Temporal changes in monocyte and macrophage subsets and microglial macrophages following spinal cord injury in the lys-egfp-ki mouse model

Highlights

• The microglial response dominates over the myeloid response in the spinal lesion

• In SCI lesions classical macrophage dominate acutely, decline then return chronically

• Non-classical macrophages arise subacutely and persist chronically in spinal lesions

Abstract

The role of hematogenous (hMΦ) and microglial (mMΦ) macrophages following spinal cord injury (SCI) remains unclear as they are not distinguished easily from each other in the lesion area. We have recently described the temporal and spatial response to SCI of each MΦ population using the lys-EGFP-ki mouse that enables EGFP⁺ hMΦ to be distinguished from EGFP⁻ mMΦ at the lesion site. In the present study, we characterized the response of monocyte and hMΦ subsets and mMΦ to SCI. We describe, for the first time, the responses of circulating classical (pro-inflammatory) and non-classical monocyte subsets to SCI. Additionally, we show the presence of classical and non-classical hMΦ at the SCI lesion. Importantly, we demonstrate that the ‘classical pro-inflammatory’ hMΦ respond in the acute (1 d, 3 d) stages of SCI while the ‘non-classical’ hMΦ respond in the sub-acute (7 d, 14 d) phase of SCI. At later time points (6 weeks post injury) classical hMΦ return to the injury site. Our study offers new insight into the cellular inflammatory response that occurs after SCI and suggests that the timing and targets of anti-inflammatory therapies may be crucial to maximize neuroprotection at the acute and more chronic stages of SCI.

Introduction

Immediately following a spinal cord injury (SCI), an inflammatory response is initiated by neurons and glia within and surrounding the spinal lesion; this response includes the release of a cascade of inflammatory cytokines and chemokines (Popovich and Longbrake, 2008, Pineau et al., 2010, David and Kroner, 2011, Hawthorne and Popovich, 2011). Microglia, the resident innate immune cells of the CNS, participate in this response, reacting to trauma by activating, proliferating and migrating to the lesion (David and Kroner, 2011). In addition the release of pro-inflammatory cytokines, and in particular chemokines, leads to the subsequent accumulation of hematogenous leukocytes such as neutrophils and monocyte-derived hematogenous macrophages (hMΦ) that infiltrate the lesion via diapedesis (Hawthorne and Popovich, 2011).

The role of the responding inflammatory immune cells in SCI is controversial as they can have beneficial as well as harmful effects on the injured spinal cord (Bethea and Dietrich, 2002, Donnelly and Popovich, 2008, Popovich and Longbrake, 2008, David and Kroner, 2011). For example, the damaging and toxic effects of neutrophils in the acute period after SCI are well established (Hawthorne and Popovich, 2011), yet depletion of neutrophils impairs neurological outcomes (Stirling and Yong, 2008). Although significant numbers of hMΦ are present in the lesion, their roles at various times following SCI remain poorly understood (Gordon, 2003, Mosser, 2003, Mantovani et al., 2004, Gordon and Taylor, 2005, David and Kroner, 2011). Factors such as the timing of cord entry and subtype identity of the infiltrating monocyte/hMΦ may be key determinants of their role in the injured spinal cord (Shechter et al., 2009, David and Kroner, 2011, Donnelly et al., 2011).

A major impediment to understanding the exact role of hMΦ in SCI is our inability to distinguish hMΦs from mMΦs (macrophages derived from microglia) in the lesion as these populations, when activated, are morphologically and phenotypically similar (Davoust et al., 2008, David and Kroner, 2011). Whether these two MΦ populations work synergistically or antagonistically is unclear. Recently, we demonstrated that hMΦ and mMΦ may be distinguished from each other after SCI in the lys-EGFP-ki mouse, based on EGFP transgene expression in hMΦ (Mawhinney et al., 2012). These transgenic mice express EGFP under the LysM promoter; thus, in these mice, EGFP is expressed in mature myelomonocytic cells [neutrophils, monocytes, and hMΦ (Faust et al., 2000) and weakly in some dendritic cells (G.A. Dekaban, unpublished data)]. We have confirmed that resting microglia and activated mMΦ lack EGFP expression in the lesion of SCI lys-EGFP-ki mice. We also demonstrated that the spatial organization and temporal appearance of hMΦ and mMΦ in the spinal cord lesion can be independently tracked because EGFP expression allows these two populations to be examined separately (Mawhinney et al., 2012). The lys-EGFP-ki mouse provides a good model for studying the MΦ response to SCI.

An additional degree of complexity in understanding the role of hMΦ in the inflammatory responses to SCI is the existence of subsets of circulating monocytes and of functionally different hMΦ (Tacke and Randolph, 2006, David and Kroner, 2011). In both the mouse and human, two subsets of hMΦ, designated the classical inflammatory and the non-classical or resident monocytes, have been identified in the blood. The classical pro-inflammatory monocytes are short-lived and home to inflamed tissues, mediated by expression of CCR2 and CD62L, where they participate in an ongoing immune response (Geissmann et al., 2003, Gordon and Taylor, 2005, Tacke and Randolph, 2006, Auffray et al., 2007). The non-classical, resident monocytes have a longer half-life in the blood where they generally remain and are thought to patrol blood vessel surfaces and/or to home to normal tissues to become ‘resident’ cells that maintain tissue homeostasis (Geissmann et al., 2003, Gordon and Taylor, 2005, Tacke and Randolph, 2006, Auffray et al., 2007, Tacke et al., 2007, Auffray et al., 2009). These non-classical, resident blood monocytes also move into sites of inflammation, with somewhat delayed kinetics compared to the classical, pro-inflammatory monocytes (Auffray et al., 2009). More recently the non-classical, resident monocyte subset has been further subdivided into the intermediate and non-classical subsets as defined by intermediate expression levels of key surface markers (Ly6C/G^int) and functional markers in mouse and humans (Tacke and Randolph, 2006, Ziegler-Heitbrock et al., 2010, Wong et al., 2012).

Macrophages are also classified into two major types, M1 and M2 (David and Kroner, 2011). Classically activated M1 MΦ are pro-inflammatory, phagocytic and cytotoxic. (Gordon, 2003, Mosser, 2003, Mantovani et al., 2004, Gordon and Taylor, 2005, David and Kroner, 2011) By contrast, M2 MΦ generally promote immunoregulation and tissue repair/remodeling, and suppress inflammation (Gordon, 2003, Mosser, 2003, Mantovani et al., 2004, Gordon and Taylor, 2005, David and Kroner, 2011). These two types of macrophages are found in SCI lesions (Kigerl et al., 2009, David and Kroner, 2011, Donnelly et al., 2011, de Rivero Vaccari et al., 2012). Whether M1 and M2 macrophages phenotypes are distinct cell populations, or constitute a continuum of physiological responses of a single cell population, remains unclear (Gordon, 2003, Mosser, 2003, David and Kroner, 2011). Accordingly, the conflicting evidence for the beneficial versus harmful effects of MΦ after SCI (Gordon, 2003, Mosser, 2003, Mantovani et al., 2004, Gordon and Taylor, 2005, Shechter et al., 2009, David and Kroner, 2011) may be due to MΦ heterogeneity.

In the present study, we examined the monocyte/hMФ response to SCI using the lys-EGFP-ki mouse model to investigate the temporal presence of monocyte and hMΦ subsets in the blood and spinal cord lesion during the acute (1 d, 3 d), sub-acute (7 d, 14 d), and chronic (6 wk) phases of SCI. In parallel, by taking advantage of the lack of EGFP expression in cells of microglial origin within the lesion, we also examined the microglial/mMΦ response. By distinguishing clearly between monocyte/hMϕ and microglia/mMϕ, we demonstrated that pro-inflammatory M1 monocyte/hMΦ appeared first within the spinal lesion followed by increasing numbers of non-classical monocyte/M2-like hMΦ. In the chronic phase of SCI, the numbers of the pro-inflammatory monocyte/M1-like hMΦ increased again and were greater than the non-classical monocyte/M2-like hMΦ. The microglia/mMΦ were always more numerous than the hMΦ.