Both CCR5 and CXCR4 are G-protein coupled, 7-transmembrane receptors (GPCR) that belong to a family of 20 proteins known as chemokine receptors. These receptors are expressed on a wide range of cell types and are classified into four subgroups depending on which subfamilies of chemokines they bind: CXC, CC, XC, and CX3C receptors [28, 30, 183]. While chemokines are known to act as chemoattractants, chemokine receptor activation also mediates a broad array of physiological processes, including cell migration, cytoskeletal rearrangement, cell survival, and inflammation [28, 184]. These receptors are often upregulated in response to tissue damage or in diseases such as cancer and HIV, attracting immune cells to respond to the damage or insult [30, 185, 186]. While many different chemokines and chemokine receptors play a role in HIV neuropathogenesis [183, 187, 188], CCR5 and CXCR4 are considered the major co-receptors used by HIV-1 strains and are thought to mediate many of the resulting chemokine responses [189, 190], making them appealing targets for antiretroviral development.
Like other GPCRs, chemokine receptors transmit signals resulting from ligand binding via an associated complex of three distinct G-proteins; Gα, Gβ, and Gγ subunits [191]. In the inactive state, GPCRs are coupled to a trimer of these G-proteins, one of each type. Ligand binding initiates a conformational shift in the GPCR, activating the Gα subunit by exchanging a bound guanine diphosphate (GDP) for a guanine triphosphate (GTP). This causes the Gα subunit to dissociate from the receptor to initiate distinct downstream signaling cascades [192]. There are four distinct subtypes of Gα proteins: Gαs, Gαi, Gαq, and Gα12/13, each associated with distinct signaling cascades, though distinct G-proteins can activate the same downstream effectors or have overlapping functions [191]. The signal transduction pathways initiated by Gα proteins are more frequently studied, but Gβ and Gγ subunits, which form an obligate dimer (Gβγ), can also activate several signaling mechanisms [193,194,195]. There are also several G-protein independent signaling cascades, which vary with chemokine receptor, cell type and effector [196, 197]. These G-protein independent pathways are mediated by β-arrestins 1 and 2, which can serve as scaffolds for downstream signaling molecules [196, 198,199,200]. This often involves the recruitment of Src kinases [199, 201], which are also activated by Gαi signaling [202], and can lead to downstream activation of mitogen activated protein kinase (MAPK) signaling pathways [203,204,205] as well as the PI3 kinase (PI3K) signaling pathway [199, 206].
Chemokine receptor signal transduction is also regulated by blocking receptor signaling. This involves GPCR kinases (GRKs) phosphorylating several different residues on the active carboxyl terminus of the GPCR, recruiting β-arrestins [207, 208]. Following their recruitment, β-arrestins carry out two main functions. First, they occlude the G-protein binding site on the chemokine receptor, blocking signaling and desensitizing the receptor [208, 209]. Next, β-arrestins can recruit the cellular machinery required for clathrin-mediated internalization, initiating either receptor recycling [209, 210] or lysosomal degradation of the receptor [211]. Some studies suggest that the strength of the interaction between the arrestin and the receptor regulates this choice, with weaker interactions resulting in degradation and stronger interactions leading to receptor recycling [212]. Irrespective of the precise mechanism, GRK/β-arrestin-mediated desensitization and internalization works in conjunction with ligand stimulation to mediate chemokine receptor signaling.
The primary function of chemokine receptors is to mediate cell migration, and much of the signaling originating from these receptors promotes cell movement. Classically, these receptors couple to Gαi [4, 31, 36, 213,214,215]. Activation of this G protein inhibits adenylate cyclase, decreasing cyclic AMP (cAMP) production [29, 35, 215,216,217]. Gαi can also activate Src and Lyn kinases [202, 218, 219], which in turn activate the small GTPases Ras and/or Raf resulting in MEK and ERK phosphorylation [220,221,222] and leading to chemotaxis [223,224,225,226,227,228]. Release of the Gβγ subunits can also stimulate cell movement [195, 228, 229], activating PI3K and the serine/threonine kinase AKT [224, 230]; the GTPases Rac and Rho [231]; or phospholipase C-β (PLCβ) [232, 233], which hydrolyzes phosphatidylinositol 4,5-bisphosphate (PIP2) to diacylglycerol (DAG) and inositol triphosphate (IP3), initiating intracellular calcium (Ca2+) release from the endoplasmic reticulum as well as activation of protein kinase C (PKC). Several of these cascades, including activation of PI3K and AKT [199, 206], or Src and ERK1/2 [203,204,205], can also be triggered by recruitment of β-arrestins and are associated with a number of downstream functions, including chemotaxis and apoptosis [196, 198,199,200].
These are just some of the vast network of signaling pathways associated with chemokine receptor activation, and both CCR5 and CXCR4 have a number of important pathways specifically associated with their activity. The complexity of these and other signaling networks initiated by co-receptor binding is often amplified by the use of different types of cell systems. The use of different types of cells has enabled more precise interrogation of specific signaling processes, but it has also led to confusing and contradictory data, as co-receptors in distinct cell types often show different responses to similar stimuli. This is particularly problematic when comparing transfected cells and cell lines to primary cells and in vivo systems, as it can create confusion about the physiologically relevant actions of co-receptors during disease. These signaling processes, their downstream effects and the caveats associated with their study are discussed below, and those involved in HIV pathogenesis are described in detail in the subsequent section. Further discussion on the functions and pathways activated by chemokine receptor signaling can be found in additional reviews [29, 34, 234,235,236].
CCR5
The CCR5 receptor was first identified as a human monocyte chemokine receptor in 1996, following the discovery of its primary ligands CCL3 (MIP1α), CCL4 (MIP1β), and CCL5 (RANTES) [4,5,6, 237]. Only CCL4 binds exclusively to CCR5, with CCL3 also binding to CCR1 and CCL5 acting as a ligand for both CCR1 and CCR3 [28, 238]. Several other ligands including CCL2 (MCP-1), CCL7 (MCP-3), CCL8 (MCP-2), CCL13 (MCP-4) and CCL11 (eotaxin), have shown some affinity for CCR5 in vitro in binding studies or competition assays [237], but the in vivo relevance of these interactions is unclear [37, 186]. CCR5 is expressed on a number of cell types, including macrophages [18, 239], microglia [18, 240], T-cells [18, 20, 21], and numerous other immune cells [18] as well as astrocytes [114, 240] and neurons [49, 185, 241].
Under homeostatic conditions, chemokine signaling through CCR5 is associated with numerous physiological processes including leukocyte migration [218, 223, 242,243,244], regulation of inflammation through cytokine and chemokine release [245, 246], and the activation of cell survival pathways [219]. These functions are thought to be particularly important in the context of viral infection or cellular injury. For example, CCR5 is upregulated during inflammation on CD8 + T-cells, promoting the migration of these cells towards the site of infection and thereby increasing the likelihood of encountering antigen specific cells to enhance the adaptive immune response [247]. This receptor may also play a role in the recruitment of immune cells to the CNS [248], the formation of atherosclerotic plaques [186], and tumor cell migration and survival [249, 250]. These effects require a number of intracellular signaling cascades, some of which may be distinct to CCR5 activation by specific ligands, such as CCL5 [35, 251].
Due to its broad influence, CCR5 signaling is tightly regulated by several processes, generally triggered by phosphorylation of serine residues on its C-terminus and the Asp-Arg-Tyr (DRY) motif leading to GRK/arrestin mediated receptor desensitization and internalization [34, 252]. The early steps of CCR5 desensitization are similar to that of other chemokine receptors and involve C-terminal phosphorylation, recruitment of β-arrestins, and clathrin-dependent endocytosis [34, 207, 253]. Upon internalization, CCR5 is directed through the endosomal recycling compartment to the trans-golgi network (TGN), where it is then recycled back to the cell surface upon resensitization [254, 255]. Importantly, different ligands can induce different fates for CCR5 once it is internalized. Some, like the chemokine analogues PSC-RANTES and AOP-RANTES, promote sequestration of CCR5 in the endosome recycling compartment or TGN [254, 256, 257], while others, like the physiological ligand CCL5, induce recycling back to the cell surface [255, 258]. This appears to be due to the ability of these ligands to alter the structure of the intracellular CCR5 loops via GRK recruitment and/or binding of different β-arrestins [256, 259]. Given that the sequestration of CCR5 is a promising method of blocking viral entry [38, 212, 258, 260,261,262,263], further examination of the processes mediating CCR5 desensitization and recycling may have important implications for HIV infection and potential antiretroviral activity.
CCR5 signaling
The CCR5 receptor can signal through several distinct G-protein mediated pathways. Signaling through Gαi inhibits the activation of adenylate cyclase, regulating the production of cAMP [35, 215, 216, 251] and MAPKs [191, 221, 264], and functions such as T-cell proliferation and chemotaxis [34, 214, 230, 265,266,267]. Signaling through Gαi can be influenced by the formation of oligomers, changing the responses to select ligands [268]. For example, in HEK293 cells co-transfected with CCR5 and/or CCR2, the Gαi inhibitor pertussis toxin (PTX) block CCL5-induced Ca2+ release in cells expressing only CCR5, but not in cells co-expressing CCR5 and CCR2 and treated with CCL2 and CCL5 [268]. Homodimerization or dimerization with either CCR2 or CXCR4 may also prevent gp120 binding [244, 269, 270], although whether this has any in vivo relevance is unclear.
However, several CCR5 signaling pathways are insensitive to pertussis toxin, indicating the involvement of alternative G-proteins and/or G-protein independent signaling mechanisms [271,272,273]. Much CCR5-initiated signal transduction can also occur through intracellular Ca2+ release and PLCβ activation, which can be mediated by Gβγ [32, 215, 219, 232, 233, 242, 274]. Moreover, many studies rely solely on pertussis toxin, which may have Gαi-independent effects via its B-oligomer [32, 275], or do not make use of specific G-protein inhibitors, making it is difficult to define the specific G-proteins initiating each pathway. Further, while not specific to CCR5, studies have suggested that the release of Gβγ, rather than Gαi is essential for chemotaxis mediated by Gαi-coupled receptors [195, 229]. Both the PLCβ and PI3K signaling cascades mediate the chemotactic response to CCR5 binding by activating a number of Ser/Thr protein kinases, particularly members of the PAK and FAK families, in macrophages [230, 242, 273, 276] and T-lymphocytes [213, 214, 276]. Signaling via the PLCβ signaling cascade also activates MAPKs [221, 267], which is associated with the production of inflammatory mediators [31, 267, 277] and chemotactic responses [223, 278, 279]. Activation of PI3K is specifically required for CCL5 mediated chemotaxis in macrophages and T-lymphocytes and cytoskeletal rearrangement induced by Rho GTPases [230, 242, 280], and can also activate AKT and MAPK signaling [221, 267]. These data demonstrate the importance of PLCβ and PI3K in CCR5-chemokine signaling, suggesting the release of Gβγ may be more critical in mediating the effects of CCR5 than Gαi.
CCR5 may also couple to Gαq, which can also initiate IP3-mediated intracellular Ca2+ release [36, 215, 251, 281]. A recent study showed that both CCL4 and CCL5 could initiate Ca2+ flux in HEK.CCR5 cells, and this was inhibited by the Gαq inhibitor YM-254,890 [215]. The physiological relevance and extent of Gαq-specific CCR5 signaling is unclear, due to the extensive overlap between Gβγ and Gαq signaling. However, the centrality of Gβγ-mediated PI3K signaling to chemotaxis suggests CCR5 mainly acts through Gαi and Gβγ in response to physiological ligands [195, 229]. In addition, the binding of HIV gp120 may push CCR5 towards Gαq signaling, as siRNA against Gαq, but not Gαi, prevented CCR5-mediated changes in viral fusion [281] and gp120 induces several Gαi independent signaling effects [273, 282,283,284,285].
In addition to Gαi, Gαq, and Gβγ signaling, several signaling processes may be independent of G-protein activity. Stimulation of CCR5 by either CCL2, CCL3, or CCL5 leads to activation of janus kinases 1 and 2 (JAKs) [214, 271, 272], and activation of JAK2 was insensitive to both pertussis toxin and U73122, an inhibitor of PLCβ [272]. This suggests JAK2 activation is mediated via G-protein independent mechanisms, potentially through a direct JAK-CCR5 interaction, which can lead to the phosphorylation and dimerization of the receptor. Activation of the PI3K/AKT signaling could also be initiated by G-protein independent signaling through β-arrestins, as CCL4 stimulates macrophage chemotaxis by inducing the formation of a β-arrestin signaling complex comprised of PI3K, Pyk2 and Lyn, leading to downstream ERK activation [218].
The responses to CCR5 binding are mediated by a number of effectors, including but not limited to MAPKs [219, 223, 271, 273, 276, 277, 282, 286], signal transducer and activator of transcription (STAT) proteins [214, 272, 287], AMP-activated protein kinases (AMPKs) [242], and small GTPases (Rac, Rho) [266] or FAKs like Pyk2 [213, 243, 272, 273, 276, 282], which play a critical role in chemokine-mediated cellular migration in both lymphocytes and macrophages [223, 230, 243, 266, 276]. Many of these effectors are activated by one or more endogenous CCR5 ligands, such as the phosphorylation of the MAPK ERK 1/2 by both CCL3 and CCL4 [223]. Multiple overlapping pathways are linked to MAPK signaling, including the release of pro-inflammatory chemokines, cell survival, cell death, the activation of STATs, and the activation of matrix metalloproteinases (MMPs) [279, 288, 289]. For example, induction of CCL5 in response to influenza infection leads to Gαi-mediated activation of both the MEK/ERK and PI3K/AKT signaling cascades, reducing apoptosis in mouse macrophages [219]. Stimulation with CCL5 also leads to Ca2+ mobilization and the activation of JAK1/STAT5, triggering cell polarization and migration. However, this did not occur with AOP-RANTES, a synthetic CCL5 derivative that can also bind to CCR5 but does not induce chemotaxis [261, 290], due to differences in the length of Gαi-association and subsequent release of Gβγ [214]. Many of these pathways may also be regulated by PI3K and PLCβ activity. For example, both of these proteins were necessary to promote CCL5-mediated chemotaxis through AMPK in RAW264.7 rodent macrophages [242]. This demonstrates that activation of CCR5 by a single ligand can simultaneously activate several different G-proteins and downstream pathways, regulating multiple cellular functions, a commonality among GPCRs. Indeed, CCR5-mediated activation of JAK/STAT pathways may play a role in T-cell activation and proliferation, although it is unclear whether this is solely due to CCR5 or a combination of CCR receptors, including CCR2 [34, 268, 271, 287, 291]. These data show that CCR5 can initiate signaling through both G-protein dependent and independent pathways and demonstrate the substantial overlap between CCR5 signaling cascades activated by distinct G-proteins (Fig. 1A). These overlaps, combined with the likelihood that the coupling of CCR5 to certain G-proteins and pathways may be different in distinct cell types and species, shows the challenges involved in defining how the specific ligands activate certain signaling pathways and highlights an important area for future studies.
Conformation mediated changes in CCR5 signaling
In addition to complications due to overlapping pathways and distinct effects in different cell types [214, 271, 272], analysis of CCR5 signaling is complicated by the existence of conformationally heterogenous populations of CCR5 on the cell surface [257, 292,293,294]. The conformational differences center on the degree of exposure of binding sites in the N-terminal (NT) and second extracellular loop (ECL2) regions [40, 292], and the interaction of the N-terminus with the transmembrane helix bundles [34, 295]. These differences seem to have a broad impact on receptor function, changing the affinity for different ligands, the coupling to specific G-proteins and the functional potency of 2nd messenger induction [215, 257, 294, 296, 297]. In CCR5 transfected L1.2 lymphocytes, antibodies against the ECL2 region of CCR5 efficiently blocked the binding and functional response of CCL3, CCL4, and CCL5, whereas an antibody against the N-terminal region blocked only CCL4 binding and signaling, not CCL5 [292, 298]. Moreover, an CCR5 antibody mapped to multiple ECL domains was able to increase Gαi-associated GTPγS binding, but not Ca2+ flux, suggesting stabilization of specific conformations of CCR5 is associated with distinct signaling cascades [292].
Other studies show differences in the β-arrestin mediated internalization of specific CCR5 populations, suggesting subpopulations of CCR5 conformations may be resistant to chemokine-induced internalization in macrophages, but not T-cells [299]. This is consistent with studies showing subpopulations of CCR5 may differ between cell types [257, 300]. These differences may involve changes in receptor occupancy based on the ability of different ligands to bind to distinct CCR5 conformations [215]. Notably, natural chemokines only bind to a few distinct conformations, but gp120 may interact with a much wider number of CCR5 populations, potentially due to the interaction of the flexible V3 region of gp120 with the N-terminus of CCR5 [215, 257, 293, 300, 301]. Indeed, changes in the exposure of the NT or ECL2 regions can increase the binding affinity for specific sequences in CCR5 tropic viral envelopes [300], likely playing a role in the maintenance of gp120—CCR5 binding in the presence of natural ligands and some antagonists. Thus, defining not only the CCR5 signaling pathways, but the shifts in ligand binding and signaling associated with those structural changes, could initiate or enhance the development of a number of novel antiretroviral effectors [40, 293, 301, 302].
CXCR4
The CXCR4 receptor was first identified on peripheral blood lymphocytes in 1994 [7] and is highly expressed in a variety of cell types including lymphocytes, hematopoietic stem cells, neural cells, and stromal fibroblasts [30]. Expression of CXCR4 is relatively low on myeloid cells such as macrophages and microglia [303]. The primary ligand for CXCR4 is the chemokine CXCL12 (SDF-1), which is ubiquitously expressed in both embryonic and adult tissues, including liver, pancreas, spleen, brain, and heart [304]. The CXCL12/CXCR4 signaling axis was long considered monogamous until 2007 when it was discovered that macrophage migration inhibitory factor (MIF) was an alternative ligand for CXCR4 [305]. Additionally, both pharmacologic antagonism and knockdown studies have shown that extracellular ubiquitin (eUb), is also a natural ligand for CXCR4 [306].
In adults, CXCL12-CXCR4 binding promotes the survival and trafficking of immune cells [307, 308]. During embryogenesis, CXCL12 is important for proliferation, migration, and differentiation of immature progenitors [309, 310]. The CXCL12/CXCR4 signaling axis is also important in CNS homeostasis, where it regulates the migration of neural precursors [311, 312], establishment of neural circuitry [313, 314], modulation of NMDA subunit composition [315], and alterations in dendritic spine number and morphology [316, 317]. MIF also plays an important role in cell recruitment and arrest through binding to CXCR2 and CXCR4 [305], and can exert protective functions in liver fibrosis, myocardial ischemia–reperfusion injury and in the developing cerebral cortex upon tissue damage [318]. In contrast to CXCL12, MIF is not associated with homeostatic function, but rather pro-inflammatory and pro-atherogenic activity [305, 319], and is considered an inflammatory cytokine [320]. Similar to other CXCR4 ligands, eUb can also mediate chemotaxis, but the chemotactic activity is weaker than induced by CXCL12 [321]. Furthermore, the interaction of eUb with CXCR4 is independent of the N-terminal receptor domain used by CXC12 and instead relies on binding sites in the 2nd and 3rd extracellular loop [321].
Like CCR5 and other GPCRs, CXCR4 signaling is regulated by desensitization (homologous and heterologous), internalization, and degradation. Direct activation of PKC by phorbol esters [322, 323], T or B cell receptor engagement [324, 325], CXCR1 activation [326], or CCR5 activation [327] are able to induce CXCR4 internalization. CXCR4 can recycle back to the plasma membrane following PKC-mediated internalization [322]; however, the receptor recycles poorly following CXCL12 stimulation [328]. CXCR4 has been shown to be ubiquitinated, sorted to the lysosome, and degraded [329], which is mediated by the E3 ubiquitin ligase AIP4 [330]. Targeting CXCR4 with specific agonists or molecules that promote the internalization and sequestration has not been as widely explored as for CCR5, owing in part to the physiological requirements for CXCR4 binding. However, there has been some progress in finding drugs that can induce signaling while preventing viral binding [43, 331].
CXCR4 signaling
Like CCR5 and other chemokine receptors, CXCR4 signals primarily through G-protein dependent pathways, primarily Gαi mediated inhibition in cAMP production. Recombinant overexpression systems, such as HEK293T and Sf9 cells show that CXCR4 can activate different Gαi proteins, including Gαi1, Gαi2, Gαi3, and Gαo in response to CXCL12 stimulation [332], although it seems that CXCR4 couples more efficiently to the Gαi1 and Gαi2 rather than Gαi3 and Gαo [333, 334]. Activation of CXCR4 coupled to Gαi triggers activation of MAPK and PI3K pathways [335], mediating effects on migration [336,337,338] as well as cell survival and proliferation [313, 339]. CXCR4 can also act through other G proteins, such as the noncognate G protein Gα13 [340] or Gαq [341, 342], although this may be context specific. CXCR4 only couples to Gαq in dendritic cells and granulocytes but not T and B cells [343], and coupling to Gα13 may have particular relevance in cancer, where Gα13 is overexpressed [344, 345]. As with CCR5, activation of CXCR4 coupled to Gαq can induce IP3 mediated Ca2+ release through PLCβ [341, 342]; this pathway can also be activated via the released Gβγ subunit [235, 346]. Activation of CXCR4 coupled to Gα13 in Jurkat T cells mediates cell migration via activation of Rho [340], and also mediates CXCR4 trafficking into Rab11 + vesicles during CXCL12-induced endocytosis in T cells [347].
Like CCR5, CXCR4 can also signal through G protein independent mechanisms, such as β-arrestin-mediated signaling. Both β-arrestin-1 and -2 enhance CXCR4-mediated ERK activation [338], and β-arrestin-2 is involved in p38 activation and migration following CXCL12 stimulation [348]. Additionally, upon CXCL12 stimulation, CXCR4 can dimerize and become phosphorylated at intracellular tyrosines by rapid recruitment and activation of JAK2 and JAK3 [349]. This leads to STAT dimerization and activation of the STAT pathway, which is unaffected by pertussis toxin treatment [349]. The JAK/STAT pathway leads to diverse cellular effects, including mobilization of Ca2+ from intracellular stores, and after its nuclear translocation, the transcription of several target genes [349]. Studies using a JAK-specific inhibitor have shown that in a T-lymphoblast cell line, the association of Gαi with CXCR4 is dependent on JAK, further supporting a co-dependent mechanism between members of the JAK/STAT pathway and G-protein coupled signaling [350] (Fig. 1B). Activation of the JAK/STAT pathway may also be affected by the oligomerization state of this receptor, as homodimerization of CXCR4 is necessary to elicit G protein independent activation of JAK/STAT and enhance the response of CXCR4 to CXCL12 [349]. The heterodimerization/oligomerization of CXCR4 to other receptors may have a number of other potential effects on chemokine signaling responses as well, interfering with the binding to one receptor in the oligomer or altering its ability to interact with or signal through intracellular mediators [351]. This is exemplified by the potential ability of ACKR3 to interact with CXCR4 and affect CXCR4 trafficking and/or coupling to other proteins [352].