All animal studies were approved by the University of California, San Francisco, Institutional Animal Care and Use Committee. Male and female C57BL/J mice were used in equal proportions for all experiments except where indicated otherwise. Mice were housed in temperature- and humidity-controlled environments on a 12 h/12 h light/dark cycle with free access to standard chow and water. For chronic hypoxia experiments, mouse pups with their lactating mothers were subjected to chronic sublethal hypoxia (10% fraction of inspired oxygen [FiO2]) from P3 through P10. Mouse pups were treated daily from P3 through P10 with vehicle (saline) or clemastine fumarate (Selleckchem) by oral gavage. Doses of clemastine (weighed and dosed based on the clemastine fumarate salt) were: 0.5, 2, 7.5, or 10 mg per kg of body weight. On P10, mice were returned to normoxic (21% FiO2) conditions. At P14 and 10 weeks of age, OL differentiation and myelination were compared in hypoxic mice treated with vehicle or clemastine and in mice exposed to normoxia from P0 through P14.

P14 and 10-week-old mice were euthanized and perfused transcardially with ice-cold Phosphate Buffered Saline (PBS) followed by ice-cold 4% paraformaldehyde (Electron Microscopy Sciences) diluted in water. Brains and optic nerves were removed and stored overnight in 4% paraformaldehyde at 4 °C. Tissues were then placed in 30% sucrose in PBS at 4 °C for 1–2 days followed by freezing in O.C.T. compound (Tissue-Tek) and generation of 30 um sections on a microtome (HM 450 Sliding Microtome, Epredia™, Richard-Allan Scientific) or cryostat (Leica CM 1850). Sections were blocked and permeabilized for 2 h at room temperature in blocking solution (PBS with 0.1% Triton X-100, and 10% donkey or goat serum), and subsequently incubated overnight at 4 °C in blocking solution with primary antibody added. After washing, secondary antibody incubation was performed for 2 h at room temperature in 10% donkey or 10% goat serum in PBS with secondary antibody added. Primary antibodies used were: rat anti-MBP (1:500, MCA409S, Serotec), mouse anti-Olig2 (1:200, EMD Millipore, MABN50), rabbit anti-Cleaved Caspase-3 (CC3, 1:300, Cell Signaling 9661 S), rabbit anti-SOX10 (1:500, EMD Millipore AB 5727), mouse anti-CC1 (1:300 Calbiochem OP80-100uG), rabbit anti-PDGFRα (1:500, Source: W.B. Stallcup), rabbit anti-Caspr (1:500, Abcam AB34151), chicken pan anti-Neurofascin (1:500, R&D Systems, AF3235), and rabbit anti-Neurofilament heavy subunit (1:500, Abcam AB8135). Secondary antibodies were: donkey or goat AlexaFluor 488 (1:1,000), 594 (1:1000), or 647 (1:500)-conjugated IgG. Sections were counterstained with DAPI (Themofisher/Invitrogen D1306). Fluorescent images were obtained using a Zeiss Imager M2 (1024994428) microscope or a Zeiss LSM700 inverted confocal microscope. MBP and neurofilament fluorescence signal intensities were measured using Imaris (v9.3.1) in a fixed area of striatum or cortex in coronal brain sections, or in equivalent areas of longitudinally sectioned optic nerves, and normalized within each section to the area of lowest fluorescence. Cell counts were performed using Imaris (v9.3.1) software, Count Spots function, manually adjusted to remove incorrect spots (for example double counting of single cells or counting of fluorescent debris). All fluorescence intensity quantifications and cell counting were performed by an investigator blinded to the experimental condition.

P14 mice were euthanized and perfused transcardially with ice-cold 0.1 M sodium cacodylate buffer (0.1 M sodium cacodylate trihydrate [Electron Microscopy Sciences, 12310], 5 mM calcium chloride dihydrate [Sigma, 223506], pH adjusted to 7.3–7.4), followed by ice-cold EM fixative (1.25% glutaraldehyde [Electron Microscopy Sciences, 16220], 2% paraformaldehyde [Electron Microscopy Sciences, 19210], 0.1 M sodium cacodylate buffer). Brains and optic nerves were removed and stored for 8 days in EM fixative at 4 °C, and subsequently placed in 30% sucrose in PBS at 4 °C for 1–2 days. Samples were mounted in O.C.T. compound (Tissue-Tek) and 500 um sections were generated on a microtome (HM 450 Sliding Microtome, Epredia™, Richard-Allan Scientific). For analysis of the corpus callosum, the section at approximately +1.1 mm anterior to bregma (joining of corpus callosum) was placed in PBS and under a dissecting microscope, the middle 1/3 of the corpus callosum was dissected using a razor blade. Samples were stained with osmium tetroxide, dehydrated in ethanol and embedded in TAAB resin. Brain sections were cut perpendicular to the angle of the fibers of the corpus callosum and optic nerves were cut perpendicular to the nerve, at 1-mm intervals. Axons were examined using electron microscopy, and g-ratios were calculated as the diameter of the axon divided by the diameter of the axon and the surrounding myelin sheath. We could not reliably identify unmyelinated axons, particularly in vehicle-treated animals, which would have confounded comparisons of unmyelinated axon counts and diameters. Thus, all quantifications were performed on myelinated axons only. Measurements were performed using ImageJ (v1.54i) by an investigator blinded to the experimental condition.

Male and female C57BL/J mice were dosed by oral gavage with 7.5 mg/kg/day clemastine daily from P3 to P10. Blood was sampled before (time=0) the 8th dose and at 0.5, 1, 1.5, 2, 3, 4, 6, 9, 12, and 24 h after the 8th dose. Terminal blood collections were performed by cardiac puncture after deeply anesthetizing the animal. At least 3 animals were sampled per timepoint and at least one animal per sex was included for every time point. Blood was collected into 1 ml syringes using 25ga needles coated with 0.5 M EDTA (ThermoFisher, #15575020), and immediately gently transferred into K2EDTA tubes (Sarstedt #41.1395.105) followed by centrifugation at 6700 rcf for 5 min at 4 °C. Plasma was transferred into cryotubes (Nalgene™ Cryogenic Tube, #5012-0020) and stored at −80 °C until sample analysis.

Clemastine was quantified in mouse EDTA plasma using high-performance chromatography- tandem mass spectrometry (LC-MS/MS) at iC42 Clinical Research and Development (University of Colorado, Aurora, CO). The assay followed the principles described by ref. 28. Clemastine reference material was from Toronto Research Chemicals (North York, ON, Canada) and the internal standard diphenhydramine-D3 from Sigma Aldrich (St. Louis, MO). Isotope-labeled clemastine as internal standard was commercially not available at the time of the study. Two hundred (200) µL of a protein precipitation solution (0.2 M ZnSO4 30% water/ 70% methanol v/v) containing the internal standard (1.0 ng/mL diphenhydramine-D3) was added to 50 µL of study samples, quality control samples, calibrators and zero samples. Samples were vortexed for 2.5 min, centrifuged at 4 °C and 16,000 g for 10 min. The supernatants were transferred into 2 mL glass HPLC injection vials. The samples were then further extracted online and analyzed using a 2D-LC-MS/MS system composed of Agilent 1100 HPLC components (Agilent Technologies, Santa Clara, CA) and a Sciex API 5000 MS/MS detector (Sciex, Concord, ON, Canada) connected via a turbo flow electrospray source run in the positive ionization mode (4500 V, 550 °C source temperature).

Ten (10) µL of the samples were injected onto the extraction column (Zorbax XDB C8, 4.6 · 50 mm, Agilent Technologies). The mobile phase was 80% 0.1% formic acid in HPLC grade water (mobile phase A) and 20% methanol containing 0.1% formic acid (mobile phase B). Samples were cleaned with a solvent flow of 3 mL/min and the temperature for the extraction column was set to room temperature. After 0.7 min, the switching valve was activated and the analytes were eluted in the backflush mode from the extraction column onto a 4.6 · 150 mm analytical column filled with C8 material of, 5 μm particle size (Zorbax XDB C8, Agilent Technologies). The analytes were eluted using a gradient starting with 50% mobile phase B that increased to 98% within 2.3 min and was held for 1.0 min. The system was then re-equilibrated to starting conditions at 50% B for 0.8 min. The flow rate was 1.0 mL/min and the analytical column was kept at 60 °C. The MS/MS was run in the multiple reaction mode (MRM) and the following ion transitions were monitored: clemastine m/z = 346.2 [M(37Cl) + H]+ → 217.0 (quantifier), m/z = 344.2 [M(35Cl) + H]+ → 215.0 (qualifier) and diphenhydramine-D3 (internal standard) m/z = 259.0 [M + H]+ → 167.3. Declustering potentials were set to 51 V and collision energies were set to 23 V for clemastine quantifier and qualifier transitions. For the internal standard diphenhydramine-D3 the declustering potential was 56 V and the collision energy 19 V.

Clemastine concentrations were quantified using the calibration curves that were constructed by plotting nominal concentration versus analyte area to internal standard area ratios (response) using a quadratic fit and 1/x weight. All calculations were carried out using the Sciex Analyst Software (version 1.7.3). The quantification range for clemastine was 0.0025 (lower limit of quantification) – 20.0 ng/mL and study sample were diluted 1:50 and 1:250 as necessary for the detector response to fall within the calibration range as necessary. All results reported were from runs that met the following acceptance criteria: >75% of the calibrators had to be within ±15% of the nominal values (except at the lower limit of quantification: ±20%) and >2/3 of the quality controls had to be within ±15% of the nominal values. The imprecision of the results was better than 15%. Significant carry-over and matrix effects were excluded and dilution integrity was established.

Non-compartmental PK analysis was conducted using the geometric mean of each timepoint. Steady-state conditions were assumed after 7 days of dosing. The maximum concentration (Cmax), time of the maximum concentration (Tmax), and the concentration at 24 h post dose (C24h) were taken directly from the observed data. Area under the curve during the 24 h dosing interval (AUC24) was calculated using the trapezoidal method. Oral Clearance (CL/F) was then calculated as Dose/AUC24. The terminal elimination rate constant (ke) was calculated using linear regression of log transformed concentrations over the terminal log-linear decline phase and from this the terminal elimination half-life calculated (t1/2 = 0.693/ke).

Statistical significance between groups was determined with GraphPad Prism 5 software. Initial assessments of myelination (MBP intensity) and the OL lineage (CC1+ and SOX10+ cells) split by sex were performed using two-way analysis of variance (ANOVA) tests followed by post hoc Dunnett’s tests for pair-wise comparisons between the hypoxia plus vehicle group and all other treatment groups. All data sets used for two-way ANOVAs passed Shapiro-Wilk tests for normality of the residuals and Spearman’s tests for heteroscedasticity (to assess for equal variance) except for the data used for analysis of CC1+ cell density (Supplementary Table 1). The data set used for CC1+ cell density analysis was log transformed, and subsequently passed both tests for normality and equal variance; a two-way ANOVA was then performed on the log transformed data. The data sets for comparisons of immunohistochemistry and electron microscopy quantifications not split by sex were tested for normality using a Shapiro–Wilk test and for equal variance using a Brown–Forsythe test (Supplementary Table 1). Normally distributed datasets with equal variance were compared using one-way ANOVA followed by Tukey’s or Dunnett’s post hoc tests (depending on the number of comparisons made) for pair-wise comparisons. Datasets that were not normally distributed were compared using a Kruskal–Wallis test with post hoc Dunn’s tests. Datasets that were normally distributed and had unequal variance were compared using a Brown–Forsythe and Welch one-way ANOVA followed by Dunnett post hoc tests. Pairwise comparisons were performed between the hypoxia plus vehicle group and all other groups for experiments that tested all 6 treatment conditions (hypoxia plus vehicle, hypoxia plus clemastine at 0.5, 2, 7.5 or 10 mg/kg/day, and normoxia). Pairwise comparisons were performed between all groups for experiments that tested only the hypoxia plus vehicle, hypoxia plus clemastine MED, and normoxia conditions. We made the assumption of independence. When possible, we assigned different treatments to mice within the same litter (after an identifying toe clip at P3) to control for batch/litter effects. A probability of p < 0.05, adjusted for multiple comparisons, was considered significant for all statistical comparisons.