Complement C6 deficiency exacerbates pathophysiology after spinal cord injury

Historically, the membrane attack complex, composed of complement components C5b-9, has been connected to lytic cell death and implicated in secondary injury after a CNS insult. However, studies to date have utilized either non-littermate control rat models, or mouse models that lack significant C5b-9 activity. To investigate what role C5b-9 plays in spinal cord injury and recovery, we generated littermate PVG C6 wildtype and deficient rats and tested functional and histological recovery after moderate contusion injury using the Infinite Horizon Impactor. We compare the effect of C6 deficiency on recovery of locomotor function and histological injury parameters in PVG rats under two conditions: (1) animals maintained as separate C6 WT and C6-D homozygous colonies; and (2) establishment of a heterozygous colony to generate C6 WT and C6-D littermate controls. The results suggest that maintenance of separate homozygous colonies is inadequate for testing the effect of C6 deficiency on locomotor and histological recovery after SCI, and highlight the importance of using littermate controls in studies involving genetic manipulation of the complement cascade.

Traumatic spinal cord injury (SCI) causes a series of secondary degenerative events thought to exacerbate injury pathology [1][2][3] . One component of secondary injury after SCI is activation of the classical and alternative pathways of the complement cascade 1,4 . The complement cascade is composed of more than 30 soluble and membranebound proteins. The classical and alternative pathways converge on the terminal complement pathway, leading to the assembly of a multimeric protein complex, C5b-9, for which C6 is a necessary component 5,6 .
To investigate the role of C6 in SCI, we initially acquired Piebald-Viral Glaxo (PVG) C6-deficient (C6-D) rats from the University of Wales and PVG WT rats from Harlan laboratories. PVG C6-D rats carry a spontaneous mutation resulting in a lack of the C6 protein and are incapable of C5b-9 formation 2,9 . While heterozygous colony breeding to generate littermate controls for mouse transgenic studies is the standard for the field, this approach is not commonly applied for comparison of C6 deficiency in PVG rats. A literature search identifying papers that

Results
Confirmation of C6-deficiency. A PCR genotyping protocol was established for C6 deficiency by blast identification of a novel restriction site susceptible to Ava I cleavage and development of primer sequences enabling identification of restriction fragment lengths that could distinguish WT, heterozygous, and C6-D genotypes, as described under methods. In this protocol, WT (C6 +/+ ) DNA is not cut and runs as a single band of ~ 200 bp, heterozygous (C6 +/− ) DNA exhibits incomplete digestion with two bands at ~ 150, and 200 bp, and homozygous C6D (C6 −/− ) DNA exhibits complete digestion with one band at ~ 150 bp (Fig. 1b). The genotyping protocol was validated using rats sourced from homozygous and heterozygous breeding colonies, and by comparison with CH50 assay for dose-dependent lysis of IgG sensitized sheep erythrocytes (Fig. 1a). To further confirm that C5b-9 was present only in F2-C6-WT animals and not in F2-C6-D animals immunohistochemistry using an anti-C5b-9 antibody was conducted 54,55 . C5b-9 was observed only in the gray and white matter of F2-C6-WT SCI animals, and not in either F2-C6-D animals or sections for which the primary was omitted (Fig. 2).
C6 deficiency impairs locomotor function in F2-C6 PVG rats but not H-C6 PVG rats. For our initial experiment, rats were maintained as separate homozygous colonies as previously reported by many investigators (H-WT or H-C6-D PVG; homozygous colony cohort). In a subsequent experiment, we generated experimental animals from F2 crosses to yield littermate controls (F2-WT and F2-C6-D PVG; littermate control cohort). Animals from both cohorts received SCI at 4-6 months of age and were within 10% variation in weight (range 169. 2-195.8 g). Multiple tasks were used to assess functional locomotor recovery after SCI. BBB is an open-field locomotor test that assesses gross functional improvements such as stepping and coordination. The higher a BBB score, the more functional the animal. Analysis in the homozygous colony cohort showed that on average, H-C6-D rats exhibited improved outcome vs. H-WT rats, as seen by higher BBB scores, suggesting that Variation in behavioral recovery kinetics and plateau was observed across these sequential studies, with WT littermate control cohort animals performing significantly better than WT homozygous colony animals in BBB and worse in horizontal ladder beam testing. While this could suggest that homozygous colony and littermate control WT rats responded differently to SCI, these cohorts were injured more than 2 years apart, and could therefore have been affected by variables that cannot be excluded. For this reason, a conservative data analysis approach was taken, and no statistical comparisons between cohorts are made in these figures.
Horizontal ladder beam is most sensitive to recovery of function after SCI in the range of stepping recovery 56 . Because of the high performance plateau in the littermate cohort, we employed a supplemental assessment, Catwalk kinematic analysis, to assess fine parameters of gait after SCI in addition to stepping function (Fig. 4). Data were normalized to pre-injury data (dashed lines). One-sample t tests were used to compare baseline (dashed line) and post-injury performance; the closer the group mean is to baseline the better recovery post-SCI. Using this analysis, F2-C6-D rats consistently exhibited group means that showed decrements in recovery in comparison with WT animals. Both F2-C6 WT and F2-C6-D rats exhibited a significant decrease in print area, indicative of weight bearing, the higher the number, the more weight bearing, from pre-injury baseline (Fig. 4a, one sample t test # p < 0.05), suggesting less weight support; however, F2-C6-D rats exhibited a significant further decrement in comparison with F2-C6-WT (Fig. 4a, Student's t test *p = 0.02). A second parameter analyzed was mean swing time (the time interval between each consecutive placement of the same paw), for which a higher number indicates a deficit. While F2-C6 WT rats showed no difference in mean swing time vs. pre-injury baseline (Fig. 4b, one sample t test p > 0.05), F2-C6-D rats exhibited an increase in comparison with both pre-injury baseline (Fig. 4b, one sample t test # p < 0.05) and F2-C6-WT animals (Student's t test *p = 0.04). These data indicate a slowed step cycle and decrement in recovery compared to F2-C6 WT. A third parameter analyzed was duty cycle (the percentage of time spent standing within a step cycle), which is indicative of weight bearing. F2-C6-WT rats did not exhibit a significant decrease from pre-injury baseline for duty cycle, but F2-C6-D rats again demonstrated www.nature.com/scientificreports/ a significant decrement in comparison with both pre-injury baseline (Fig. 4c, one sample t test # p < 0.05), and F2-C6 WT rats (Fig. 4c, Student's t test *p = 0.03). The fourth parameter analyzed was regularity index (a measure of coordination). Both F2-C6-WT and F2-C6-D rats were significantly impaired vs. pre-injury baseline levels for regularity index (Fig. 4d, one sample t test # p < 0.05); while not reaching significance, a strong trend for a further decrease in F2-C6-D rat was observed in comparison with F2-C6 WT rats (Fig. 4d, Student's t test p = 0.08).
C6 deficiency alters white matter sparing and lesion volume in homozygous strain H-C6 PVG but not in littermate control F2-C6 PVG rats. We assessed whether C6 deficiency affected percent white matter sparing (Fig. 5a) or lesion volume (Fig. 5b) within the homozygous colony and littermate control cohorts using unbiased stereology. As noted above, statistical comparisons between cohorts should be approached with caution, because these cohorts were run at different times; however, no statistical differences were observed between H-WT and F2-WT rats for either SCI-induced lesion volume (H-WT mean 2.360 ± 0.3300 SEM N = 8), F2-WT 2.873 ± 0.2020 SEM N = 8; p = 0.2061; 2-tailed t test) or percent spared white matter (H-WT mean   (Fig. 5c). In contrast, this measurement was unchanged between F2-C6-D and F2-C6 WT rats (Fig. 5d, Student's t test, p = 0.4). Parallel analysis of central lesion volume in the homozygous strain cohort found no difference between H-C6-D vs. H-C6-WT rats (Fig. 5e). However, again in contrast, animals in the littermate control cohort exhibited a strong trend towards increased lesion size in F2-C6-D vs. F2-C6-WT rats (Fig. 5f, Student's t test, p = 0.08). Taken together these There was also a postinjury trend towards a decrease in (d) regularity index score, indicative for coordination, for F2-C6-D rats compared to F2-C6 WT rats. Collectively, these data suggest that F2-C6-D rats, when normalized to baseline levels, and compared to F2-C6 WT rats, have decreased weight bearing ability on affected paws and decreased gait coordination. Data points represent group means ± SEM. www.nature.com/scientificreports/ www.nature.com/scientificreports/ histological data demonstrate divergent responses in lesion pathogenesis as well as locomotor recovery between homozygous colony and littermate control cohorts.

Discussion
Here, we have compared the recovery differences of rats that were either maintained as separate homozygous colonies or generated from F2 crosses to yield littermate controls. Within cohort comparisons demonstrate that H-C6-WT vs. H-C6-D rats and F2-C6-WT vs. F2-C6-D rats have different results for both recovery of function and lesion pathogenesis after SCI. These results highlight the importance of the use of littermate controls in studies involving genetic manipulation of the complement cascade. While future studies should include comparisons between the different cohorts within the same experiment to also test for baseline differences, the current data indicate that there are functional differences due to maintenance in separate homozygous colonies versus heterozygous colonies to generate littermate controls. The standard protocol in the literature has been to utilize animals from separate homozygous colonies for comparisons of C6 deficiency. A literature search identifying papers utilizing C6 WT and C6-D rats as a component of the experimental work identified 18 studies, a majority of which reported obtaining WT and C6 deficient rats from separate sources, highlighting a need to review findings using this model. Because of identified differences in complement sufficiency and activation between species, the C6-D PVG rat has been an important model for the study of complement and inflammation. C57Bl/6 mice, the most common background for genetically manipulated mice, have dramatically reduced C5b-9-hemolytic capacity in comparison with rats 57,58 . Indeed, we have previously reported that male C57Bl/6 mice fail to exhibit hemolytic activity in CH50 assays after SCI in comparison with male BUB mice or Sprague Dawley rats 58 . Accordingly, while transgenic mice can give insight into innate immune mechanisms in inflammation and injury, mouse models may have gaps with regard to the initiation of downstream immune activation. Together, these points highlight the importance of conducting studies in a rat model in which bred-in genetic variation between control and knockout/deletion animals is addressed.
C6-D PVG rats have been previously characterized and reported to derive from a specific autosomal recessive spontaneous mutation 59,60 ; in parallel, tissue antigens, serum levels of other complement proteins, and spleen cell proliferation of C6-D rats have been reported to be identical to that of WT rats sharing the PVG background. Finally, these animals also do not exhibit either evidence of increased vulnerability to infection within a controlled vivarium environment, or evidence of differences at the level of immunity based on mixed lymphocyte reaction and alloantibody testing. However, it should be noted that the data presented cannot exclude the possibility of potential variabilities in blood or other immune parameters at baseline between colonies. Additionally, strain differences in SCI lesion pathogenesis are well recognized [61][62][63][64][65][66][67] . It would thus not be surprising that nonlittermate WT and C6-D rats maintained in separate homozygous colonies for generations have diverged, resulting in different functional strain characteristics. In addition to genetic drift with colony maintenance practices, there are a number of reasons why maintenance of animals in separate colonies could yield such a result. For example, obtaining animals from different sources or housing in separate locations may lead to differences within the gut microbiome. Villarino et al. observed that C57Bl/6 mice that were sourced from different vendors, when exposed to the malaria parasite, differed in their severity response 68 . Gut microbiota has also been implicated in response to experimental autoimmune encephalomyelitis 69 , SCI 70,71 , and other CNS injury such as TBI 72 . In these studies, however, rats for both C6 WT and C6-D were bred in the same vivarium and fed the same chow; accordingly, the more likely scenario is that the differences in behavior and histology between the homozygous C6 WT and C6-D rat colony and heterozygous rat colony are due to the accumulation of genetic differences.
The results reported here suggest a surprising positive role for C5b-9 after nervous system injury, and are in contrast to studies have evaluated complement deficiency or knockout in SCI mice that are not complement sufficient or not bred to establish littermate controls. A positive role for C5b-9 in recovery after SCI could be consistent with the established role of complement in debris clearance, which is critical for regeneration and plasticity in the CNS. Studies have suggested C5b-9 presence can amplify the complement cascade through a positive feedback loop with C1q 10,19 . Amplification increases the recruitment of macrophages by an increase in C3a and C5a, which would be predicted to promote clearance of myelin debris that could otherwise inhibit remyelination. The presence of myelin and other debris may also indirectly increase stress or toxicity to, or impair the function of, surviving cells, leading to increased cell death as well as delayed remyelination [73][74][75][76][77][78] , negatively impacting motor recovery.
In sum, these results suggest that maintenance of separate homozygous colonies is inadequate for testing the effect of C6 deficiency on locomotor and histological recovery after SCI, and that comparison of F2-WT and F2-C6-D PVG rats provides a more appropriate model for assessing the role of complement in CNS injury. Moreover, these data point to a previously unidentified finding, that C5b-9 formation can exert beneficial functions after SCI, and presumably other types on CNS injury and damage.

Materials and methods
Animal numbers and exclusion criteria. For experiments comparing PVG rats maintained as separate homozygous colonies, PVG wildtype rats were purchased from Harlan (Indianapolis, Indiana) and C6 Deficient (C6-D) rats on a Piebald-Viral Glaxo (PVG) background were generously provided by Paul Morgan (University of Wales College of Medicine, U.K.). These strains are abbreviated as homozygous wildtype (H-WT) or homozygous C6-D-PVG), respectively. From these homozygous litters, H-C6 WT (n = 14) and H-C6-D (n = 12) females were used to assess motor function after spinal cord injury. For experiments comparing PVG rats generated as littermate controls, PVG rats WT purchased from Harlan were crossed with the homozygous PVG C6-D rats from the University of Wales, to generate F1 heterozygote C6-D rats at the University of California, Irvine; heterozygous males and females were then crossed to produce F2 litters with wildtype (F2-WT), heterozygous, www.nature.com/scientificreports/ and C6-D PVG rats (F2-C6-D PVG). F2-WT (n = 11) and F2-C6-D (n = 10) females were used to assess motor function after spinal cord injury. Only females were used in this study to facilitate manual bladder expression after SCI. 3 (WT, n = 2 and C6-D, n = 1) animals were excluded from final analysis; all exclusions were based on Grubbs test identification of outliers in behavioral data and/or post-injury weight monitoring.
Genotyping. WT and C6-D genotypes were determined using PCR of DNA samples extracted from tail clips (Qiagen DNeasy Tissue Kit) followed by restriction enzyme digest and agarose gel electrophoresis. PCR was conducted using Taq PCR Master Mix (Qiagen) for 38 cycles: 4 m 94 °C; 45 s 94 °C; 1 m 50 °C; 1 m 72 °C; 10 m 72 °C. Primer sequences were as follows: C6 forward: TGC AGT AGG AAT GGG GCT AA; C6 reverse: GAG AAA AGA GGC ATT CCC AGT. PCR products were purified (Qiagen QIAquick PCR purification kit, DNA eluted in 30 µL NEB buffer, and 15 µl of eluted DNA product digested with 1 µl Ava I (New England Biolabs) overnight at 37 °C. PCR product digests were loaded onto a 1.5% agarose gel for electrophoresis and genotype identification relative to 100 bp DNA ladder as shown in Fig. 1a. Injury model. All experiments were conducted in accordance in an AALAC accredited vivarium with the UCI Institutional Animal Care and Use Committee (IACUC) approval. Rats were anesthetized with isofluorane gas anesthesia. A laminectomy was performed at thoracic vertebrae 9 (T9) and given 200 kilo dyne (kd) contusion injuries at T9 using the Infinite Horizon Impactor (Precision Systems and Instrumentation, Lexington, KY). After the contusion injury, a small piece of gel-foam was placed over the laminectomy location and the muscles overlying the injury was sutured and skin was closed using metal wound clips. Post-operative care included twice daily manual bladder expression until animals regained voluntary micturition, as well as administration of antibiotics and buprenorphine, as described previously 79 .
Behavioral assessments. All behavioral testing and scoring was performed by individuals blinded to the experimental group and genotype.
Open field locomotion. The Basso Beattie Bresnahan (BBB) scale was utilized to evaluate gross locomotor recovery in injured animals weekly starting from 7 days post-injury (dpi) up to 42 dpi 80 . The BBB scale goes from 0 to 21 with the higher the BBB score the more functional the animal.
Horizontal ladderbeam. The horizontal ladder beam task was used as a terminal supplemental quantitative task with higher the number of errors, the less functional the animal 56,81-85 . Animals were tested at the end of the study prior to sacrifice. Prior to testing and video acquisition, rats were trained in a 20-30 m acclimation session. The apparatus length was 94.5 cm with 20 rungs spaced 2.54 cm apart. During the testing session, rats were videotaped from below; quantification of hindlimb errors was scored by slow motion video playback. Hindlimb errors from three test trials per rat were averaged for subsequent statistical analysis.
CatWalk gait assessment. Catwalk XT (version 7.1 Noldus Information Technology) was used to assess fine details of gait changes and allowed for easy quantification of a large number of gait parameters while animals are recorded crossing a walkway. Prior to testing and video acquisition, rats were trained in an acclimation session. Animals were recorded crossing the catwalk in the dark and data from three test trial per rat were averaged for statistical analysis.
Tissue collection, sectioning and immunohistochemistry. At time of sacrifice, animals were perfused with 4% phosphate-buffered saline (PBS), followed by 4% paraformaldehyde in PBS. The T6-T12 segment of the spinal cord was dissected by the dorsal root to provide an anatomically defined region for stereology. Cords were further post-fixed in 20% sucrose/4% paraformaldehyde in PBS overnight at 4 °C. Spinal cords were frozen in isopentane at − 65 °C and stored at − 80 °C. Serial 30 μm thick coronal cryostat sections were cut on a CryoJane tape-transfer system (Leica Microsystems Inc., Cuffalo Grove, IL) at − 21 °C and collected in sets of 12 for immunohistochemistry.
White matter sparing and lesion volume. Slides were processed for antigen retrieval using a Buffer A (pH 6) in the Retriever 2100 system (PickCell Laboratories, Amsterdam, The Netherlands) before immunohistochemical staining. To assess white matter sparing, slides (1/12 section sampling) were dehydrated through a series of ethanol solutions (50%, 1 min; 70%, 3 min; 95%, 3 min; 100%, 5 min) and then rehydrated back through the alcohol series with fresh ethanol starting at 100% to 50% prior to incubation with anti-myelin basic protein primary antibody overnight (EMD Millipore). Anti-MBP was used at 1:750 and visualized with diaminobenzidine (DAB, Vector Labs, Carpinteria, CA). To assess central lesion volume, slides (1/12 section sampling) were incubated with anti-GFAP overnight (Dako). Anti-GFAP was used at 1:10,000 and visualized with diaminobenzidine (DAB, Vector Labs, Carpinteria, CA).
C5b-9 formation. Slides were processed for antigen retrieval using a trypsin buffer before immunohistochemical staining. To assess C5b-9 formation, mouse monoclonal antibody to Human SC5b-9 (Quidel A239) was used at 1:100 and was visualized with Donkey anti-mouse Alexa-flour 488 Fab (Jackson Immunoresearch) at a 1:500 dilution. A Hoescht counter stain at a dilution of 1:1000 was used to visual nuclei. Images were taken in the ventral horn using the Keyence BZ-X810 and a 60 × objective. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.