Bipartite left-right sided endocrine system: processing of contralateral effects of brain injury

The crossed descending neural tracts set a basis for contralateral effects of brain injury. In addition, the left-right side-specific effects of the unilateral brain lesions may be mediated by neurohormones through the humoral pathway as discovered in animals with disabled descending motor tracts. We here examined if counterparts of the endocrine system that convey signals from the left and right brain injuries differ in neural and molecular mechanisms. In rats with completely transected cervical spinal cords a unilateral injury of the hindlimb sensorimotor cortex produced hindlimb postural asymmetry with contralateral hindlimb flexion, a proxy for neurological deficit. The effects of the left and right side brain lesions were differently inhibited by antagonists of the δ-, κ- and µ-opioid receptors suggesting differential neuroendocrine control of the left-right side-specific hormonal signaling. Bilateral deafferentation of the lumbar spinal cord eliminated hormone-mediated effects of the left-side brain injury but not the right-side lesion suggesting their afferent and efferent mechanisms, respectively. Analysis of gene-gene co-expression patterns identified the left and right side-specific gene regulatory networks that were coordinated across the hypothalamus and lumbar spinal cord through the humoral pathway. The coordination was ipsilateral and perturbed by brain injury. These findings suggest that the neuroendocrine system that conveys left-right side-specific hormonal messages from injured brain is bipartite, contributes to contralateral neurological deficits through asymmetric neural mechanisms, and enables ipsilateral coordination of molecular processes across neural areas along the neuraxis. GRAPHICAL ABSTRACT


INTRODUCTION
In rats with transected cervical spinal cords, UBI induced HL-PA (Figure 1D,E). The PAS and 156 PA in rats with left and right UBI were at least 7-fold greater than in rats before UBI or sham 157 surgery, or after sham surgery. The PA and MPA did not differ between the left-and right-side 158 UBI groups. Differences in the PAS and PA between the UBI sham surgery groups were strong 159 and highly significant (Figure 1F,G). The response to the injury was developed on the 160 contralesional side; the left or right UBI induced the right and left hindlimb flexion, respectively. 161 The PAS and PA in the UBI rats with transected cervical spinal cords were similar to those that 162 were previously reported for the UBI animals with intact and transected spinal cords (Lukoyanov (Figure 1-figure supplement 3). The 172 W and AI may differently depend on the stretching distance, and therefore were both analyzed. 173 Difference in the work between contra-(C) and ipsilesional (I) hindlimbs were compared by  Differences in the work applied to stretch the left and right hindlimbs WLR and AILR were highly 183 significant after left and right UBI while sham surgery did not produce the asymmetry ( Figure   184 1I; Figure 1-figure supplement 3B). Both the left and right UBI groups significantly differed 185 from the sham group in the WLR ( Figure 1J) and AILR (Figure 1-figure supplement 3C). 186 The WCI strongly correlated with the MPA (Figure 1K; WCI vs. the MPA). Thus in rats with 187 completely transected cervical spinal cord both the left and right UBI produced robust and highly 188 significant increase in the musculo-articular resistance to stretching of the contralesional vs. 189 ipsilesional hindlimbs. We concluded that the UBI-induced asymmetries were not mediated by 190 the sympathetic system or descending neural tracts, and that the endocrine pathway may be only 191 an option for the left-right side-specific signaling from injured brain in these experiments.  (Figure 2). In rats with left UBI, the PAS and PA were strongly, 3.0-and 3.5-fold, respectively, 201 reduced after rhizotomy (Figure 2B,C). In contrast, in the right-side UBI rats, the PAS and PA 202 demonstrated only small, approximately 1.3-fold decrease after deafferentation. No signs of the 203 asymmetry were revealed in the sham surgery rats after rhizotomy. 204 Contrast in both the PAS and PA was strong and highly significant between the UBI groups and 205 sham surgery group before the rhizotomy, and between the right UBI and sham surgery group 206 after rhizotomy (Figure 2D,E). In opposite, the left UBIsham surgery contrast was negligible 207 in rats analyzed after deafferentation. Relative impact of rhizotomy (before vs. after it) on the 208 effects of left and right UBI (UBI vs. sham surgery) was analyzed as contrast of contrasts ( Figure   209 2F,G). Contrast in each the MPA and PA between the animal groups (left UBI vs. sham surgery; 210 right UBI vs. sham surgery; and left UBI vs. right UBI) was compared between two time points 211 that were before vs. after rhizotomy. Contrast was high and significant for comparison of the left 212 UBI group with sham surgery group, whereas it was much smaller when the right UBI rats were 213 compared with sham rats. Contrast of contrasts in both the MPA and PA for the left UBI vs. right 214 UBI was strong and significant.

Stretching resistance analysis 216
The stretching resistance of the contra-and ipsilesional hindlimbs in rats with transected cervical 217 spinal cords was analyzed before and after bilateral rhizotomy that was performed 3 hours after rhizotomy abolished the differences between the hindlimbs in the resistance. In contrast, the 220 asymmetry was only slightly decreased after rhizotomy in rats with the right UBI. No rhizotomy 221 effects on the left-right differences were evident in rats with sham surgery. Contrast in both the 222 WLR and AILR between UBI groups and sham surgery group was strong and highly significant 223 before rhizotomy, while after it no differences in the left UBI group were evident ( Figure 2J;   238 We previously demonstrated that naloxone, a nonselective opioid antagonist, blocked the 239 asymmetric effects of the left UBI on hindlimb posture in rats with transected spinal cords, and 240 that the selective opioid antagonists differentially inhibited formation of HL-PA after the left and 241 right UBI in rats with intact spinal cords (Lukoyanov et al., 2021;Watanabe et al., 2021). We 242 here examined if the UBI effects mediated through the humoral pathway are controlled by the 243 opioid system, and if this control is side-and receptor subtype-specific. 244 The effects of the µ-, and -opioid antagonists β-Funaltrexamine (FNA), naltrindole (NTI) 245 and nor-Binaltorphimine (BNI), respectively, and naloxone on the asymmetry in hindlimb to the rats 24 h before the surgeries. NTI and naloxone were administered 3-4 h after UBI to rats 252 that displayed HL-PA with the MPA > 1.5 mm before the injection, and then HL-PA was 253 analyzed 1 h later.

254
In the left UBI group, a marked decrease in the MPA was induced by NTI and BNI (3.2-and 2.5-255 fold, respectively) while no substantial effects of FNA on HL-PA were evident (Figure 3C,D). 256 In contrast, in the right-side UBI group administration of FNA, but not NTI and BNI, resulted in 257 substantial MPA reduction (2.3-fold). Naloxone inhibited the effects of both the left-and right 258 side injuries (2.2 and 3.7-fold, respectively). The effects of the antagonists were significantly 259 different between the left and right UBI groups ( Figure 3E). NTI and BNI preferentially responsiveness of these circuits to a unilateral impact. We previously reported that the 280 "decoding" lumbar spinal cord is characterized by asymmetric gene expression, and that the UBI  284 We here examined if the UBI targets the hypothalamus and pituitary gland as the "encoding" 285 areas by analysis of gene expression as a readout; if the UBI effects are the ipsi-or contralesional; 286 and if expression of neurohormonal and neuroplasticity-related genes is lateralized in this area.

287
To reveal regulatory humoral interactions between the "encoding" hypothalamus and "decoding" 288 spinal cord, we then characterized gene-gene co-expression patterns between these regions in 289 rats with complete spinal cord transection along with their perturbations produced by UBI.

290
The hypothalamic-pituitary system 291 We reasoned that the "encoding" system that mediates the neuroendocrine UBI effects in the   and right hypothalamus) were identified between UBI and sham groups, and these group were 317 combined for analysis of lateralization. Comparison of the AI with zero identified three genes 318 (Pomc, P = 0.009; Trh, P = 0.014; and Ghrh, P = 0.014) with higher expression in the left 319 hypothalamus while other three genes (Grin2b, P = 0.002; GluR1, P = 0.010; and Arc, P = 0.010) 320 showed higher expression on the right side ( Figure 4E)  were higher with nominal significance in the UBI group.

331
Thus in the hypothalamus expression of a subset of the neurohormonal, neuropeptide and 332 neuroplasticity-associated genes was lateralized and affected by UBI on the ipsilesional side.

333
Among genes responded to UBI in the neuroendocrine system were Avp, Avpr1b and Pomc that 334 give rise to Arg-vasopressin, the vasopressin receptor V1B and β-endorphin that mediate the 335 effects of the left UBI on HL-PA through humoral pathway. gene-gene co-expression patterns differ between these networks and for both network between 347 the sides in the hypothalamus and spinal cord; if the patterns are coordinated between these 348 regions; and if the coordination is ipsi-or contralateral and perturbed by a unilateral brain lesion.

509
Patients with stroke and cerebral palsy often do not relax their musclesthey are tonically 510 constricted without any voluntary command. This phenomenon is defined as "stretchand effort- suggests plasticity in the Arg-vasopressin system that signals from the injured brain.  (Figures 6 and 7C). These ensembles may be defined as "allo-symmetric". In     The "left" extract induced flexion of left hindlimb, while the right hindlimb was flexed after 594 administration of peptides from the right hemisphere. No asymmetry was formed after 595 administration of the total peptide pool prepared from the whole brain. Thus, peptides with side-596 specific actions were lateralized in the brain, and the integral activity of the "left" and "right"   The T-NES is bipartite and consists of the left and right counterparts that mediate the effects of 622 the left-side and right-side brain injury, respectively (Figure 7). These counterparts enable 623 overall mirror-symmetric functional responses, e.g., flexion of the right and left hindlimbs,   We are grateful to Dr. Michael Ossipov for discussion and manuscript processing, and Ms. Karen 666 Rich for assistance with histochemical analysis.

Competing interests 668
Vladimir Galatenko is affiliated with Evotec International GmbH, and has no other competing    The animals were anesthetized with sodium pentobarbital anesthesia (intraperitoneal, I.P.; 60 688 mg/kg body weight, as an initial dose and then 6 mg/kg every hour). Core temperature of the 689 animals was controlled using a feedback-regulated heating system.

690
The experimental design included rats with UBI which was preceded by a complete spinal cord were analyzed before UBI, 3 h after UBI before rhizotomy, and 0.5 h after rhizotomy.

724
Histological analysis of brain injury 725 Localization and size of cortical lesions were analyzed in rats with left side (n = 5) and right side 726 (n = 5) UBI 3 -5.5 hours after the injury. After perfusion with 4% paraformaldehyde the brain 727 was removed and postfixed in the same fixative overnight. Then the brain was soaked in 728 phosphate-buffered saline with 30% sucrose for 48 hours, dissected into blocks which were then 729 sliced into 50 µm sections with a freezing microtome. Every fourth section was stained with     genes Actb and Gapdh. GeNorm software was used to analyze the gene expression stabilities (M 822 value) of the ten candidate reference genes (Actb, B2m, Gapdh, Gusb, Hprt, Pgk, Ppia, Rplpo13a, 823 Tbp, and Tfrc). The calculation of M value was based on the pairwise variation between two 824 reference genes. If the M value was less than 1.5, it could be considered a suitable reference gene.   The following genes were selected as neuroplasticity-related including Arc, activity-regulated 845 cytoskeletal gene implicated in numerous plasticity paradigms; Bdnf, brain-derived neurotrophic  (Figure 4-figure supplement 4). 867 Genes coding for pituitary hormones (Figure 4-figure supplement 7  In the spinal cord, neuroplasticity-related genes (Figure 4-figure supplement 4) along with 873 the neuropeptide and their receptor genes (Figure 4-figure supplements 2,3) were analyzed 874 besides the Avp, Avpr1b, Avpr2, Oxt and Pomc genes that were expressed at low levels.    (Figures 4,5), and compared between UBI and sham surgery groups using Mann-922 Whitney test followed by a Bonferroni correction for multiple tests (n = 28 and 20 for the 923 hypothalamus and spinal cord, respectively). Because no significant differences between UBI 924 and sham surgery groups were revealed, the groups were combined, and the pooled data were 925 used for analysis of lateralization of gene expression. One-sample version of non-parametric 926 Wilcoxon signed-rank test was applied to compare the AI with zero followed by Bonferroni 927 multiple testing correction (28 genes for hypothalamus). Data for the spinal cord were acquired 928 and analyzed in our previous study (Lukoyanov et al., 2021).

959
The R/boot.pval package was used to compute the P-value (Thulin, 2021). P-values were 960 adjusted using Benjamini-Hochberg family-wise multiple test correction that was applied 961 separately for each set of 6 sets of the tasks (Source data: The EXCEL source data files " Table   962 III-S6 23 05 10.xlsx"). Three sets were designed to compare the coordination strength, and other  (Figures 4,5).  (Figures 4,5). 980 Task 3. Pairwise gene-gene inter-modular correlations internal for each the left and right gene 981 networks, were compared between these networks with each other (lmLdN-rmLdN vs. lmRdN-982 rmRdN correlations) in each area, to assess differences in the inter-modular coordination between 983 the networks (Figure 4-figure supplement 11; Figure 5-figure supplement 2). 984 Tasks 4 and 5. The ipsilateral hypothalamus (HPT) and spinal cord (SpC) correlations made  the C6-7 level that was followed by L-UBI (n = 8), R-UBI (n = 11), or sham surgery (Sh; n = 7).

1041
The asymmetries were analyzed three hours after UBI or sham surgery (Post) and in the same                    1573 P values shown for the contrasts that are significant (P ≤ 0.05) after the correction for all three 1574 categorization variants, or for two of them while in the third variant P was < 0.05 and < 0.10 1575 before and after the correction, respectively.

1576
Source data: The EXCEL source data file "Hypoth_SO_UBI.xlsx; Table III- 1628 P values were determined by permutation testing with Benjamini-Hochberg family-wise multiple 1629 test correction, and are shown for the median AI of the combined sham surgery and UBI group.
1630 P values shown for the contrasts that are significant (P ≤ 0.05) after the correction for all three categorization variants, or for two of them while in the third variant P was < 0.05 and < 0.10 1632 before and after the correction, respectively.