DCs secrete BMP ligands and express type-I receptors (BMPR1a, ACVR1, and to a lesser extent BMPR1b), type-II receptors (BMPRII), and the BMP downstream signaling components SMAD1, SMAD5, SMAD8, and SMAD4 (10). Numerous evidence supports the notion that BMP signaling is involved in the process of DC maturation. Treatment of monocyte-derived (mo)DCs with BMP ligands promoted the expression of maturation markers CD80, CD83, and CD86 and promotes IL-8 secretion (10). Besides increasing the expression of co-stimulatory markers, it was also described to increase the expression of the co-inhibitory markers programmed death-ligand (PD-L1) and PD-L2 (11, 12). By inhibiting autocrine BMP signaling during moDC maturation with the small molecule BMP inhibitor dorsomorphin (DM) Martinez et al., could show that this resulted in lower expression of PD-L1 and PD-L2 without impacting the expression of other co-stimulatory markers in a process that is mediated through the IRF-1 transcription factor (11). Moreover, blockade of autocrine BMP signaling in mature moDCs resulted in stronger T cell and NK cell stimulatory ability (11). These observations could be significant for the improvement of DC-based immunotherapies.

On top of regulating DC maturation, BMPs have also been reported to influence the differentiation of different DC subsets. The addition of conditioned media from BMP4-overexpressing acute lymphoblastic leukemia (ALL) cells inhibited the differentiation of moDCs from monocytes. Furthermore, moDCs treated with BMP4 displayed increased immunosuppressive functionality by expressing higher levels of IL-10 and TGFβ (13). In mice, BMP9 was shown to promote CD8α⁺DC differentiation and inhibit plasmacytoid DC (pDC) development in a process that is mediated through ACVRL1 (14).

The most studied and documented effect of BMPs on DC differentiation is probably represented by the role of BMP7 in the context of Langerhans cell (LC) differentiation. Before this observation, TGF-β1 was believed to be the key cytokine promoting LC differentiation in vitro and in vivo. LCs can be differentiated in vitro by culturing CD34⁺ hematopoietic progenitor cells (HPCs) in the presence of GM-CSF, SCF, FLT3L, TNFα, and TGF-β1 (15), and BMP7 was shown to be able to replace TGF-β1 in this process (16). Similarly to TGF-β1, BMP7 can also promote LCs from CD1c⁺ DCs (17). The observation that LC precursors are already present in developing skin that is devoid of active TGF-β1 but expresses BMP7 further supports the recently described role of BMP7 in LC differentiation. In addition, BMP7-deficient mice exhibited lower frequencies of LCs in the epidermis and these cells appeared to possess lower amounts of dendritic processes compared to their wild-type counterparts (16). Mechanistic studies performed by inhibiting TGFβ type I receptor (TGFβR1) and overexpressing BMPR1a additionally described that LC differentiation is driven by BMPR1a activation and that TGF-β1 promotes LC differentiation by signaling through BMPR1a. Blockade of TGFβR1 during in vitro LC differentiation led to increased numbers of LCs, whereas the overexpression of BMPR1a resulted in higher percentages and numbers of LCs (16). Follow-up studies focused on characterizing better these LCs differentiated in the presence of BMP7 (BMP7-LCs). BMP7-LCs expressed the hallmark LC markers CD207, CD1a, and E-cadherin like their TGF-β1 counterpart (TGF-β1-LCs). However, a more detailed analysis revealed that BMP7-LCs are phenotypically and functionally different from TGF-β1-LCs. BMP7-LCs exhibited a CD207⁺CD1c⁺CD206⁺TLR2⁺ phenotype which resembles that of LCs in psoriatic skin lesions (18). Functionally, BMP7-LCs secrete higher amounts of pro-inflammatory and anti-inflammatory cytokines in response to bacterial stimuli and possess a greater allostimulatory capacity of CD4⁺ T cells (16). Additionally, BMP7-LCs are stronger in promoting regulatory T (Treg) cell differentiation and naïve CD4⁺ T cells cultured in the presence of BMP7-LCs secrete lower amounts of IL-22 in comparison to naïve CD4⁺ T cells cultured with TGF-β1-LCs (12).

The role of BMPR1a in murine LC differentiation in vivo has still not been elucidated. A study performed on LCs from the oral mucosa, described a two-step sequential role for BMP7 and TGF-β1 in the process of LC differentiation. BMP7, expressed in the lamina propria, promoted the recruitment of LC precursors into the mucosal epithelium where TGF-β1 could finalize their differentiation (19). An analysis performed by us in mice lacking BMPR1a in the CD11c compartment (BMPR1a^fl/flCD11c^cre mice) revealed no significant differences in LC counts in the epidermis between knockout and wild-type mice in steady-state and during inflammation (12). However, it must be taken into consideration that the CD11c-cre is a useful tool for studying dendritic cell function and homeostasis. To have a better understanding of how BMPR1a affects dendritic cell differentiation, one should also investigate the effect of BMPR1a when its expression is deleted in earlier hematopoietic progenitors.

Early studies investigating the role of BMPs in macrophages were performed on the murine macrophage cell line RAW 264•7. It was demonstrated that macrophages express ACVR1, and BMPR1a, and all three BMP type-II receptors (BMPR2, ACVR2A, and ACVR2B) (20). Functionally, treatment of RAW 264•7 cells and peritoneal macrophages with BMP6 promoted the expression of iNOS and IL-6. However, this effect was only detected with BMP6 and the same effect was not achieved with BMP2, BMP4, and BMP7 (21).

BMPs were also described to influence macrophage polarization. In murine models of prediabetic cardiomyopathy and atherosclerosis, BMP7 was observed to skew macrophage differentiation into the anti-inflammatory M2 phenotype with higher detectable levels of IL-10 and lower levels of IL-6, and TNFα (22, 23). BMP4 was also described to skew macrophage towards an M2 phenotype. In this case, its effect was investigated in human monocyte-derived macrophages and the context of tumor immunology and immunoevasion. BMP4 was detected in both conditioned media from bladder cancer cell lines and from, as previously described in the DC section, ALL cell lines. The addition of BMP4 and BMP4-containing conditioned media inhibited the pro-inflammatory M1 phenotype and promoted the M2 phenotype characterized by high IL-10 and low TNFα expression (13, 24). Pharmacological inhibition of BMP signaling in the cell lines THP-1 and RAW 264•7 differentiated towards the M2 macrophages caused distinct changes in markers associated with M1 and M2 phenotypes. This suggests a possible involvement of autocrine BMP signaling in macrophage polarization however, the obtained results were not consistent and only the M2 marker MMP12 showed a consistent decrease in the different cell lines treated with the BMP inhibitor DMH1 during M2 differentiation (25). This underlines the complexity of the BMP signaling pathway and the complexity of the polarization process of macrophages. Nevertheless, interesting effects of BMP signaling in macrophages were described particularly in the context of tumors where macrophages were described to play a role in tumor progression. These observations could in the future be exploited to develop new therapeutic targets.

T cells represent probably the immune cells in which the effects of BMPs have been the most studied. Numerous studies have focused on the role of BMP signaling in T cell activation, homeostasis, and polarization. The first investigations reported a role for BMPs in thymus organogenesis and thymocyte development. Thymic stromal cells were reported to secrete BMP2 and BMP4, and the BMP receptors and intracellular signaling components are expressed by CD4^-CD8^- double-negative (DN) thymocytes (26). Inhibition of BMP signaling in thymic epithelial cells by transgenic expression of the BMP antagonist Noggin led to a smaller thymus with a low number of thymocytes that did not descend but remained at the level of the hyoid bone. The few reported thymocytes nevertheless, were phenotypically the same as the thymocytes in healthy fully developed thymuses (27). In addition, loss of BMP4 signaling in pharyngeal epithelial cells resulted in defective thymus morphogenesis (28). This suggests that BMP signaling plays a central role in the development of thymic stroma which indirectly regulates thymocyte differentiation. BMP signaling can also regulate thymocyte differentiation by direct signaling. The addition of exogenous BMP4 to fetal thymocyte organ cultures has been shown to negatively regulate thymocyte differentiation by inhibiting their transition from DN to CD4⁺CD8⁺ double-positive (DP) (29). In vivo studies performed using conditional knock-out mice in which BMPR1a expression is selectively deleted from the hematopoietic compartment (BMPR1a^fl/flvav^cre mice) revealed that BMP signaling is required for thymocyte development and that the effects are dose-dependent. Physiological levels of BMPR1a signaling are necessary for the correct differentiation however, elevated levels of BMPR1a signaling, obtained through the addition of exogenous BMP4, produced an opposite effect characterized by inhibition of the DN to DP stage (30).

BMP signaling was also described to be involved in the activation and homeostasis of mature CD4⁺ T cells. Naïve CD4⁺CD45RA⁺ T cells express mRNA for BMPR1a, ACVR1a, and the BMP intracellular components SMAD1, SMAD5, SMAD8, and SMAD4 (31). Despite this, at a protein level naïve CD4⁺ T cells BMPR1a is poorly expressed. Following TCR activation, BMPR1a expression increases significantly alongside CD25 expression suggesting that CD4⁺ T cells can respond to BMPs following activation (12, 31). Moreover, BMP signaling influences T cell activation and function by acting on the expression of IFNγ, IL-2, and CD25. Inhibition of the BMP pathway in vitro with DM resulted in lower secretion of IL-2 and inhibition of CD25 expression (12, 31, 32). In contrast to these observations, a recent study showed that activated CD4⁺ T cells treated with BMP7 secrete lower amounts of IL-2 (33). This suggests that the role of BMP signaling in IL-2 signaling is not fully elucidated and different BMPs or the activation of the different receptors could have varying effects on IL-2 production and secretion. BMP4 and BMP7 instead were both described to reduce IFNγ expression in activated CD4⁺ T cells (33, 34).

Besides their involvement in T cell activation and homeostasis, BMP signaling is additionally involved in Treg and Th17 differentiation. In vitro studies using murine naïve CD4⁺ T cells and DM indicated that BMP signaling promotes Th17 cell differentiation (32). However, more recent studies performed with murine BMPR1a-deficient CD4⁺ T cells and by treating murine CD4⁺ T cells with exogenous BMPs revealed that BMP signaling inhibits TGF-β-mediated Th17 cell differentiation (35). Regarding Treg differentiation, it was shown that BMPs can promote Treg differentiation. In support of this, it was observed that the treatment of murine naïve CD4⁺ T cells with DM under Treg-promoting conditions inhibited CD4⁺CD25⁺FoxP3⁺ Treg differentiation (32). In addition, BMP2 and BMP4 were described to have a synergistic effect in combination with TGF-β1 in the induction of Tregs (36). Murine studies using conditional knock-out mice in which BMPR1a is deleted from the CD4⁺ T cell compartment (BMPR1a^fl/flCD4^cre mice) further confirmed the role of BMP signaling in Treg differentiation. These mice displayed lower Treg numbers and in addition, CD4⁺ T cells isolated from these knock-out mice did not efficiently convert into Tregs (37, 38). The involvement of BMPs in Treg differentiation was also observed in the human setting with BMP7 being described to promote Treg differentiation from human naïve CD4⁺ T cells (12). This overall suggests that BMPs can influence the Th17/Treg cell balance and could be useful in Th17-driven autoimmune disorders. In fact, it was already shown to have a role in limiting inflammation in the DSS-induced colitis model (35) and would be of interest in studying whether it can also have a protective effect in other inflammatory diseases.

Very few studies investigating the role of BMPs in B cells have been performed. To date, only studies reporting data from in vitro experiments are available. Both naïve and memory B cells express the BMP type-II receptor ACVR2B. However, resting naïve B cells do not express BMP type-I receptors. They gain expression of ACVR1 following stimulation with CD40L/IL21. BMP2, BMP4, BMP6, and BMP7 were all demonstrated to have an inhibiting effect on CD40L/IL21-induced immunoglobulin (Ig) production. BMPs inhibited the production of IgM, IgA, and IgG by both naïve and memory B cells and also inhibited the differentiation of B cells into CD27⁺CD38⁺ plasmablasts (39). The effect of BMP7 was mainly due to its ability to promote apoptosis in the cells suggesting differential effects between BMP2/BMP4/BMP6, and BMP7 (40).

ILCs is one of the newest described family of immune cells. BMP signaling has also been implicated in ILC function and differentiation. Group 2 ILCs have been described to express BMPR1a, and BMPR1b (41) and to secrete BMP2 and BMP7 (42). Not much is known about the role of BMPs secreted from ILCs however, ILC-derived BMPs were suggested to regulate adipogenesis (41).

NK cells have been recently classified in the ILC family and represent the most studied ILC subset in terms of BMP signaling with studies pointing towards a role for BMPs in NK cell differentiation. Early CD34⁺BMPR1a⁺ thymic precursors express typical transcription factors (NFIL3, ID3) and cell surface markers (IL-2Rβ, IL-15RA, CD161) which are associated with NK cells (43–45). Reconstitution of thymic lobes from fetal SCID mice with CD34⁺BMPR1a⁺ precursors gave rise to high amounts of functional NK cells (about 60% CD3^-CD161⁺CD56⁺ cells), whereas their reconstitution with CD34⁺BMPR1a^- counterparts failed to reach 10% CD3^-CD161⁺CD56⁺. Moreover, in vitro differentiation experiments with the BMP inhibitor DM clarified that during their differentiation process NK cells produce BMP4 and that autocrine BMP signaling is involved in their differentiation (46). In addition to NK cell differentiation, the BMP signaling pathway also contributes to NK cell function. Blood circulating mature NK cells secrete BMPs (BMP2 and BMP6), express type I and type II receptors, and the BMP downstream signaling components SMAD1, SMAD5, SMAD8. Autocrine BMP signaling in NK cells was described to be important for optimal NK cell function by promoting IFNγ secretion, activation markers (CD69 and NKp46), and their ability to activate autologous DCs (47).