The Journal of Neuroscience, September 9, 2015 • 35(36):12325–12327 • 12325

Journal Club

Editor’s Note: These short, critical reviews of recent papers in the Journal, written exclusively by graduate students or postdoctoral fellows, are intended to summarize the important findings of the paper and provide additional insight and commentary. For more information on the format and purpose of the Journal Club, please see http://www.jneurosci.org/misc/ifa_features.shtml.

The Complement C5aR Has a Dual, Time- Dependent Effect on the Outcome of Spinal Cord Injury

X Antigona Ulndreaj,1,3 XJessica M. Marbourg,2 and Pia M. Vidal3 1Institute of Medical Science, Faculty of Medicine, University of Toronto, Toronto, Ontario M5T 2S8, Canada, 2Neuroscience Graduate Program and Center for Brain and Spinal Cord Repair, The Ohio State University, Columbus, Ohio 43210, and 3Division of Genetics and Development, Toronto Western Research Institute and Spinal Program, Krembil Neuroscience Centre, University Health Network, Toronto, Ontario M5T 2S8, Canada Review of Brennan et al.

Thecomplementsystemconsistsofmorethan dents and humans (Peterson and Anderson, signaling is detrimental during acute in- 30 plasma and membrane-bound proteins 2014), but its role in CNS pathologies remains jury, but is necessary for protective glial that promote host defense by inducing cell ly- unclear. Some data suggest a protective effect scar formation chronically (Fig. 1). sis, disposing of immune complexes, and aug- of C5a signaling (Weerth et al., 2003; Beck et C5aR antagonism restricted to the first menting the adaptive immune response. al., 2010; Guo et al., 2013) while other data 7 dpi improved functional outcomes, in- Under homeostatic conditions, these comple- suggest the opposite (Li et al., 2014). These di- creased myelination, and reduced im- ment proteins are expressed by all the cells of vergent effects of C5a signaling might also be mune cell infiltration at 35 dpi (Brennan the CNS at low levels. However, upon insult time-dependent (Beck et al., 2010; Guo et al., et al., 2015, their Fig. 3). In this condition, (infection, injury, or ischemia) complement 2013). inhibiting C5aR signaling did not affect proteinsareupregulated,andifthebalancebe- A recent study by Brennan et al. (2015) astrogliosis, but it reduced immediate tween complement activation and regulation provides compelling in vivo evidence in a postinjury increases of inflammatory cy- is disturbed, the CNS parenchyma can be sub- mouse model of T9 spinal cord injury tokines, which are thought to be integral jected to detrimental complement-mediated (SCI) that resolves much of the uncer- for the recruitment of immune cells. To attack. The complement cascade is a series of tainty surrounding the effect and time de- confirm this, and in an attempt to restrict enzymatic reactions that can be activated pendency of C5a signaling after CNS the effects of C5aR signaling disruption to through at least three different pathways—the trauma. Specifically, during the acute in- cells in the CNS, Brennan et al. (2015) classical, the alternative, and the lectin path- jury phase [before 7 d post injury (dpi)], used bone marrow chimeric mice in way—all of which involve the generation of pharmacological blockade or genetic ab- which C5ar-deficient bone marrow cells C5a. Specifically, in one step of the comple- lation of the (C5aR) signifi- were transplanted into irradiated wild- ment cascade, C5 protein is fragmented into cantly improved functional recovery. type mice. In these animals, C5aR signal- C5a and C5b. While C5b participates in lysis of These functional benefits corresponded ing deficiency was restricted to cells of the pathogenic microbes by forming the mem- with reduced inflammatory cytokine lev- peripheral . Importantly, brane attack complex (MAC) on their cell sur- els and -derived myelin content and functional recovery face, C5a acts as a potent chemoattractant of infiltration early after injury. Conversely, after SCI in these mice was not different inflammatory cells at the site of its generation prolonged (C5aR blockade up to 21 dpi) from control irradiated wild-type mice (Peterson and Anderson, 2014). C5a is up- or permanent (C5ar knock-out) disrup- that received wild-type bone marrow regulated immediately after CNS injury in ro- tion of C5aR signaling impaired forma- transplants (Brennan et al., 2015, their tion of the astrogliotic scar around the Fig. 5). This suggests that, to achieve neu- lesion core, exacerbated peripheral im- roprotection, C5aR disruption must tar- Received June 22, 2015; revised Aug. 3, 2015; accepted Aug. 4, 2015. We thank Drs. Michael G. Fehlings and Phillip G. Popovich for providing mune cell infiltration and lesion size, and get CNS-resident cells. critical feedback during the preparation of this manuscript and Dr. Made- reduced functional recovery. These ef- The observation that eliminating pe- leine O’Higgins for assistance with editing. fects were shown to be mediated in part ripheral C5aR signaling does not affect Correspondence should be addressed to Antigona Ulndreaj: Krembil by C5aR-dependent activation of STAT3, SCI outcomes is critical given that infil- Discovery Tower, 60 Leonard Avenue, 7KD430, Toronto, ON M5T 2S8, Can- ada. E-mail: [email protected]. a transcription factor necessary for pro- trating immune cells have been directly DOI:10.1523/JNEUROSCI.2380-15.2015 moting astrocyte proliferation. Collec- associated with changes in pathology and Copyright © 2015 the authors 0270-6474/15/3512325-03$15.00/0 tively, these data indicate that C5aR functional recovery (Popovich et al., 12326 • J. Neurosci., September 9, 2015 • 35(36):12325–12327 Ulndreaj et al. • Journal Club

Figure1. Schematicdiagramshowingdiversetime-dependenttherapeuticeffectsofC5asignalingafterSCI.A,ImmediatelyafterSCI,C5alevelsinplasmaandthespinalcordincrease,resulting inthesubsequentupregulationofintraspinalinflammatorymediators(IL-6,IL-1␤,andCXCL1).They,inturn,recruitperipheralimmunecellstotheinjurysite.TheeffectsofC5aarerelayedthrough C5aR, which is also expressed in the CNS, although its expression kinetics have not been described. B, In wild-type (WT) mice, SCI results in the formation of a lesion core containing infiltrating immune cells. Astrocytes are also recruited to the lesion core to form the glial scar. C,InC5ar knock-out (KO) mice or upon long-term C5aR pharmacological blockade, there are significant increases inlesionsize,demyelination,andrecruitmentofTlymphocytesandgranulocytes.LackofC5aRsignalingreducesastrocyteproliferationandglialscarformationthroughSTAT3signaling.D,Blockade ofC5aRduringthesubacuteinjuryphasepresentsaneffectivetherapeutictimewindowtoreducerecruitmentofmonocyte-derivedmacrophages(mo-derivedm⌽),demyelination,andlesionsize. E, Given the dichotomous effect of C5aR in SCI, it would be useful to explore whether delayed intrathecal (i.t) administration of C5a could promote recovery by increasing proliferation of STAT3-dependent astrocytes, which in turn, will prevent the spread of secondary injury.

1999). Although Brennan et al. (2015) important for astrocyte proliferation and acute injury phase. Importantly, in vitro show that ubiquitous C5aR deficiency eventual recovery of function. data show that C5aR is involved in induc- ameliorates postinjury induction of in- The study by Brennan et al. (2015) ing proliferation of STAT3-expressing as- flammatory cytokines (Brennan et al., provides interesting insights that could trocytes, which have been shown to limit 2015, their Fig. 4A–F), and that systemic fuel future research on three main fronts, the spread of and neuronal pharmacological blockade of C5aR signif- with time after SCI as the common theme. loss in the injured spinal cord by contrib- icantly reduces peripheral immune cell re- First, astrogliosis is not always harm- uting to scar formation (Okada et al., cruitment to the injured spinal cord at 35 ful. Although it has often been viewed as 2006; Herrmann et al., 2008; Wanner et dpi (Brennan et al., 2015, their Fig. 3E,G), detrimental to the outcome of SCI due to al., 2013). This suggests that C5aR signal- their bone-marrow chimera results sug- its inhibitory effect on axonal regrowth ing might favor proliferation of protective gest that these effects are CNS-dependent. (Menet et al., 2003), recent studies have astrocytes after SCI. Together, the study From these data, one can conclude that challenged this notion, emphasizing that highlights the importance of harnessing acute C5aR signaling in CNS cells contrib- astrocytes are heterogeneous and some as- the beneficial effects of astrogliosis for de- utes to the onset of intraspinal inflamma- trocyte activation is neuroprotective veloping successful therapies for SCI. tion after SCI. Given the role of CNS cells (Faulkner et al., 2004). For example, in a Second, C5aR signaling might pro- in initiating C5aR signaling after SCI, it recent SCI study, astrocytes that were lo- mote regeneration after SCI. One of the would have been useful for the authors to cated in the scar border surrounding the obstacles to overcome in the treatment of quantify the temporal C5aR expression in lesion core were described as elongated, CNS injury is resistance of the mamma- the spinal cord postinjury. Although they proliferating cells, whereas the hypertro- lian CNS to regenerate. Interestingly, C5a include histology of C5aR expression on phic astrocytes showing less proliferation has been implicated as an instrumental microglia/ and astrocytes at were found Ͼ1 mm away from the lesion mediator of hepatocyte proliferation and 1 and 7 dpi (Brennan et al., 2015, their Fig. core (Wanner et al., 2013). In their study, regeneration via activation of the 1), similar data at later time points are not Brennan et al. (2015) allowed for protec- STAT3 pathway (Strey et al., 2003). It is provided. This would have been particu- tive C5aR-mediated astrocyte prolifera- therefore conceivable that C5aR signaling larly useful after the subacute injury phase tion without causing excessive astrogliosis could have a regenerative role after CNS when C5aR signaling is proposed to be by limiting C5aR inhibition to the sub- injury. Indeed, a study by Guo et al. Ulndreaj et al. • Journal Club J. Neurosci., September 9, 2015 • 35(36):12325–12327 • 12327

(2013) reported that C5a enhanced neu- In summary, the pivotal role of the im- Li L, Xiong ZY, Qian ZM, Zhao TZ, Feng H, Hu S, ronal outgrowth in vitro and improved mune system in the progression of SCI is Hu R, Ke Y, Lin J (2014) Complement C5a is functional recovery upon its delayed ad- well established. However, SCI-induced detrimental to histological and functional lo- comotor recovery after spinal cord injury in ministration after SCI in mice. In combi- immune activation is not an all or none mice. Neurobiol Dis 66:74–82. CrossRef nation with this literature, the findings by response that kills or repairs injured tis- Medline Brennan et al. (2015) suggesting a dual, sue. Instead, it is an intricate and complex Menet V, Prieto M, Privat A, Gime´nez y Ribotta M time-dependent role of C5aR create op- process that requires precise tuning to (2003) Axonal plasticity and functional re- timism that exogenous stimulation of harness its beneficial effects. Brennan et covery after spinal cord injury in mice defi- C5aR may further enhance recovery in al. (2015) reveal that C5aR has an impor- cient in both glial fibrillary acidic protein and tant CNS-restricted and time-dependent vimentin genes. Proc Natl Acad Sci U S A 100: SCI by direct promotion of regeneration 8999–9004. CrossRef Medline (Fig. 1E). role in recovery after experimental SCI, Okada S, Nakamura M, Katoh H, Miyao T, Third, SCI outcomes appear to be pri- and these findings should encourage fu- Shimazaki T, Ishii K, Yamane J, Yoshimura A, marily influenced by central C5aR signal- ture research to avoid traditional black- Iwamoto Y, Toyama Y, Okano H (2006) ing, and its therapeutic manipulation can or-white approaches in the treatment of Conditional ablation of Stat3 or Socs3 dis- therefore be restricted to the CNS and so SCI. closes a dual role for reactive astrocytes after spinal cord injury. Nat Med 12:829–834. should its therapeutic manipulation. References CrossRef Medline Brennan et al. (2015) present the novel Beck KD, Nguyen HX, Galvan MD, Salazar DL, Peterson SL, Anderson AJ (2014) Complement finding that inhibiting C5aR signaling Woodruff TM, Anderson AJ (2010) Quanti- and spinal cord injury: traditional and non- can be neuroprotective only when these tative analysis of cellular inflammation after traditional aspects of complement cascade effects are confined to the CNS. Thus, fu- traumatic spinal cord injury: evidence for a function in the injured spinal cord microenvi- ture therapeutic approaches could focus multiphasic inflammatory response in the ronment. Exp Neurol 258:35–47. CrossRef Medline on strategies that manipulate C5aR sig- acute to chronic environment. Brain 133:433– 447. CrossRef Medline Popovich PG, Guan Z, Wei P, Huitinga I, van naling exclusively in the CNS, to avoid Brennan FH, Gordon R, Lao HW, Biggins PJ, Tay- Rooijen N, Stokes BT (1999) Depletion of off-target effects in the periphery. More- lor SM, Franklin RJ, Woodruff TM, Ruiten- hematogenous macrophages promotes partial over, the notion among spine surgeons berg MJ (2015) The complement receptor hindlimb recovery and neuroanatomical re- that “time is spine” regarding the manage- C5aR controls acute inflammation and astro- pair after experimental spinal cord injury. Exp ment of human SCI is also reflected in the gliosis following spinal cord injury. J Neurosci Neurol 158:351–365. CrossRef Medline Strey CW, Markiewski M, Mastellos D, Tudoran work by Brennan et al. (2015), where im- 35:6517–6531. CrossRef Medline Faulkner JR, Herrmann JE, Woo MJ, Tansey KE, R, Spruce LA, Greenbaum LE, Lambris JD mediate and transient blockade of C5aR Doan NB, Sofroniew MV (2004) Reactive (2003) The proinflammatory mediators signaling resulted in persisting improve- astrocytes protect tissue and preserve function and C5a are essential for liver regeneration. J ment (Fig. 1D). Most importantly, by em- after spinal cord injury. J Neurosci 24:2143– Exp Med 198:913–923. CrossRef Medline phasizing the dichotomous role of C5aR 2155. CrossRef Medline Wanner IB, Anderson MA, Song B, Levine J, Fer- in the progression of SCI, the results of Guo Q, Cheng J, Zhang J, Su B, Bian C, Lin S, nandez A, Gray-Thompson Z, Ao Y, So- Zhong C (2013) Delayed post-injury ad- froniew MV (2013) Glial scar borders are Brennan et al. (2015) should prompt fu- ministration of C5a improves regeneration formed by newly proliferated, elongated as- ture studies to investigate whether exoge- and functional recovery after spinal cord in- trocytes that interact to corral inflammatory nous C5aR stimulation in the chronic jury in mice. Clin Exp Immunol 174:318– and fibrotic cells via STAT3-dependent mech- phase will improve outcomes after SCI 325. Medline anisms after spinal cord injury. J Neurosci 33: (Guo et al., 2013). This approach could Herrmann JE, Imura T, Song B, Qi J, Ao Y, 12870–12886. CrossRef Medline provide a dual therapeutic strategy to en- Nguyen TK, Korsak RA, Takeda K, Akira S, Weerth SH, Rus H, Shin ML, Raine CS (2003) Sofroniew MV (2008) STAT3 is a critical Complement C5 in experimental autoim- hance recovery in SCI patients and ad- regulator of astrogliosis and scar formation af- mune encephalomyelitis (EAE) facilitates re- dress the long-standing need for ter spinal cord injury. J Neurosci 28:7231– myelination and prevents gliosis. Am J Pathol treatment of chronic SCI (Fig. 1E). 7243. CrossRef Medline 163:1069–1080. CrossRef Medline