Ischemic stroke is mainly caused by cerebral vascular stenosis or occlusion. The focal blood flow declines sharply, leading to a shortage in the supply of oxygen and nutrients, eventually causing cell death and tissue damage. The main way that ischemic brain tissue compensates for insufficient oxygen supply is by building new vessels. Angiogenesis in the infarcted area determines the recovery of cerebral blood flow, neuronal regeneration, and reconstruction of synaptic connections between nerve cells, which all influence the degree of functional recovery of the patient (14). The collateral circulation in the acute and subacute phases of ischemia relies on the opening of a pre-existing vascular network, while in the chronic recovery stage, it mainly depends on the formation of new blood vessels. New blood vessels have been shown to appear on the third day and persist for at least 90 days after stroke onset (15, 16). Angiogenesis is stimulated by massive production of VEGF from hypoxic tissues after stroke (17, 18). Proteolytic enzymes, angiogenic growth factors, and inhibitors collaborate to maintain the migration and proliferation of ECs. Vascular endothelial growth factor-A (VEGF-A) has been shown to be significantly elevated, which causes ECs to proliferate and protrude filopodia by coordinating with Notch family receptors and their ligands. The migration and proliferation of ECs are then enhanced and mediated by the interaction between VEGF-A and VEGFR-2 (15, 19). The activation of VEGFR-2 also induces the release of matrix metalloproteinases (MMPs) and endothelial growth factors (EGFs), which may contribute to endothelial progenitor cell migration and basement membrane degradation (20). Previous studies have identified that various immune cells and cytokines are directly or indirectly involved in the regulation of angiogenesis. The cytokines associated with angiogenic effects are listed in Table 1 .

Post-stroke neuronal remodeling is mainly caused by residual neurons that survive ischemia-reperfusion injury and newborn neurons from neurogenesis. Axonal sprouting from residual neurons does not normally appear until 14 days after ischemia, even in the ischemic penumbra (36). The occurrence of newly formed cortical circuits can be detected as early as 3 weeks after ischemia, which is associated with functional recovery (36). Factors released by newly formed blood vessels secrete signals, such as VEGF, artemin, and neurotrophins, which are critical components for axonal sprouting (17, 37). In addition, cytokines from the microglia and invading peripheral immune cells activate astrocytes, which promote angiogenesis and synchronize neuronal activity to induce axonal outgrowths and establish new connections (10, 38, 39). These processes constitute neural tissue repair.

The number of neurons that survive ischemia-reperfusion injury is limited, and newly formed neural connections are generated by neurogenesis. Neurogenesis is the process of generating new functional neurons from endogenous NSCs, including proliferation, migration, and differentiation into mature neurons (40). Neurogenesis becomes highly active following ischemia-reperfusion insults in mainly two distinct regions: the subventricular zone (SVZ) of the lateral ventricles and the subgranular zone (SGZ) of the dentate gyrus of the hippocampus (41). In both areas, primitive pluripotent NSCs express glial fibrillary acidic protein (GFAP). The transcription factor, sex-determining region Y-box 2 (Sox2), and Nestin are termed type B cells in the SVZ, or type I cells in the SGZ (42). These cells are largely quiescent in the physiological state, but they can be activated in response to various external stimuli, such as exercise, hypoxia, and ischemia (43, 44). However, neural regeneration is relatively inefficient, as only a limited number of immature neurons integrate into the existing circuitry and mature into functional neurons. Others undergo programmed cell death and are cleared by microglia (45, 46) In addition, overactivation of transit-amplifying cells by factors secreted by immune cells, glial cells, and newborn ECs during pathological states may lead to the exhaustion of the stem cell pool and early termination of neurogenesis.

The microvasculature brings oxygen, nutrients, and growth factors to the area of ischemic injury, which creates an appropriate microenvironment for cell migration. Neuroblasts from the SVZ or SGZ expand and migrate to the peri-infarct region, where post-ischemic angiogenesis occurs. Compelling evidence suggests that several growth factors released from ECs are also indispensable for NSC proliferation, such as basic fibroblast growth factor, brain-derived neurotrophic factor, EGF, and VEGF (47). In addition, angiogenesis promotes vascular-guided NSC migration. Recent studies suggest that Laminin/β1 integrin signals support the synergistic migration of newly formed microvessels and neuroblasts toward injury sites. β1 integrin signals on the surface of microvasculature have been shown to play a critical role in bridging neuroblasts to laminin and promoting cell body movement during migration (48, 49). The relief of hypoxia induced by angiogenesis in the cortex triggers a switch from NSC expansion to neuronal differentiation (50). Together, the factors released by ECs from the nearby microvessels (choroid plexus) are essential for neurogenesis and neuronal maturation during the nerve repair process after stroke ( Figure 1 ).

Considering that neurogenesis usually begins over 1 week after stroke onset, infiltrated immune cells and cytokines have been found to affect neurogenesis during this period. Cytokines reported to have potential regulatory effects on NSCs are listed in Table 2 . Numerous studies have identified a variety of immune cells, such as neutrophils, monocytes, macrophages, microglia, T lymphocytes, and B lymphocytes, that play crucial roles in inflammation-mediated angiogenesis and neurogenesis post-stroke (64, 65). The molecular mechanisms of immune cell-mediated nerve repair in stroke are discussed below.

Neutrophils are recognized as the first line of defense against pathogens and antigens in the peripheral immune system. As members of the innate immune system, the defense mechanisms of neutrophils have minimal specificity and always cause tissue damage after their activation. Similar to deployed soldiers, neutrophils are fully armed when exiting the bone marrow and require little transcriptional or translational modification to function. Based on these characteristics, neutrophils are the first group of immune cells to infiltrate the brain after ischemic stroke (66). Using in vivo two-photon microscopy combined with immunohistochemistry, researchers demonstrated that neutrophils rapidly attach to inflammatory brain ECs within a few minutes and peak at 1–3 days, as seen in permanent and transient middle cerebral artery occlusion (tMCAO) models in mice (67, 68). Neutrophils are characterized by three main functions including phagocytosis, respiratory burst, and formation of neutrophil extracellular traps (NETs). These mechanisms coordinate with one another and cause neuronal and BBB injuries (69). Stimulated by DAMPs, neutrophils release nuclear and granular contents to form a wide network of DNA complexes (NETs), which are used to capture, neutralize, and kill pathogenic microorganisms and prevent their spread. Verified by the expression of Ly6G and protein arginine deiminase 4 (PAD4), Kang et al. found that the formation of intravascular and intraparenchymal NETs peaked at 3–5 days post-stroke in a mice MCAO model (69). PAD4 has been shown to act as an enzyme essential for the NET formation, which contributes to enlarged thrombosis, BBB injury, and reduced neovascularization in the latter stage of stroke (70, 71). Overexpression of PAD4 induces an increase in NET formation, which is accompanied by increased BBB damage and reduced angiogenesis (72). Disruption of NETs by DNase I, inhibition of NET formation by genetic ablation, or pharmacological inhibition of PAD increases revascularization and improves functional recovery. In summary, the above findings suggest that stroke-induced activation of neutrophils form excessive NETs around the BBB and cerebral parenchyma post-stroke, which may affect vascular remodeling during the recovery phase of stroke.

Phagocytosis triggers degranulation, including the release of antimicrobial peptides, proteases, and superoxide anions (O⁻²), which act as the “blasting fuse” of reactive oxygen species (ROS). Although ROS plays an important role against pathogens, excess ROS can damage surrounding tissues, such as the BBB, which promotes blood leakage into the brain parenchyma and causes a more severe inflammatory response. Moreover, ROS was shown to mediate the expression of MMP-9 via the phosphatidylinositol-3 kinase-mediated signaling pathway, which further impaired BBB integrity (73). Moreover, ROS promotes NET formation by activating PAD4, which induces the unwinding of DNA strands (70). A recent study found that neutrophils mediated the recovery of the endothelium by secretion of Cathepsin, which deposit along the injured arterial lumen induces the recruitment of circulating endothelial progenitor cells in an N-formyl peptide receptor 2-dependent manner in ischemic limb (74, 75). However, only a very small number of neutrophils are present in the brain parenchyma beyond 7 days after stroke onset, which suggests that their function in the latter phase of stroke is mainly secondary to their function in the acute stage (76). In a recent study, skewing neutrophils towards N2 phenotype (CD206⁺) before experimental stroke facilitated the clearance of neutrophils by macrophage, which relief further inflammatory damage and therefore improved long-term recovery (77). The interaction between neutrophils and endothelium are depicted in Figure 2 .

There are also a large number of CNS-resident macrophage-like cells, namely microglia, which have phagocytic abilities during tissue damage. Similar to peripheral macrophages which come in with dozens of varieties, microglia also exhibit tremendous heterogeneity, as evidence by advanced transcriptomic and proteomic profiles. Therefore, it is reluctance to paraphrase the microglia with M1/M2 paradigm since the definition are derived from exposing isolated cells to certain stimuli in vitro, which were drastic differences from the microenvironment in pathological state in vivo, such as tumor, infection, aging, physical trauma and stroke.

Microglia have been previously observed to engulf neutrophils in an ischemic tissue model in vitro (93, 94). Subsequently, microglia were also found to activate ECs and capture invasive neutrophils in a cooperative manner (67, 95). Similar to infiltrated macrophages, activated microglia release various pro-inflammatory cytokines, including IL-1β, TNF-α, IL-18, IL-23, and iNOS (96). About 30% of brain resident microglia observed in adult mice are capillary-associated microglia which constantly survey the influx of blood-borne components into the CNS (97). Microglia have been shown to have extensive interaction with ECs and pericytes, which may regulate capillary diameter and cerebral blood flow by altering pericyte or astrocyte coverage (98). In the acute stage of stroke, microglia were attracted to blood vessels with BBB leakage and contributed to the disintegration of blood vessels by phagocytosis of endothelial cells (99). In addition, pericytes were shown to detach from capillary and participate in inflammatory-immunological response mediated by the interaction between DAMPs and TLR4 expressed on the surface of pericytes (100, 101). Interestingly, activated pericytes were found to express microglial markers in both experimental stroke brain and human stroke brain tissue, which may be involved in BBB damage and brain edema (102, 103).

In the chronic recovery stage of stroke, microglia contribute to debris clearance and tissue repair as they engulf newborn neurons that fail to integrate into the neural circuit. Using Rag^-/-γc^-/- mice with ischemic stroke, microglial depletion was shown to exacerbate stroke severity and impair long-term outcomes (104). In contrast, microglial depletion by a CSF1R inhibitor (PLX3397) was also shown to contribute to endogenous neurogenesis and improve functional recovery in a traumatic spinal cord injury model (105). Considered together, both peripheral macrophages and microglia represent a biphasic regulatory mechanism in the context of stroke. Therefore, to achieve therapeutic goals, future studies should focus on elucidating the mechanism of subtype-transdifferentiation rather than simple activation or depletion.

While immediate and early proinflammatory responses to stroke are mainly mediated by innate immune cells, B cell-mediated adaptive immune responses exhibit delayed pro-injury functions in the latter stages. The initiation of an adaptive immune response specific for CNS antigens can be observed at around week 4, when primed antigen-presenting cells present cell fragments as antigens to T cells and B cells (125). B cells mainly function in three ways: antigen presentation, antibody production, and cytokine secretion. Naïve B cells express the primary effector antibodies, IgM and IgD. After receiving antigens, B cells initiate isotype transformation, express plasma cell markers, and produce antigen-specific IgG or IgA antibodies, which may be indirectly associated with some stroke-related risk factors, including hypertension, diabetes, and atherosclerosis (126). Stroke may lead to subsequent vascular dementia. In a distal pMCAO model, mice developed short-term memory deficits between weeks 1 and 7 following cerebral infarction. B cells were found to accumulate in the infarct region and secreted IgA and IgG at 4–7 weeks after stroke (127). Genetic deficiency and pharmacologic ablation of B-lymphocytes using an anti-CD20 antibody prevented delayed-onset cognitive deficits, suggesting that immunoglobulin synthesis by B-lymphocytes may be involved in long-term injury to neuronal cells after stroke. In a retrospective cohort study, immunoglobulin synthesis (IgG, IgM, and IgA) in the cerebrospinal fluid of stroke patients was found several months after stroke onset (128). The antibodies may activate the complement pathway and cause further damage to neuronal cells, which is a main cause of neuronal death in multiple sclerosis (129). Concrete evidences have proved that massive memory B cells and antibody-producing plasma blasts were exist in the CSF of patients with multiple sclerosis (130, 131). In addition to the production of IgG, B cell and its subtypes were also found to be involved in the pathogenesis of MS by secretion of proinflammatory cytokines, such as lymphotoxin-alpha, CXCL12, and CXCL13 (132). Therefore, it is reasonable to speculate that B cells may also play a key role in the chronic stage of stroke. Up to date, there is no explicit evidence of the beneficial role of B cells in the chronic stages of stroke. In gene ablation mice models, mice deficient in lymphocytes (Rag^−/−), or specific depletion of CD4⁺ T cells, CD8⁺ T cells, B cells, or IFN-γ, was used to determine the contribution of different lymphocyte subgroups to ischemia reperfusion injury and recovery. Smaller infarction volumes and amelioration of neurological disorders were found in mice lacking lymphocytes (Rag^−/−), but no improvements were observed in mice lacking B cells. Furthermore, B cell transfer in Rag^−/− mice did not shown any significant improvement, indicating that B cells may not be play a protective role in the chronic stage of stroke (133). On the contrary, B cells may have negative effects on post-stroke neuroprotection. In an MCAO model in μMT^–/– mice (B cell depletion), enlarged areas of infarction, neurological disorders, and higher mortality were observed (134, 135). A higher number of invading circulating immune cells, including activated T cells, macrophages, microglial cells, and neutrophils were found in the ischemic hemispheres of μMT^−/− mice. It is worth noting that beneficial effects were observed after adoptive transfer of B cells from WT mice into μMT^−/− mice prior to MCAO modeling. Interestingly, it is reasonable to speculate that these protective effects were mediated by IL-10 since no obvious change was observed when B cells were transferred into IL-10-deficient mice.

In addition to the above immune cells, DC and NK cells have also been reported to be involved in the pathological process of stroke. However, their main role is to coordinate with other types of cells to facilitate an immune response. DCs are the most efficient antigen-presenting cells. Under physiological conditions, DCs mainly patrol near the cerebrospinal fluid, such as the meninges and choroid plexus, and are barely present in cerebral parenchyma (136). In a pMCAO model, DCs (OX6⁺) were found to invade the ischemic core within the first few hours and gradually increased until 6 days after ischemia (137). However, in addition to acting as antigen-presenting cells, their functions remain unclear. NK cells are innate lymphocytes that can be swiftly mobilized during the earliest phases of immune responses. Recruited by CXCL8, NK cells have been shown to accumulate in the ischemic hemisphere and aggregate in the peri-infarction area through the release of IFN-γ and ROS in the acute stage of stroke (138, 139). However, the spatiotemporal and phenotypic profiles of NK cells after stroke are yet to be elucidated. In addition to peripheral immune cells, the meningeal immune cells have recently been recognized as a potential immune repertoire which may have an impact on the neuro-immune crosstalk during stroke (140). Study have shown that the meningeal mast cells, including granulocytes and activated macrophages, may contribute to stroke-related pathology in MCAO mice (141). Although the properties of meningeal immunity have yet to be fully elucidated, its potential threat or benefit to the CNS are worth being addressed. The roles of immune cells in angiogenesis, neuronal remodeling, and neurogenesis are summarized in Table 3 .