(684c) Regionally-Dependent Cellular and Mechanistic Signatures for Screening Neuroprotective Therapies in the Preterm Injured Brain

Preterm birth, accompanied by intrapartum-related complications, is one of the leading causes of neonatal deaths. Perinatal hypoxia, hyperoxia, and hypoxia-ischemia (HI) are common during preterm birth, contributing to inflammation and oxidative injury. Improving the outcomes of extremely preterm (EP) infants born before 28 weeks’ gestation is an important target, with at least 10% dying and >60% of the survivors currently developing at least one disability such as cerebral palsy, autism, ADHD, or cognitive, hearing, or visual impairment.¹ However, there are currently no targeted neuroprotective interventions for preterm infants. As preterm survival rate continues to improve, it is critical to develop targeted antioxidant, anti-inflammatory, or immunomodulatory therapies to optimize outcomes for this population.

In this study, we utilize organotypic whole-hemisphere (OWH) brain slices from the postnatal day (P)14 extremely preterm-equivalent ferret to evaluate neuroprotective efficacy of two therapeutics of interest: erythropoietin (Epo) and azithromycin (Az). Epo and Az have potential neuroprotective effects and are used in preterm infants for other reasons – anemia of prematurity and treatment of ureaplasma infections, respectively.² The underlying biochemical and cellular responses against preterm brain injury are yet to be explored for Az and Epo. We have previously shown that ferrets are a highly relevant animal model to investigate treatments for preterm brain injury since they are born lissencephalic and develop a gyrencephalic cerebral cortex postnatally.^3–5 The ferret brain also undergoes postnatal white matter maturation and complex cortical folding in a similar pattern as the human brain during the third trimester. OWH slice models bridge the gap between in vitro cell lines and in vivo models by maintaining the brain’s structural integrity and capturing regional variability in susceptibility to injury and therapeutic responses. We have previously shown in a term-equivalent ferret OWH slice model of oxygen glucose deprivation (OGD) that neuronal injury has regional dependence in response to injury and treatment, while mirroring trends seen after injury in vivo.^3,4,6 Here, we expose extremely preterm-equivalent ferret OWH slices to OGD to investigate injury responses through global and regional cellular cytotoxicity, microglial morphological parameter and shape mode distribution changes, and evaluate potential neurotherapeutic effects from Az and Epo with regional transcriptomics.

Methods

300μm whole-hemisphere ferret brain slices were obtained from P14 kits, comparable to extremely preterm (<28 weeks’) human gestation and cultured for 72 h. After 72h in vitro, slices were subjected to 2h of OGD followed by treatment with Az (15mM), Epo (3 IU/mL), or Az+Epo combination for 24h. Global and regional cell death were determined using lactate dehydrogenase (LDH) release and confocal microscopy of DAPI-stained pyknotic nuclei counts, respectively. Immunofluorescence staining with Iba-1 highlighted microglia in different regions. Microglial morphology was assessed using the Visual Aided Morpho-Phenotyping Image Recognition (VAMPIRE) machine learning tool by capturing morphological parameters and reducing each cell to determine common morphological groupings using principal component (PC) analysis.⁴ Global (whole slice) and regional (microdissected slice) transcriptomics were determined using a ferret-specific NanoString nCounter panel of 255 genes. Transcripts upregulated by OGD and normalized by Az+Epo were determined through fold changes and statistical significance.

Results and Discussion

Global cell death according to LDH release was significantly increased after OGD. Treatment with the Az+Epo combination, but not either therapy individually, significantly decreased LDH release to a level comparable to that of the normal control (NC) group. Using counts of pyknotic nuclei to indicate cell death, OGD decreased cell viability significantly in the cortex (p < 0.0001) and basal ganglia (p = 0.0004). In the basal ganglia, Az (p = 0.011) and Epo (p = 0.0076) individually restored cell viability from injury, whereas the Az+Epo combination (p < 0.0001) reduced regional cell death to a greater extent (Figure 1A). Az (p = 0.001) and Epo (p = 0.004) as standalone treatments were also protective in the hippocampus. Az+Epo combination therapy was significantly protective in the basal ganglia (Figure 1A).

In vitro mechanisms of microglia in injury repair have been explored in rodent models but have not yet been thoroughly investigated with in the OWH ferret model. We interpreted regional phenotypic changes and shape mode shifts to predict microglial activation and neuroprotective mechanisms. The OGD group showed a significant decrease in Iba-1+ cell numbers below those in the NC group in all regions except the subcortical white matter (ScWM). Treatment with Az significantly increased microglial counts in the subcortical white matter (p = 0.0073), the basal ganglia (p = 0.0007), and thalamus (p = 0.0345). Epo treatment significantly increased Iba-1+ cell counts in the thalamus (p = 0.037) (Figure 1B). In addition to changes in cell number, morphological shifts in microglia can also provide insights into post-OGD treatment responses. Globally, OGD significantly decreased microglial cell area and circularity, with slight increase in perimeter. These trends can indicate a shift to more ramified morphology of microglia. While the global trend in perimeter and circularity appears the same in all treatment groups compared to OGD slices, the Az+Epo combination treatment more significantly increased cell area coverage (p < 0.0001). This can signify the presence of rounder and expanded microglial cell bodies with retracted processes within the combination treatment group. Regional variability was also apparent in microglial shape mode (SM) distribution. In the cortex, SM1 was more likely found across all groups, except for in the Az+Epo group. When compared to OGD, the likelihood of any individual microglia being in SM5 was significantly increased by Az+Epo in the cortex (Figure 1C). The dominance of SM5 decreased presence of the more extended and branched shapes (SMs 2 and 3) correlated with increased area coverage and circularity of microglia.

We also quantified biomarkers that are indicative of the presence and severity of preterm brain injury, though expression outcomes have varied clinically and are an ongoing research area for potential translation. From our analysis, after OGD, n=92 (globally), n=96 (white matter), n=77 in (deep gray matter) and n=64 (cortex) genes were differentially expressed. Out of these, n=7 globally, n=22 in white matter, n=19 in deep gray matter and n=4 genes in the cortex were differentially normalized by Az+Epo but not Az or Epo individually. The top transcripts normalized by Az+Epo were Caspase 8, MERTK, TFEC, and PARP1 globally, anti-inflammatory cytokine IL10 and CD86 in the deep gray matter, and the macrophage scavenger receptor MARCO, TNF, and GSN in white matter (Figure 1D). We highlighted region-specific transcriptomics level changes to Az and Epo, especially in the white matter and deep gray matter where injury is commonly seen in preterm infants.

Conclusions

In the extremely preterm-equivalent ferret brain, we demonstrated region-specific responses to Az, Epo, and their combination with cellular level changes. Az+Epo ameliorated cytotoxicity globally but didn’t provide neuroprotection in all brain regions. Microglia morphological changes were also highly heterogenous across regions. The complexity of interpreting regional phenotypic changes and shape mode shifts calls for complementary methods to predict microglial activation and neuroprotective mechanisms. Spatial transcriptomics showed Az+Epo normalizing cell death pathways (Caspase 8, MERTK) and inflammatory responses (IL10, MARCO, TNF) that are not seen with either drug individually. This platform warrants further investigation with diverse cellular and mechanistic signatures in neurotherapeutic targets for regional and global neuroprotection for the extremely preterm brain.

Reference

1. Juul SE, Wood TR, Comstock BA, Perez K, Gogcu S, Puia-Dumitrescu M, Berkelhamer S, Heagerty PJ, PENUT Consortium. Deaths in a Modern Cohort of Extremely Preterm Infants From the Preterm Erythropoietin Neuroprotection Trial. JAMA Netw Open. 2022 Feb 1;5(2):e2146404. PMCID: PMC8822378
2. Juul SE, Pet GC. Erythropoietin and Neonatal Neuroprotection. Clin Perinatol. 2015 Sep;42(3):469–481. PMCID: PMC4529536
3. Wood T, Moralejo D, Corry K, Snyder JM, Traudt C, Curtis C, Nance E, Parikh P, Juul SE. A Ferret Model of Encephalopathy of Prematurity. Dev Neurosci. 2018;40(5–6):475–489. PMCID: PMC6658350
4. Wood TR, Hildahl K, Helmbrecht H, Corry KA, Moralejo DH, Kolnik SE, Prater KE, Juul SE, Nance E. A ferret brain slice model of oxygen–glucose deprivation captures regional responses to perinatal injury and treatment associated with specific microglial phenotypes. Bioengineering & Transla Med [Internet]. 2022 May [cited 2022 Aug 31];7(2). Available from: https://onlinelibrary.wiley.com/doi/10.1002/btm2.10265
5. Wood T, Moralejo D, Corry K, Fisher C, Snyder JM, Acuna V, Holden-Hunt A, Virk S, White O, Law J, Parikh P, Juul SE. A Ferret Model of Inflammation-sensitized Late Preterm Hypoxic-ischemic Brain Injury. J Vis Exp. 2019 Nov 19;(153). PMID: 31814608
6. Kolnik S, Corry K, Hildahl K, Filteau J, White O, Brandon O, Farid L, Shearlock A, Moralejo D, Juul SE, Nance EA, Wood TR. Vitamin E Decreases Cytotoxicity and Mitigates Inflammatory and Oxidative Stress Responses in a Ferret Organotypic Brain Slice Model of Neonatal Hypoxia-Ischemia. Dev Neurosci. 2022;44(4–5):233–245. PMID: 35134797