Improved locomotor recovery after contusive spinal cord injury in Bmal1−/− mice is associated with protection of the blood spinal cord barrier

The transcription factor BMAL1/ARNTL is a non-redundant component of the clock pathway that regulates circadian oscillations of gene expression. Loss of BMAL1 perturbs organismal homeostasis and usually exacerbates pathological responses to many types of insults by enhancing oxidative stress and inflammation. Surprisingly, we observed improved locomotor recovery and spinal cord white matter sparing in Bmal1−/− mice after T9 contusive spinal cord injury (SCI). While acute loss of neurons and oligodendrocytes was unaffected, Bmal1 deficiency reduced the chronic loss of oligodendrocytes at the injury epicenter 6 weeks post SCI. At 3 days post-injury (dpi), decreased expression of genes associated with cell proliferation, neuroinflammation and disruption of the blood spinal cord barrier (BSCB) was also observed. Moreover, intraspinal extravasation of fibrinogen and immunoglobulins was decreased acutely at dpi 1 and subacutely at dpi 7. Subacute decrease of hemoglobin deposition was also observed. Finally, subacutely reduced levels of the leukocyte marker CD45 and even greater reduction of the pro-inflammatory macrophage receptor CD36 suggest not only lower numbers of those cells but also their reduced inflammatory potential. These data indicate that Bmal1 deficiency improves SCI outcome, in part by reducing BSCB disruption and hemorrhage decreasing cytotoxic neuroinflammation and attenuating the chronic loss of oligodendrocytes.


Supplementary Methods
Experimental design: sample size determination. For analysis of locomotor recovery, a priori power calculations were performed. From BBB data on hindlimb function recovery in rats that were collected at Louisville over the past 27 years, power analyses show the ability to detect a significant difference is >95% with a sample size of 10/group, based on the observed standard deviations between 1.2-1.9 and an effect size of 10% (e.g. difference between group scores). As the variance of BBB and BMS evaluations are similar 1,2 , functional analyses using n=10 Bmal1 -/and n=14 WT mice were adequately powered to detect at least 10% change in BMS (99.9% power to detect the observed BMS difference of 1.59 at p<0.05). For all other studies, no statistical tests were used to predetermine sample size. Instead, sample sizes were rationalized by considering sufficient replication (weighing the level of biological variation) as well as censoring due to inadvertent losses of animals or samples. In most of those cases a priori design was not possible due to lack of reliable prior data to define variability.
Experimental design: gender. In studies with WT mice (Figs. 1-3), females were used. Females are predominantly used in SCI literature due to lower incidence of urinary tract infections and, therefore, have better survival as compared to males. Thus, use of females helps to reduce the overall animal number needed for SCI experiments. As Bmal1 -/mice are sterile, those animals have to be generated by breeding heterozygous parents thereby limiting availability of suitable experimental subjects. Therefore, to properly power experiments that involved Bmal1 -/mice (Figs. 4-8) we used both sexes. Such a strategy is justified by a documented lack of significant sex effects on locomotor recovery, neuroinflammation or lesion volume after contusive injury in C567Bl6 mice 3 . However, sex unbalanced groups emerged in our studies on Bmal1 -/mice due to (i) limited availability of males of comparable age, (ii) peri-operative loss of animals, (iii) initial genotyping errors that were later corrected by terminal genotyping and resulted in group re-assignments.
Group allocation and blinding. Group allocation was not random. Instead, biological controls were used in all experiments. Such a strategy was necessary due to limited number of Bmal1 -/animals and the need to pair them with comparable WT control animals including similar age, sex and, if possible, same origin. For each study surgeries were performed on the same day or two consecutive days (ZT3-ZT5) with random order of animals and without knowledge of group assignment. Likewise, all behavioral and histological analyses were blinded.
RNASeq analysis. Spinal cord tissue that was isolated from Bmal1 -/-(n=3, 2 females, 1 male) and WT mice (n=4, 2 females, 2 males) was homogenized on ice with Tissue-Tearor (BioSpec Products, Bartlesville, OK). Total RNA was extracted using RNeasy Lipid Tissue Minikit (Qiagen #74804) according to the manufacturer's instructions. The quality of RNA was accessed by capillary electrophoresis using an Agilent Bioanalyzer. One microgram of total RNA was used for poly-A enrichment. First and second cDNA strands were synthesized followed by 3' end adenylation. Libraries were prepared using the TruSeq Stranded mRNA Library Prep Kit along with TruSeq RNA Index Set A according to manufacturer's instructions (Illumina, San Diego, CA). Samples were barcoded with Illumina TruSeq Adapters, DNA fragments were enriched by PCR reaction and their quality was validated on an Agilent Bioanalyzer. All 1.8 pM libraries were then denatured and sequencing was performed at the University of Louisville Genomics Core Facility Illumina NextSeq 500 using the NextSeq 500/550 75 cycle High Output Kit v2.5 (Illumina, Carlsbad, CA). Raw read number ranged from 50,185,248 to 57,727,458 with an average of 52,643,643 reads/sample. The quality control of the raw sequence data was performed using FastQC v.0.10.1 and the sequences were directly aligned to the Mus musculus reference genome assembly (GRCm38.p6.fa) using Tophat 2v.2.0.13. At least 97.6% reads were aligned to the mouse reference genome (average read alignment was 98.1%). For the DESeq2 analysis of differential gene expression raw counts were obtained from the Tophat aligned bam format files using HTSeq v.0.10.0. The raw counts were normalized using DESeq2's default method, relative log expression (RLE). A q-value < 0.05 was used as a criterion for defining differentially expressed genes.
WMS analysis. WMS was evaluated as described previously 4,5 . Briefly, serial transverse sections spanning -1 mm rostral to +1 mm caudal from the lesion epicenter were stained for myelin using iron eriochrome cyanine (EC) with an alkali differentiator modification. Images were captured with a 4x objective on a Nikon Eclipse Ti inverted microscope. Myelinated, EC + WM was traced using Nikon Elements software. The injury epicenter was identified based on the least relative content of WM as defined by EC + area/total section area. Percent sparing was calculated after normalizing the data to average EC + WM content in naïve female C57BL6/J mice (n=4, T8-T10 level, 8-10 week old). All imaging and analysis was performed without knowledge of genotype or treatment.
OL content and axonal density analysis. The procedure followed previously described methodology with minor modifications 6 . Briefly, transverse 20 μm sections were co-stained with the CC1 (OL marker) and anti-NFH (axonal marker) antibodies. Nuclei were counterstained with Hoechst-33258. The sections were selected based on results of WMS analysis that was done on a set of adjacent sections to define the injury site. Following image capture (Zeiss Observer.Z1, 10x objective, identical exposures) digitalized pictures were saved as gray scale jpeg files. A grid of 0.01 mm 2 squares (100 µm x 100 µm) was overlaid on the images over the ventral/ventrolateral WM. The numbers of CC1 + OLs and all cells (defined by Hoechst-stained nuclei) were counted manually in ventral (V) and ventrolateral (VL) WM on each side at the injury epicenter and -1 mm rostrally or +1 mm caudally. At -1 mm or +1 mm V WM was defined as WM located between the tips of the ventral horns and VL WM was defined as WM lateral to the ventral horns and ventral to a coronal plane line through the middle point of the central canal. In the epicenter region, V WM was defined as an area between transverse or sagittal plane lines that were tangential to ventral or lateral borders of the grey matter lesion, respectively. The epicenter VL WM was located between the coronal and sagittal plane lines that intersected the middle point of a section and were tangential to a lateral border of the grey matter injury site, respectively. The grey matter injury site was identified as an area that contained densely packed cells of the fibrotic scar, as visualized with Hoechst. For each animal, two sections from each region were analyzed; in each section all cells in the V-or VL WM were counted. The average total number of cells counted per animal was 466 or 426 or 502 for V WM and 692 or 629 or 738 for VL WM at the epicenter, -1 mm, or, +1 mm locations, respectively.
Axonal density was quantified using the same sections and identical region definitions as described for cell counting. Each region was traced on both sides and the areas of the positive NFH signal and Hoechst-stained nuclei were determined. The area measurements were performed using the NIH ImageJ after adjusting the "threshold" parameter to separate a positive signal from the background. The threshold parameter was manually adjusted for each section to cover exclusively positive signals of NFH or Hoechst. The ratio of pixels above the threshold to all pixels was calculated for both NFH and Hoechst and then presented as the NFH/Hoechst ratio (axonal density). The imaging, as well as image processing and analysis including cell counting, were done without knowledge of the genotype.   and SCI mice at dpi 1 (animals and sections were as described for Fig. 2; SCI images are from the injury penumbra region, 0.5-1 mm from the injury epicenter). Arrows and arrowheads point to BMAL1 + neurons. Figure S3. BMAL1 is expressed in spinal cord astrocytes. Low power images of coimmunostaining for BMAL1 and the astrocyte marker GFAP in WT spinal cord of sham and SCI mice at dpi 1 (animals and sections were as described for Fig. 2; SCI images are from the injury penumbra region, 0.5-1 mm from the injury epicenter). Arrows and arrowheads point to BMAL1 + astrocytes. Figure S4. BMAL1 is expressed in spinal cord OLs. Low power images of co-immunostaining for BMAL1 and the OL marker epitope CC1 in WT spinal cord of sham and SCI mice at dpi 1 (animals and sections were as described for Fig. 2; SCI images are from the injury penumbra region, 0.5-1 mm from the injury epicenter). Arrows and arrowheads point to BMAL1 + OLs. In SCI tissue, positive CC1 staining in grey matter neuron-like cells is likely non-specific (asterisks).       Table S1. List of mRNAs whose expression is altered in Bmal1 -/mice on dpi 3 (a separate MS Excel spreadsheet is provided).  1a) WT sham@6h WT sham@24h WT SCI@6h WT SCI@24h 2 (0:2) 3 (0:3) 4 (0:4) 4 (0:4) None Western blot analysis of BMAL1 expression after SCI (fig 1b, c) WT sham@12h WT sham@24h WT SCI@12h WT SCI@24h