Jürg Gertsch¹

• inflammation
• cannabinoids
• endocannabinoid system

The majority of organs in the body contains tissue-resident macrophage populations that are distinctly activated and polarized, depending on their microenvironment. Tissue macrophages not only phagocyte microorganisms and serve as APCs or to clear cell debris but are also involved in numerous biologic and pathophysiological conditions, including chronic inflammation, tissue remodeling, and cancer [1]. In the lung, resident macrophages are important for the innate and adaptive immune responses. These macrophages primarily consist of PAMs (>95% in bronchoalveolar lavage), originating from yolk sac erythro-myeloid progenitors and fetal liver monocytes, and to a minor extent, pleural, interstitial, and intravascular macrophages [2]. Moreover, the lung harbors DCs, which share markers and some functionality with alveolar macrophages. Monocyte-derived or classic DCs have myeloid precursors present in the bone marrow. DCs have inferior phagocytic activity and superior antigen-processing capacity. The lung is constantly exposed to potentially noxious airborne stimuli, such as foreign particles and invading microorganisms, which are sensed by lung-resident macrophages.

PAM activation and the initiation of inflammation involve a complex interplay between activating and repressing signals [2]. The innate immune response is characterized by activation of TLRs through pathogen-associated molecular patterns, IFN-γ or TNF-α, and the subsequent activation of transcription factors, such as NF-κB and the AP-1. This typically leads to macrophage activation and expression of proinflammatory cytokines, such as IL-1β, IL-6, IL-12, TNF-α, and TGF-β. Numerous receptors and signaling molecules have been reported in this context [2], but apart from PGs (e.g., PGE2), leukotrienes, and lipoxins, relatively little is known about the role of immunomodulatory lipids produced by PAMs and macrophages in general. The evolution of the innate immune system is paralleled by the evolution of the ECS, a lipid signaling network comprising cannabinoid receptors and the enzymes involved in the generation and metabolism of the arachidonic acid-derived endocannabinoids, which were discovered, thanks to the groundbreaking research on THC from cannabis [3]. Monocytes/macrophages are able to express all of the components of the ECS and apparently, integrate stimuli generated via CB1 and CB2 receptors. However, it remains unclear what exactly the roles of the ECS in monocytes/macrophages are, whether they depend on the activation state of the macrophages, and whether the differential pharmacological effects of cannabinoids related to inflammation, fibrosis, and cancer reported could be attributed to their action on macrophages.

In this issue of the Journal of Leukocyte Biology, Staiano et al. [4] convincingly show that human lung resident macrophages constitutively produce endocannabinoids and N-acylethanolamines. Endocannabinoids exert protective (prohomeostatic) effects in acute and chronic inflammation, primarily through CB2 receptors [5]. They show that ex vivo LPS stimulation of PAMs leads to a significant intracellular calcium-dependent release of the major endocannabinoid 2-AG, which is the endogenous, full CB2 receptor agonist. PAMs seem to express all major proteins of the ECS. CB1 and CB2 receptor activation significantly inhibited the release of key angiogenic (Ang1 and -2) and lymphangiogenic factors (VEGF A and C) from PAMs, a finding with translational implications. Moreover, CB1 and CB2 receptor agonists induced ROS and ERK1/2 phosphorylation. The release of endocannabinoids by PAMs suggests endocrine or exocrine roles of these lipids. The authors also identify CB receptors in macrophages resident in human lung cancer tissue, indicating a potential role for the ECS in TAMs. With the use of similar protocols, human CD14⁺ MDMs showed a different response pattern to cannabinoids, thus highlighting the importance of the macrophage ontogeny and activation state. As expected from any important finding, the present study raises numerous questions.

Smoking tobacco impairs the function of PAMs, as smoking is the major risk factor for COPD, which is currently the fourth leading cause of death globally. Smoking cannabis (marihuana, hashish) is the most widespread illicit form of consuming cannabinoids recreationally and has been suggested by different studies to be distinct from smoking tobacco. However, there is a striking disagreement between the published data and study designs, and it remains largely unclear which cannabinoids are good or bad for the lung or whether it is context dependent. Moreover, the role of ECS in the lung remains poorly understood.

The first question that arises is whether the lung-resident macrophages (here, simply referred to as PAMs) and the peripheral MDMs studied are homogenous or mixed populations that uniformly respond to the culture conditions. To study the differential responses to cannabinoids in distinct macrophage phenotypes, phenotypic markers, such as CD11b, CD11c, CD68, F4/80, mannose receptor, and sialic acid-binding Ig-type lectin F, help to characterize macrophage populations better. The lung-resident macrophages studied by Staiano et al. [4] were characterized as FSchiSSchiCD45⁺HLA-DR⁺ cells, which homogeneously expressed CD206, indicative of PAMs or a lung-resident intermediate population. CD11c expression would unambiguously categorize these cells as PAMs and DCs. Activation and polarization of macrophages (in vivo and in vitro) are poorly standardized, and the nomenclature and basic experimental guidelines have only recently been revised [6]. For MDMs, IL-1 and IFN-γ treatment of CSF-1-generated macrophages induces a defined and comprehensive population, reflecting the M1 activation state, according to the M1/M2 paradigm. On the other hand, IL-4 and IL-10 triggers an alternative M2 activation state. CSF-1 alone might favor the M2 state. Prolonged LPS stimulation will eventually favor a M1 phenotype, but LPS, together with PGE2, can also generate a M2-like state, expressing high levels of IL-10. Thus, the activation state of macrophages is confused by their outrageous plasticity, which is a hallmark of phagocytes.

In different studies, it has been convincingly shown that CB2 receptor agonists can inhibit the production of TNF-α and other proinflammatory factors from primary CD14⁺monocytes and possibly M1 macrophages, increasing the expression of IL-10. Likewise, CB2 agonists reduce the inflammatory damage in rodent models of sepsis. The suppression of the M1 activation state or general interference with macrophage activation programs might explain why in fibrotic processes, CB2 agonists are strongly protective [5]. This is currently best exemplified in the liver through the action of CB2 agonists in Kupffer cells [7]. Moreover, endocannabinoids drive the acquisition of a M2 state in microglial cells via CB2 receptors. Paradoxically, some data suggest that CB1 receptor activation leads to increased TNF-α and cyclooxygenase 2 expression in certain macrophage populations, which is in agreement with the emerging evidence that CB1 overactivation increases fibrosis in liver and kidney (vide infra). However, these mechanisms are far from being understood. In the study by Staiano et al. [4], both CB1 and CB2 receptors act in harmony, indicating that the role of the ECS depends on the polarization state of the macrophages and that the role of CB1 receptors can be switched (Fig. 1).

View larger version:

• In this page
• In a new window

• Download as PowerPoint Slide

Figure 1. Hypothetical model of the involvement of CB receptors in the activation of macrophages.

The arachidonic acid-derived endocannabinoid 2-AG, by activating CB2 receptors and possibly CB1 receptors in different monocyte/macrophage activation states (M1, M2, and TAM), may exert different modulatory effects. It remains unknown whether cannabinoids can reprogram M2 or TAM states.

The second question is whether the inhibition of VEGF and angiopoietin from PAMs is a good thing or not. PAMs, together with DCs, are responsible for protecting the lung against infectious microorganisms but have also been associated with inflammatory and fibrotic processes [2]. PAMs can release VEGF, necessary to control allergic airway inflammation in asthmatic conditions and the clearance of apoptotic cells. In the study by Staiano et al. [4], the human PAM population produced LPS-stimulated TNF-α that was not modulated by CB receptor agonists, which is unexpected. Yet, this might be beneficial, as TNF-α is involved in the resolution of established pulmonary fibrosis via a mechanism involving reduced numbers of profibrotic M2-like macrophages. The blocking of VEGF under normal conditions could favor persistent airway inflammation, but this may depend on the level of TNF-α and other factors. PAMs have also been found to contribute to angiogenesis and tumor growth via the secretion of IL-8 and VEGF. Intriguingly Staiano et al. [4] showed that LPS-stimulated VEGF is inhibited in PAMs but not in MDMs, again pointing toward macrophage plasticity or origin (yolk sac-derived erythro-myeloid progenitors vs. adult hematopoietic stem cell derived). To complicate the matter, both M1 and M2 macrophages can produce VEGF. IL-10 has been reported to inhibit only VEGF in M1 macrophages [8]. As all primary monocyte/macrophage populations robustly express CB2 receptors and 2-AG, the potentially autocrine mechanism uncovered in this study might reflect a fundamental biologic CB receptor-mediated modulatory mechanism of macrophage activation and programming (Fig. 1), which should be studied in more detail.

This leads us to the question of whether cannabinoids might target TAMs and whether this could be a reasonable strategy to reprogram M2 tissue macrophages or attenuate TAM function. In cancer cells, cannabinoids generally block the activation of the VEGF pathway [9]. VEGF and the active forms of its main receptors (VEGFR1 and VEGFR2) are down-regulated on cannabinoid treatment in animal models of cancer [9]. However, so far, little attention has been drawn to the ECS in TAMs. To inhibit or even reprogram TAMs specifically has become a holy grail in recent anticancer strategies. PAMs contribute to lung cancer progression and metastasis through a M2 activation state, which was apparently not induced by cannabinoids in the study by Staiano et al. [4], as TNF-α expression was not inhibited. TAMs are derived from immature or specific subsets of monocytic progenitor cells (Fig. 1), with the spleen as an extramedullary site of their origin. How monocytes/macrophages become TAMs remains largely unclear. Yet, TAMs affect virtually all aspects of tumorigenesis, including angiogenesis, invasion, and metastasis. One exciting finding by Staiano et al. [4] is the fact that CB receptor activation leads to ROS production and ERK1/2 phosphorylation in human PAMs. The roles of ROS in PAMs and MDMs might be fundamentally different. In human macrophages derived from PMA-stimulated monocytes, ROS production has been shown to be differentially modulated by CB1 and CB2 receptor agonists, and the generation of TAMs from monocytes has been shown to require ROS [10]. To shed light on the differential effects of CB receptor ligands, we are forced to standardize our experiments with macrophages [6]. It would be interesting to study the effect of CB2 receptor-selective agonists on well-defined M2-like macrophages or TAMs, as such molecules could have multiple protective effects on different macrophage subsets. CB2 receptor-selective agonists are expected not only to be antifibrotic but also could target TAMs (Fig. 1), although this remains speculative at this point. Nevertheless, various reports have shown that cannabinoids can reduce tumor growth and progression in animal models of cancer [9], and therefore, the modulation of TAMs by specific CB receptor ligands or modulators of the ECS needs to be studied in more detail.

In returning to the epidemiologic data from cannabis smokers, habitual smoking of cannabis is associated with abnormalities of PAMs, including impairment in microbial phagocytosis associated with defective production of immunostimulatory cytokines and NO, thereby potentially predisposing to pulmonary infection. However, cannabis smoking does not seem to contribute to the development of COPD beyond the effects of tobacco. Moreover, a recent meta-analysis by members of the International Lung Cancer Consortium concluded that there is little evidence for an increased risk for lung cancer among habitual or long-term cannabis smokers [11] beyond the general high risk for lung cancer for smokers. Importantly, cannabis preparations vary and might have effects different from purified cannabinoids. For instance, THC pharmacologically acts as an agonist at CB1 and antagonist (weak partial agonist) at CB2 receptors, but certain cannabis strains produce more CB2 receptor-selective cannabinoids than others.

As nicely demonstrated in the paper by Staiano et al. [4], cannabinoid receptors have different roles in different macrophage subtypes. Thus, it will be important to explore the potential of differential CB1 and CB2 receptor-dependent activation/blockage in well-characterized macrophage populations from different disease states, including acute inflammation, chronic inflammation, and tumor progression. In light of the promising data emerging from cannabinoid-based preclinical and clinical studies on inflammatory pain, neuropsychiatric diseases (microglial cells), and also cancer, macrophages could be essential mediators of some of the therapeutic effects observed with cannabinoids.

• Received September 11, 2015.
• Revision received October 21, 2015.
• Accepted October 22, 2015.

• © Society for Leukocyte Biology