Critical reagents and instruments used in this study are listed in Table A1. The bicinchoninic acid (BCA) kit, MaxiSorp 96-well plates, horseradish peroxidase (HRP)-conjugated secondary antibodies, and Superblock (TBS) were purchased from Thermo-Fisher Scientific (Bedford, MA, USA). The soluble fluorophores, Amplex UltraRed and 4-Methylumbelliferyl phosphate (4-MUP) were from Life Technologies (Carlsbad, CA, USA). Vector Laboratories Inc. (Newark, CA, USA) was the source of the Proton Biotin Protein Labeling Kit and Alkaline Phosphatase-conjugated Streptavidin. The rat Akt Pathway Total and Phospho Multiplex panels were from MilliporeSigma (Burlington, MA, USA) (Table A2). Primary antibodies used for direct binding duplex enzyme-linked immunosorbent assays (ELISAs) are listed in Table A3. Fine chemicals were purchased from CalBiochem (Carlsbad, CA, USA) or Sigma-Aldrich (St Louis, MO, USA).

The FASD model was generated in Long Evans rats in accordance with our protocol, which was approved by the Brown University Health Institutional Animal Care and Use Committee (IACUC) of Rhode Island Hospital [16]. Females purchased from Charles River Laboratories (Willmington, MA, USA) were mated, and 6 days after detecting sperm in the vaginal canal, isocaloric control (0% ethanol) or ethanol-containing (26% caloric pharmaceutical grade) liquid diets (BioServ, Frenchtown, NJ, USA) were initiated. To study the dose effects of dietary soy, both the control and ethanol diets were supplied with protein consisting of 0% soy/100% casein, 10% soy/90% casein, 30% soy/70% casein, 50% soy/50% casein, or 100% soy/0% casein. The liquid diets were initiated on gestation day 6 to minimize fetal demise and failed implantation, which frequently accompany earlier gestation ethanol feeding [37]. The liquid diets continued until postnatal day 1 (P1), after which the dams were maintained on standard chow without further ethanol exposure. Upon weaning, the pups were fed chow and provided free access to water. Apart from fetal demise marked by fetal resorption [9], prenatal alcohol exposure does not reduce longevity into adulthood.

On P16, cerebellar motor function was assessed by subjecting the offspring to 10 consecutive trials of incrementally increased rod rotation speed from 1.5 to 6.0 rpm using a Rotamex 5 instrument (Columbus Instruments, Columbus, OH, USA) [15]. The latencies to fall were automatically recorded with photocells positioned over the rod. To prevent exercise fatigue, all trials were halted after 45 seconds of work, and the rats were allowed 10 minutes of rest between each trial. The data were analyzed using two-way ANOVA with the Tukey post hoc test.

P35 was the experimental endpoint. The rats were sacrificed with terminal isoflurane anesthesia. Fresh cerebella were hemisected through the vermis. One hemisphere was fixed in 10% neutral buffered formalin, paraffin-embedded, and sectioned for staining with hematoxylin and eosin. Image analysis with Stereologer software (Stereology Resources, Inc., Chester, MD, USA, available at https://srcbiosciences.com/stereologer-software, accessed on 24 February 2025) was used for unbiased stereological quantification of granule and Purkinje cell densities, as described [17]. Gray matter volume was measured with the volume probe, and granule and Purkinje cell abundances were assessed using dissector probes. Cortical volumes were measured using the area point-count method at a 4x magnification. Purkinje and granule cells were counted, respectively, using frame sizes of 40% and 0.5% of the screen height, whereas granule cells were counted using a frame size that was 0.5% of the screen height at 40x magnification.

The non-fixed cerebellar hemisphere was snap-frozen on dry ice and stored at −80˚C for biochemical studies. For protein assays, 100 mg tissue samples were homogenized in 5 volumes of buffer (150 mM NaCl, 50 mM Tris-Base pH 7.5, 0.1% Triton X-100, 5 mM EDTA pH 8.0, 10 mM EDTA, 50 mM NaF) supplemented with protease (1 mM PMSF, 0.1 mM TPCK, 2 µg/ml aprotinin, 2 µg/ml pepstatin A, 1 µg/ml leupeptin, 1 mM NaF, 1 mM Na₄P₂O₇) and phosphatase (2 mM Na₃VO₄) inhibitors [16]. The samples were homoginzed using a TissueLyser II instrument (Qiagen, Germantown, MD, USA) and 5-mm diameter stainless steel beads. Supernatant fractions obtained by centrifuging the samples at 14,000xg for 10 minutes at 4˚C were used in immunoassays. Protein concentrations were measured with the BCA assay.

Magnetic multiplex bead-based platform kits (Invitrogen, Carlsbad, CA, USA) were used to measure total and phosphorylated protein levels in the insulin/IGF/IRS-Akt pathway (Table A2). In brief, 100 μg aliquots of cerebellar protein homogenate were incubated with antibody-coated beads. The captured antigens were detected with biotinylated secondary antibodies and phycoerythrin-conjugated Streptavidin. Immunoreactivity was measured in a Luminex MAGPIX instrument (Diasorin, Austin, TX, USA) with xPONENT software. MAGPIX calibration and verification standards were used throughout, and standard curves were generated for each analyte. Data are expressed in arbitrary fluorescence light units (FLU) calculated from the standard curves.

Duplex ELISAs: Direct binding duplex ELISAs measured immunoreactivity to neuronal and glial proteins [38]. The antibodies, their sources, characteristics, validations, and final concentrations are listed in Table A3. Triplicate samples containing 50 ng protein in 50 μL bicarbonate-binding buffer were first added to 96-well MaxiSorp plates for overnight adsorption at 4˚C. Non-specific binding sites were masked with Superblock TBS. The samples were then incubated with primary antibodies (0.2 - 5.0 µg/mL) overnight at 4˚C. Immunoreactivity was detected with horseradish peroxidase (HRP)-conjugated secondary antibodies and the Amplex UltraRed soluble fluorophore. Fluorescence intensity was measured (Ex 530 nm/Em 590 nm) in a Spectra-Max M5 Multimode Plate Reader (Molecular Devices, Sunnyvale, CA, USA). The results were normalized to large acidic ribosomal protein (RPLPO) (Proteintech Group Inc., Chicago, IL) as a control for sample loading [39][40]. The second phase of the duplex ELISA was accomplished by incubating the samples with biotin-conjugated anti-RPLPO, followed by streptavidin-conjugated alkaline phosphatase. RPLPO immunoreactivity was detected with 4-Methylumbelliferyl phosphate (4-MUP), and fluorescence was measured in a SpectraMax M5 (Ex 360 nm/Em 450 nm).

Statistical analysis was performed, and graphs were generated using GraphPad Prism 10.6 (GraphPad Software Inc., Boston, MA). Inter-group comparisons were made using one-way or two-way analysis of variance (ANOVA) with Tukey post hoc multiple comparisons. Software-generated statistically significant (P ≤ 0.05) or trendwise (0.05 < P < 0.10) differences are included in the Tables and Graphs.

The 175 offspring, comprised of 90 males and 85 females, were divided among 5 control and 5 ethanol diet groups. The diets contained 0%, 10%, 30%, 50% or 100% soy as the protein source, with casein added to achieve 100% of the protein requirements. Within each group, there were no significant differences in the proportions of males and females that were gestationally exposed to the control or ethanol diet, as assessed by chi-square analysis (Table 1).

Table 1.Rat treatment groups.

Long-Evans pregnant rat dams were maintained on isocaloric control or ethanol-containing (26% caloric) liquid diets from gestation day 6 through delivery. The sources of dietary protein included from 0% to 100% soy with casein added to achieve 100% of the required protein composition. The number and percentage of male and female pups in each group are listed. Chi-square tests show no significant differences in sex distribution across groups.

Rotarod performance was assessed by comparing mean latency-to-fall across 10 trials at progressively higher rod rotation speeds. The two-way ANOVA tests demonstrated significant dietary soy effects in all 10 trials and significant ethanol effects in all trials except Trial #1 (Table 2). Significant interactive effects of significant dietary protein source (soy dose) x ethanol were detected in Trials #4-#6 and #9-#10. In addition, statistical trend effects of dietary protein x ethanol interactions on performance were observed for Trials #3 and #8.

Table 2.Two-way ANOVA for trial-based rotarod performance-dietary protein source and ethanol factors.

Long Evans rat offspring of pregnant dams that were maintained on isocaloric control or ethanol-containing (26% caloric) liquid diets from gestation day 6 through delivery. Maternal sources of dietary protein included from 0% to 100% soy, with casein added to achieve 100% of the required protein composition. On postnatal day 15, the offspring were subjected to 10 consecutive progressively challenging rotarod performance tests to evaluate cerebellar motor function. Results were analyzed by two-way ANOVA with post hoc Tukey tests. Significant (P ≤ 0.05) differences are highlighted with bold font. Statistical trend effects (0.05 < P < 0.10) are italicized. The graphs and significant post hoc Tukey test results are shown in Figure 1 and Figure 2.

To clearly illustrate the within-group and between-group exposure effects, graphs were generated to show the clustered and overall performance of the control and ethanol groups (Figure 1) or to present direct comparisons for each trial (Figure 2). The most striking observation was that among controls, dietary soy dose impacted performance such that exposures between 30% and 100% soy significantly increased the latencies to fall relative to the effects of 0% or 10% soy (Figure 1). In contrast, the effects of dietary soy dose were modest and limited to the least challenging trials, i.e., #1-#3. The control-versus-ethanol paired comparisons by soy dose demonstrated that inter-group differences were mainly observed at 50% and 100% dietary soy, and largely attributable to the positive effects in controls versus minimal effects in the ethanol group (Figures 2(B)-(J)). Intergroup differences were not detected for Trial #1 (Figure 2(A)) or for most of the more challenging trials in rats exposed to 30% or lower levels of dietary soy.

Figure 1. Rotarod Performance-Soy dose effects on performance in control and ethanol-exposed offspring. (A)-(J) Rotarod performance was measured over 10 consecutive trials of increasing rod rotation speed (See Table 2). The latencies to fall off the rod (seconds) were optically recorded. Graphs depict the mean ± S.D. of latencies in rats exposed to 0% (control) or 26% caloric ethanol in liquid diets containing 0%, 10%, 30%, 50%, or 100% soy as the protein source, with casein added to meet 100% of the dietary protein requirement. Results were analyzed by trial using ANOVA and Tukey’s test for within-group comparisons. Significant differences are displayed.

Figure 2. Rotarod Performance-Ethanol exposure effects on performance with increasing dietary soy dose. (A)-(J) Rotarod performance was measured over 10 consecutive trials of increasing rod rotation speed (See Table 2). The latencies to fall off the rod (seconds) were optically recorded. Graphs depict the mean ± S.D. of latencies in rats exposed to 0% (control) or 26% caloric ethanol in liquid diets containing 0%, 10%, 30%, 50%, or 100% soy as the protein source, with casein added to meet 100% of the dietary protein requirement. Results were analyzed by trial using ANOVA and Tukey’s test for between-group comparisons at each soy dose. Significant differences are displayed.

The experimental endpoint was P35. Two-way ANOVA detected significant soy dose, ethanol, and soy dose x ethanol interactions with respect to body weight (Table 3). However, regarding brain weight, significant soy dose and soy dose x ethanol interactive effects were observed, whereas ethanol itself was not a significant factor (Table 3).

Table 3.Two-way ANOVA table for body and brain weights.

Long Evans rat offspring of pregnant dams that were maintained on isocaloric control or ethanol-containing (26% caloric) liquid diets from gestation day 6 through delivery. Soy comprised 0%, 10%, 30%, 50%, or 100% of the protein source, and with casein added to provide 100% of the required protein. Postnatal day 35 (P35), experimental endpoint body and brain weight comparisons were made using two-way ANOVA. Significant results are represented with bold font. The corresponding graphs with significant (P < 0.05) post hoc Tukey test results are shown in Figure 3.

Graphs were generated to optimally illustrate the within- and between-group effects of exposure on body weight (Figure 3(A), Figure 3(B)), brain weight (Figure 3(C), Figure 3(D)), and the brain weight/body weight ratios (Figure 3(E),Figure 3(F)). Body weight distributions varied within each group, but the main effect was that rats exposed to 100% soy in utero were significantly heavier than the other groups (Figure 3(A)). The paired comparisons demonstrated that within the 10% and 30% soy groups, ethanol-exposed rats were significantly heavier than controls, whereas the opposite was observed in the 50% dietary soy group (Figure 3(B)). The effects of soy dose and ethanol were clearer for brain weight. Among controls, the lowest mean brain weight was observed in the 10% soy group, and rats exposed to 100% soy protein had significantly higher brain weights than rats exposed to 10% or 30% soy (Figure 3(C)). In the ethanol group, the lowest mean brain weight was in the 0% soy group. Exposure to 10% soy or higher significantly increased brain weight. The largest effect occurred with 100% soy, such that the mean brain weight was significantly higher than in all the other ethanol groups (Figure 3(C)). The paired control versus ethanol paired comparisons demonstrated significantly lower mean brain weight in ethanol+ 0% soy and ethanol + 50% soy groups (Figure 3(D)).

Figure 3. Offspring Metrics. Effects of dietary soy dose and ethanol on body and brain weights. Pregnant Long Evans dams were maintained on isocaloric liquid diets containing 0% (control) or 26% caloric ethanol and 0%, 10%, 30%, 50%, or 100% soy as the protein source, with casein added to meet 100% of the dietary protein requirement. Ethanol and soy were withdrawn on postnatal day 1 (P1). P35 was the experimental endpoint. Scatter plots depict within-group effects of dietary soy dose on (A) body weight, (C) brain weight, and (E) brain/body weight ratios, and between-group differences in the effects of soy dose on (B) body weight, (D) brain weight, and (F) brain/body weight ratios in P35 offspring. Data were analyzed by two-way ANOVA (Table 3) and the Tukey post hoc test.

Otherwise, the control and ethanol groups exposed to 10%, 30%, or 100% dietary soy had comparable mean brain weights. The data for the brain weight/body weight ratios show progressively increasing values from 0% to 30% soy in the control and ethanol groups, but sharp reductions in the 100% soy groups, reflecting the effects of increased body weight (Figure 3(E)). Inter-group comparisons demonstrated lower mean brain weight-to-body weight ratios associated with ethanol and with dietary protein ranging from 0% to 30% soy. The combined effects of soy-mediated increases in brain and body weight rendered the brain-to-body weight ratios similar when the diets contained 50% or 100% soy as the protein source (Figure 3(F)).

Image analysis with unbiased stereology was used to quantify the densities of Purkinje and granule cell neurons in the cerebellar cortex. The control within-group one-way ANOVA tests demonstrated significant dietary soy dose effects on Purkinje cell density and a statistical trend effect on granule cell density (Table 4). The ethanol within-group one-way ANOVA test demonstrated a significant dietary soy dose effect on granule cell but not Purkinje cell density (Table 4).

Table 4. Impact of dietary soy dose on cerebellar structure within control and ethanol groups.

Image analysis of the cerebellar cortex with unbiased stereology assessed the effects of prenatal ethanol and dietary soy dose on Purkinje and granule cell densities. Cells were counted at high magnification using an optical dissector method and the results were normalized to tissue volume based on Cavalieri point grid method. The data were analyzed by one-way ANOVA to determine the effect of soy dose on neuronal cell density within the control and ethanol groups. P ≤ 0.05 was considered as statistically significant. Statistical trend effects (0.05 < P < 0.10) are italicized. NS = not significant. See the corresponding graphs in Figure 4.

Two-way ANOVA revealed significant ethanol effects on Purkinje cell density, and significant soy dose, ethanol, and soy dose x ethanol interactions on granule cell density (Table 5). The corresponding graphs of paired comparisons between the control and ethanol groups by soy dose, with post hoc Tukey test results, showed significantly lower Purkinje and granule cell neuron densities in the ethanol + 0% dietary soy samples (Figure 4(A) and Figure 4(B)). In both groups, the

Table 5.Two-Way ANOVA tests of soy dose and ethanol exposure effects on cerebellar structure.

Histological sections of cerebellar cortex were evaluated using stereology-based image analysis to assess the effects of prenatal ethanol and dietary soy dose on Purkinje and granule cell densities. Cells were counted at high magnification using an optical dissector method and the results were normalized to tissue volume based on Cavalieri point grid method. The results were analyzed by two-way ANOVA with post hoc Tukey tests. P ≤ 0.05 was considered as statistically significant. See the corresponding graphs in Figure 4.

Figure 4.Cerebellar Morphometrics. Histological sections of cerebellar cortex (vermis) from control and ethanol-exposed Long Evans rat P35 offspring were subjected to image analysis to evaluate the effects of prenatal alcohol exposure and dietary soy dose (0%, 10%, 30%, 50%, or 100% of the protein source) on Purkinje and granule cell densities. Numerical densities of (A) Purkinje and (B) granule cells were quantified per cubic micron of cortical gray matter present on the slide with the aid of an Olympus BX60 light microscope (Olympus America Inc., Center Valley, PA) with an attached MS-2000 XYZ Inverted Stage (Applied Scientific Instrumentation, Eugene, OR) and Stereologer software (Stereology Resource Center, Inc., Chester, MD). Unbiased counting frames were applied under software control. Cells were counted at high magnification (40x) using an optical dissector method and normalized to volume using the Cavalieri point grid method. The results were analyzed by two-way ANOVA (Table 4 and Table 5) and Tukey post hoc tests. Significant (P ≤ 0.05) inter-group differences for each soy dose are displayed in the panels. (C-F) Representative histological images of cerebellar cortex from the (C) Control + 0% soy, (D) Ethanol + 0% soy, (E) Control + 100% soy, and (F) Ethanol + 100% soy groups. H&E-stained slides were photographed at 200x. Insets show images of Granule cells (GC) photographed at 600x. PC = Purkinje cells. Note smaller area (volume) of the GC zone and loss of PC in Panel D versus Panel C. Ethanol-reduced GC density is highlighted by the relatively sparse populations of small dark nuclei shown in the inset in D versus C. Dietary soy had minimal effect on the PC and GC populations in control cerebella (compare E to C), but it conspicuously increased the abundance of PC and GC in the ethanol +100% soy (F) compared with ethanol + 0% soy (D). 100-µm scale bars are shown in each panel for size/magnification reference.

modest dietary soy-related increases in Purkinje cell densities abolished their inter-group statistical differences. A similar phenomenon occurred with respect to the granule cells, except at the 100% soy dose, which was associated with a significantly lower density in the ethanol group (Figure 4(B)).

Representative photographs of H&E-stained cerebellar cortex sections demonstrate the effects of ethanol and dietary soy on the Purkinje and Granule cell populations (Figures 4(C)-(F)). Control cerebella from the offspring of dams fed with liquid diets that contained no alcohol and had 0% Soy/100% casein as the protein source, exhibited abundant populations of Purkinje and granule cells (Figure 4(C)). In contrast, gestational ethanol exposure coupled with diets containing 0% soy/100% casein resulted in reduced cerebellar volumes with loss of both Purkinje and granule cells (Figure 4(D)). Large gaps in the Purkinje cell layer reflect neuronal loss. The reduced density of dark round nuclei in the granule cell layer reflects combined effects of cell loss and impaired proliferation. Control cerebella from offspring of dams fed with diets that contained 100% soy/0% casein as the protein source (Figure 4(E)) were similar to controls in the 0% soy/100% casein group. In contrast, in ethanol-exposed offspring, 100% dietary soy increased the population of Purkinje cells, resulting in fewer gaps, and increased the density of cells in the granule cell layer (Figure 4(F)). The histological features of cerebella from the ethanol-exposed, 100% Soy/0% Casein group were similar to those of the corresponding control group (Figure 4(E)), corresponding with the results obtained by image analysis.

Previous studies showed that chronic ethanol exposures impair insulin and IGF-1 signaling through IRS and Akt pathways in the brain, and correlated those effects with alcohol-related neurobehavioral dysfunctions [12][14][15][41]-[43]. To examine the extent to which dietary soy’s effects on cerebellar structure were mediated by restoration of these signaling networks, the samples were analyzed with multiplex total and phosphoprotein ELISAs (Table A2). In addition, the calculated relative levels of protein phosphorylation (p/T) were assessed. Two-way ANOVA detected significant soy dose effects on all signaling, phosphorylated signaling, and calculated relative levels of protein phosphorylation (Table 6). In addition, significant ethanol effects were detected for IRS1 and ^{pYpY1135/1136}-IGF-1R, and significant soy dose x ethanol interactive effects were observed for ^{pYpY1135/1136}-IGF-1R, ^pS636-IRS-1, ^pS473-Akt, and the relative levels of IGF-1R and Akt phosphorylation.

Table 6. Two-way ANOVA table for insulin, IGF, Akt pathway molecules.

Immunoreactivity was measured in 100 µg cerebellar protein using commercial Total Akt and Phospho-Akt Magnetic Bead-Based ELISA panels and a MAGPIX instrument. The calculated levels of relative protein phosphorylation (p/T) were also compared. Inter-group comparisons were made using mixed-model ANOVA and post-hoc Tukey tests with correction for multiple comparisons. F-Ratios correspond to measured soy dose, ethanol, and soy dose x ethanol interactive effects. Significant differences (bold font) have P-values < 0.05. Statistical trend effects (0.05 < P < 0.10) are italicized. NS = not significant. R = receptor; IGF-1 = insulin like growth factor type 1; IRS-1 = insulin receptor substrate, type 1; GSK-3β = glycogen synthase kinase 3β; pY = phosphotyrosine; pS = phosphoserine. See Table A2 for details of protein phosphorylation sites.

Box and whisker plots display the effects of ethanol and dietary soy on insulin/IGF-1/IRS-1-Akt pathway proteins, phosphoproteins, and the relative levels of protein phosphorylation (Figure 5). The main overall effects observed were that, with few exceptions, the levels of immunoreactivity at each soy dose were similar for the control and ethanol samples, and the most prominent shifts in immunoreactivity were associated with the 50% and 100% soy protein diets. To simplify the graphic presentation, the significant Tukey comparisons displayed were focused on effects relative to the 0% soy diet groups. Insulin R (Figure 5(A)) and IGF-1R (Figure 5(B)) were similarly expressed at relatively low levels in cerebellar tissue from control and ethanol-exposed rats in the 0% to 50% soy groups, and both were significantly elevated with the 100% soy diet (compared with all other groups. In contrast, IRS1 (Figure 5(C)), Akt (Figure 5(D)), and GSK-3β (Figure 5(E)) immunoreactivities exhibited downward trends from 0% to 50% soy, followed by a further significant decline in IRS-1, and moderate increases in Akt and GSK-3β in the 100% soy diet groups.

The levels of ^{pYpY1162/1163}-Insulin R (Figure 5(F)), ^{pYpY1135/1136}-IGF-1R (Figure 5(G)), ^pS636-IRS-1 (Figure 5(H)), and ^pS473-Akt (Figure 5(I)) trended upward from the 0% to 30% or 50% soy diets, then declined to levels below those in the 0% soy groups. In addition, for each of these phosphoproteins, significant ethanol-inhibitory effects were observed in the 50% soy groups. The ^pS9-GSK3β levels were relatively uniform over the full soy dose range except for a significant reduction in

Figure 5.Insulin/IGF Signaling Network. Prenatal ethanol and dietary soy dose effects on (A)-(E) protein, (F-J) phosphoprotein, and (K)-(O) the relative levels of signaling protein phosphorylation in cerebellar tissue homogenates from the offspring of dams fed with isocaloric liquid diets containing 0% or 26% caloric ethanol and 0%, 10%, 30%, 50%, or 100% soy as the protein source, with casein added to meet 100% of the dietary protein requirement. Commercial multiplex bead-based immunoassay kits (Table A2) were used to measure immunoreactivity to signaling proteins (A) insulin receptor (Insulin-R), (B) IGF-1R, (C) IRS-1, (D) Akt, (E) GSK-3β, phorphorylated (pY or pS) signaling molecules (F) pYpY1162/1163-Insulin-R, (G) pYpY1135/1136-IGF-1R, (H) pS312-IRS-1, (I) pS473-Akt, and (J) pS9-GSK-3β, and the calculated relative levels (pY/T or pS/T) of signaling protein phosphorylation (K) pY/T-Insulin R, (L) pY/T-IGF-1R, (M) pS/T-IRS-1, (N) pS/T-Akt, and (O) pS/T-GSK-3β. The data were analyzed by two-way ANOVA (Table 6) and the Tukey post hoc test. Significant (P ≤ 0.05) results are displayed within the panels.

the 30% soy control group and a significant increase in the 100% soy ethanol group relative to corresponding 0% soy samples (Figure 5(J)). The profiles observed for soy dose-related shifts in the relative phosphorylation levels of Insulin R (Figure 5(K)), IGF-1R (Figure 5(L)), Akt (Figure 5(N)), and GSK-3β (Figure 5(O)) were nearly identical to those observed for the corresponding tyrosine- or serine-phosphorylation. In contrast, the calculated relative levels of pS-IRS1 phosphorylation were uniformly low in the 0% to 50% soy groups but significantly elevated in the 100% soy compared with all other groups (Figure 5(M)).

Neuronal and glial protein expression was measured by duplex ELISA with results normalized to acidic ribonuclear protein (RPLPO). Two-way ANOVA detected significant soy dose effects on Tau, pTau, ubiquitin, and choline acetyltransferase (ChAT) (Table 7) and significant ethanol effects on Tau, ChAT, and glial fibrillary acidic protein (GFAP). Significant soy dose x ethanol interactive effects were observed for Tau, ubiquitin, ChAT, and GFAP. Box and whisker plots were configured to clearly display soy dose effects in control and ethanol-exposed cerebella (Figures 6(A)-(E)) or the comparative effects of ethanol relative to control at each dietary soy dose (Figure 6(F)-(J)), with the significant post hoc Tukey results.

Table 7. Two-way ANOVA table for neuroglial proteins.

Immunoreactivity was measured by direct-binding ELISA in 50 ng samples of cerebellar protein homogenates. Immunoreactivity was measured with primary antibodies targeting Tau, phospho-Tau (pTau), ubiquitin, choline acetyltransferase (ChAT), or glial fibrillary acidic protein (GFAP), horseradish peroxidase-conjugated secondary antibody, and Amplex UltraRed soluble fluorophore. Fluorescence intensity (Ex 560 nm/Em 590 nm) was measured in a SpectraMax M5. Results were normalized to the internal control ribosomal protein, RPLPO. The data was analyzed using mixed-model ANOVA and post-hoc Tukey tests (See Figure 6). P ≤ 0.05 was considered statistically significant (bold font). NS = not significant.

Tau immunoreactivity was significantly elevated in the 100% soy relative to most lower soy doses, while in the ethanol group, Tau was increased in the 10% soy diet relative to most other doses (Figure 6(A)). pTau (Figure 6(B)) and ubiquitin (Figure 6(C)) immunoreactivities were lowest in the 0% soy control and ethanol groups and significantly increased across the full range of dietary soy exposure. However, the peak levels were not correlated with the maximum soy dose. ChAT immunoreactivity was significantly reduced from 0% to 10% soy in control samples, but significantly reduced by 10%, 30% or 50% dietary soy in the ethanol group (Figure 6(D)). The levels of GFAP immunoreactivity were relatively consistent across the dietary soy dose range in both control and ethanol samples, although significant differences were detected due to an isolated increase in the 30% soy control group, and inhibitory effects of 10% and 30% soy in the ethanol group (Figure 6(E)).

Figure 6. Dietary Soy Dose and Ethanol Effects on Cerebellar Neuroglial Protein Expression. Cerebellar tissue from the offspring of dams fed with isocaloric liquid diets containing 0% or 26% caloric ethanol and 0%, 10%, 30%, 50%, or 100% soy as the protein source, with casein added to meet 100% of the dietary protein requirement was used to measure immunoreactivity to (A, F) Tau (BDNF), (B, G) pS396-Tau (pTau), (C, H) ubiquitin, (D, I) choline acetyltransferase (ChAT), and (E, J) glial fibrillary acidic protein (GFAP) by duplex ELISA. Results were normalized to acidic ribosomal phosphoprotein PO (RPLPO) housekeeping molecule. Graphs (A)-(E) show within-group (control or ethanol) effects of soy dose on immunoreactivity levels, and (F)-(J) show the between-group differences in responses to increasing soy dose. Data were analyzed by two-way ANOVA (Table 7) and the Tukey post hoc test. Significant differences (P < 0.05) are shown within the panels.

The graphs designed to compare the effects of ethanol revealed significantly reduced Tau (Figure 6(F)) and increased GFAP (Figure 6(J)) at 0%, 100% and either the 30% or 50% soy dose. ChAT immunoreactivity was also significantly elevated at 0% and 100% soy doses in the ethanol group (Figure 6(I)). Ubiquitin was significantly increased by ethanol only in the 50% dietary soy group (Figure 6(H)). No significant inter-group differences were observed for pTau across the soy dose range (Figure 6(G)).

Neurotrophin protein expression was measured by duplex ELISA with results normalized to acidic ribonuclear protein (RPLPO). Two-way ANOVA detected significant soy dose effects on brain-derived neurotrophic factor (BDNF), nerve growth factor beta (NGFβ), glial-derived neurotrophic factor (GDNF), Neurotrophin 3 (NT3), and NT4, but significant ethanol effect restricted to GDNF, and significant soy dose x ethanol interactive effects for BDNF (Table 8). Box-and-whisker plots were configured to display soy dose effects in control and ethanol-exposed cerebella (Figures 7(A)-(E)) or to compare ethanol relative to control at each dietary soy dose (Figures 7(F)-(J)), with the significant post hoc Tukey results shown.

Table 8. Two-way ANOVA table for neurotrophins.

Immunoreactivity was measured by direct-binding ELISA in 50 ng samples of cerebellar protein homogenates. Immunoreactivity was measured with primary antibodies targeting brain-derived neurotrophic factor (BDNF), nerve growth factor (NGF), glial-derived neurotrophic factor (GDNF), Neurotrophin 3 (NT3), or Neurotrophin 4 (NT4), horseradish peroxidase-conjugated secondary antibody, and Amplex UltraRed soluble fluorophore. Fluorescence intensity (Ex 560 nm/Em 590 nm) was measured in a SpectraMax M5. Results were normalized to the internal control ribosomal protein, RPLPO. The data was analyzed using mixed-model ANOVA and post-hoc Tukey tests (See Figure 7). P < 0.05 was considered statistically significant (bold font). NS = not significant.

In contrast to the neuroglial proteins, the neurotrophins were concordantly modulated with dietary soy dose in control and ethanol samples (Figures 7(A)-(E)). For BDNF, the levels of immunoreactivity were relatively consistent across the dietary soy dose range in both control and ethanol samples, except for the significantly reduced level observed in the 50% soy + ethanol group (Figure 7(A)). For NGFβ, cerebellar immunoreactivity was similar from 0% to 50% soy but significantly increased in both control and ethanol groups exposed to 100% soy (Figure 7(B)). In addition, a modest dose-dependent downward trend in NGFβ was detected in ethanol samples from the 10%, 30%, and 50% soy models. The profiles observed for GDNF (Figure 7(C)), NT3 (Figure 7(D)), and NT4 (Figure 7(E)) were similarly arched, with generally dose-related increases in immunoreactivity from 0% to 30% or 50% soy, followed by a tapered response to 100% dietary soy. Graphs comparing dietary soy dose effects in ethanol versus control samples (Figures 7(F)-(J)) showed relatively few significant differences. Those differences were restricted to ethanol-associated significant reductions in BDNF in samples from the 10% soy model (Figure 7(F)), and GDNF from the 10% and 100% soy models (Figure 7(H)).

Figure 7. Dietary Soy Dose and Ethanol Effects on Cerebellar Neurotrophin Expression. Cerebellar tissue from the offspring of dams fed with isocaloric liquid diets containing 0% or 26% caloric ethanol and 0%, 10%, 30%, 50%, or 100% soy as the protein source, with casein added to meet 100% of the dietary protein requirement was used to measure immunoreactivity to (A, F) brain-derived neurotrophic factor (BDNF), (B, G) Nerve growth factor beta (NGF-β), (C, H) glial-derived neurotrophic factor (GDNF), (D, I) Neurotrophin 3 (NT3), and (E, J) Neurotrophin 4 (NT4) by duplex ELISA. Results were normalized to acidic ribosomal phosphoprotein PO (RPLPO) housekeeping molecule. Graphs (A)-(E) show within-group (control or ethanol) effects of soy dose on immunoreactivity levels, and (F)-(J) show the between-group differences in responses to increasing soy dose. Data were analyzed by two-way ANOVA (Table 8) and the Tukey post hoc test. Significant differences (P < 0.05) are shown within the panels.