We hypothesized that hypoxia may induce macrophage functional changes with a potential impact on the course of the disease. Therefore, we studied the consequences of hypoxia on a set of main functions of MD-macrophages from sarcoidosis patients and healthy controls. We show that hypoxia had a significant impact on MD-macrophages from sarcoidosis patients, with the strongest effect seen in patients with high active disease. Therefore, hypoxia could play a significant role in sarcoidosis pathogenesis and fibrosis outcome.

Additional details are provided in the Online Supplement .

Factors modulating the onset and progression of sarcoidosis are still poorly known. Here we showed that hypoxia, a microenvironmental factor most likely present within granulomas and inflammatory tissues, activated HIF-1 and induced a mixed proinflammatory-profibrotic MD-macrophage phenotype particularly marked in AS, as illustrated in Figure 8 . These in vitro results were supported by immunohistochemistry data showing the expression of HIF-1α and its target PAI-1 in epithelioid cells constituting pulmonary sarcoidosis granulomas.

One strength of our study was to prospectively include patients with either high active or low active or inactive sarcoidosis, not taking any therapy that could modify macrophage biology. The aCTAS score we used to evaluate the disease activity has already been validated by two independent studies (32, 43). This score was highly correlated with the range of FVC improvement under further sarcoidosis treatment, a good marker of the lesions reversibility. The choice to specifically study MD-macrophages in these patients was supported by recent works showing that in vitro models of granuloma using PBMCs can mimic sarcoidosis events such as phagosome-regulated mTOR or IL-13 signaling (17, 44). In addition, previous studies reported that immune cells constituting recurrent sarcoidosis granulomas in lung transplant allografts originated from the recipient monocyte-macrophage lineage (19, 20). Finally, the inflammatory phenotype of MD-macrophages in pulmonary sarcoidosis and their propensity to produce TNFα has been recently highlighted (11), supporting the role of these cells in sarcoidosis pathogenesis. However, the fact that experiments were almost exclusively performed in MD-macrophages constitutes a limitation of the present study. It would certainly be interesting to study the effects of hypoxia on other cell types involved in sarcoidosis such as T lymphocytes. A comparison of our data in MD-macrophages with data obtained in lung resident alveolar macrophages could also be informative but bronchoalveolar lavages are rarely done in the follow-up of sarcoidosis patients. Another limitation of our study is the fact that it was not possible to conduct all experiments for each patient for technical reasons, i.e. relatively to the low number of retrieved monocytes per blood sample.

To the best of our knowledge, the consequences of a hypoxic microenvironment on macrophages were never investigated in sarcoidosis, although granuloma hypoxia was inferred from morphometric analysis (24). By contrast, in tuberculosis pulmonary granulomas, hypoxia was directly demonstrated to play a major role in the disease course (25, 45). Recently, 18F-fluoromisonidazole uptake, a hypoxia-sensitive PET tracer, was observed in sarcoidosis lesions and associated with ^18FFDG-PET uptake (35), suggesting that hypoxic lesions are metabolically active and associated with an active form of sarcoidosis. In addition, patients with high active sarcoidosis had more impaired lung function with lower DLCO, potentially resulting in local alveolar hypoxia.

Previous studies investigating HIF in sarcoidosis led to conflicting results (28, 29). Tzouvelekis et al. (29) reported increased expression of VEGF, a target of HIF, within sarcoidosis granulomas but failed to detect HIF-1α whereas Talreja et al. did (28). Such a difference in HIF-1α expression detection could be explained by different stages in the disease between the studies, the different antibodies used for HIF-1α detection and immunosuppressive treatments given to patients in the Tzouvelekis et al. study (28, 29). However, in line with Talreja et al. (28), we clearly immunodetected HIF-1α in pulmonary sarcoidosis granulomas. Such a finding prompted us to evaluate in vitro the effects of hypoxia on MD-macrophages from controls and untreated sarcoidosis patients.

In the present study HIF-1α nuclear expression was detected under normoxia in MD-macrophages from patients and controls. Talreja et al. also reported HIF-1α expression in freshly-isolated blood monocytes (28). The nuclear expression of HIF-1α in normoxic MD-macrophages is consistent with the fact that non-hypoxic stimuli such as TNF-α and IL-1β proinflammatory cytokines, NF-κB, reactive oxygen species, MAPK (46, 47) or mTOR pathway activation (48) are able to stabilize HIF-1α in immune cells. However, HIF-1α transcriptional activity was significantly more strongly induced by hypoxia in MD-macrophages from AS. Consistently, several HIF-target genes were induced by hypoxia in MD-macrophages from AS, such as VEGF and TGFB1 transcripts or PAI-1 protein. The TNFα and IL-1ß cytokines [involved in sarcoidosis pathogenesis (4, 49)] and CXCL8, induced by hypoxia only in AS, can also be directly upregulated by HIF in macrophages (50, 51). NF-κB activation, known to be also induced by hypoxia and leading to pro-inflammatory cytokine production (52) was decreased in AS, suggesting that it was not involved in this context. Moreover, HIF-1 can exert a negative feedback on NF-κB (53). The stronger response to hypoxia observed in MD-macrophages from AS is somewhat intriguing. It might be explained by still unexplored genetic polymorphisms, by epigenetic mechanisms and/or non-coding RNAs known to modulate HIF expression (54).

We investigated whether hypoxia could modulate phagocytosis and antigen processing/presentation, a key initial step of inflammatory lesions in sarcoidosis. We observed that phagocytosis was similar at baseline in control and sarcoidosis MD-macrophages. Interestingly, hypoxia markedly reduced phagocytosis in controls but not in sarcoidosis, suggesting that macrophage phagocytosis could be maintained under hypoxia in this disease. Previous papers reported that phagocytic activity of macrophages was increased in sarcoidosis and in an in vitro model of granuloma (17, 40, 41). An enhanced expression of phagocytosis-related genes, but a downregulation of genes involved in proteasome degradation were also observed, suggesting that the accumulation of intracellular phagocytic degradation products could participate in the chronicity of inflammation in sarcoidosis (40). We next evaluated the effect of hypoxia on the expression of CD80 and CD86 co-stimulatory molecules and on HLA-DR as these markers are associated with antigen presentation and expressed by epithelioid cells in sarcoidosis granuloma (55, 56). Hypoxia markedly decreased the expression of CD80 and CD86 on MD-macrophages in controls and AS while not significantly in IS. Hypoxia also downregulated the expression of HLA-DR in AS and IS MD-macrophages. These findings are in line with previous studies showing that hypoxia decreased the expression of CD80 in murine macrophages and CD80, CD86 and MHC class II in human dendritic cells, reducing their ability to initiate adaptive-immunity responses (57, 58). However, the role of antigenic presentation in sarcoidosis is still debated. Crouser et al. reported an enhanced presentation capacity in an in vitro model of sarcoidosis (17). Conversely, Grunewald et al. proposed that granuloma progression may occur when antigen recognition is not efficient enough, due to inadequate peptide presentation by modified HLA molecules or T cells incapacity to generate T cell clones, allowing granuloma persistence because of inefficient adaptive immune response (4). Here, we can speculate that hypoxia at the various steps towards granuloma formation may potentially exert a deleterious effect on antigen presentation processes, and therefore contribute to their persistence.

Mechanisms leading to sarcoidosis-associated pulmonary fibrosis are poorly understood. Granulomas are almost always surrounded by a concentric rim of collagen bundles (3). Macrophages are known to be involved in repair and fibrotic processes through different states of polarization (59). In sarcoidosis, macrophages may acquire a profibrotic M2-phenotype as shown in granulomas from patients’ biopsies or in vitro model (44, 60). In our study, hypoxia decreased the expression of M1 markers (CD80, CD86, HLA-DR), while M2 markers expression was maintained (CD163, CD36). However, hypoxia was also associated with a M1 pro-inflammatory cytokine response. In the literature, the effect of hypoxia on macrophage polarization remains unclear (61, 62). More than the over-simplified M1/M2 dichotomy, macrophages display remarkable plasticity and can change their phenotype in response to environmental factors.

Previous studies showed that under hypoxic conditions, macrophages can induce fibrosis in a HIF-dependent manner through the secretion of VEGF, PDGF-BB, or PAI-1 (63, 64). Here, we observed that MD-macrophages from AS secreted higher levels of PAI-1 in response to hypoxia as compared with controls or IS. The PAI-1 factor plays a key role in the development of pulmonary fibrosis by fibrotic matrix deposition (65). An increase in PAI-1 levels as well as a decrease in fibrinolytic activity has been reported in the bronchoalveolar lavage of patients with sarcoidosis (66). We clearly detected PAI-1 in granuloma epithelioid cells. Finally, we demonstrated that CM from hypoxic MD-macrophages inhibited pulmonary fibroblast migration, a phenomenon highly dependent on PAI-1. Consistently, PAI-1 binding to vitronectin has been previously shown to inhibit cell migration, leading to inefficient alveolar repair following injury and favoring the development of pulmonary fibrosis (67). Therefore, we hypothesize that PAI-1 secretion by granuloma cells could immobilize and sequester lung fibroblasts around granuloma, thus favoring the characteristic peripheral granuloma fibrosis.

In conclusion, this study shows that hypoxia exerts a significant and specific impact on MD-macrophages from sarcoidosis patients, with the strongest effect observed in patients with a high active disease. It may favor the development and persistence of granulomas in active sarcoidosis and fibrosis surrounding granulomas by promoting a mixed inflammatory/fibrosing response of macrophages, by reducing their antigen presentation capacities, leading to a deficient T cell response. Thus, the HIF pathway and PAI-1 could be involved in the pathogenesis of high active sarcoidosis, potentially representing new therapeutic targets. As a future direction, we plan to investigate interactions between HIF and other signaling pathways identified in the pathogenesis of sarcoidosis as mTOR, NLRP3 inflammasome, JAK/STAT, or heat shock proteins (49, 68, 69), already known to interfere with the HIF/hypoxia pathway (22, 70–72).

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Comité de protection des personnes (CPP) Ile-de-France X 2016-10-02. The patients/participants provided their written informed consent to participate in this study.

FJ and VB designed research studies, contributed to the experiments, analyzed the data, and wrote the manuscript. HN and MK provided human samples and reviewed the manuscript. MP contributed to the experiments and reviewed the manuscript. CP, J-FB, DV, and HN wrote the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by: Fonds de Recherche en Santé Respiratoire et Fondation du Souffle (2017), Financement de la Chancellerie des Universités de Paris-Legs Poix (2017), Bonus Qualité Recherche Université Sorbonne Paris- Nord (2016-2017).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.