Fig 2: Gal-3 deficiency results in placenta insufficiency.a Relative mRNA expression of trophoblast differentiation genes determined by quantitative PCR (qPCR) on E10 and E13 placentas (n = 6–8). b 2D pre-contrast anatomical overview of MR images through abdomen of representative Lgals3+/+ (WT) and Lgals3-/- (KO) dams. Placentae labeled with false color transparent overlay (top left panel). 3D post-contrast maximum intensity projection angiograms of the mouse abdomen to determine placental volumes based on contrast enhancement (top right panel). 3D reconstruction of individual placentas ordered by volume (largest to smallest from left to right) are shown pair wise for Lgals3+/+ (WT) and Lgals3-/- (KO) animals. For better visual comparison, placental volumes were projected onto discs shown above each reconstruction illustrating smaller placentas in the Lgals3-/- animals (bottom panel, bar = 8 mm). c Group comparison of placental volumes over all animals confirmed a decrease in Lgals3-/- animals (top panel). Series of post-contrast images were acquired and mean signal was quantified repetitively for each placenta over ~10 min for functional MRI phenotyping. Initial enhancement, representative of maternal fractional blood volume in percent was increased in Lgals3-/- placentae whereas the rate of enhancement representative of several functional parameters such as perfusion, contrast agent inflow, clearance, and uptake was unchanged. Notably, a higher variability was found in KO animals for both, initial (middle panel) and rate (bottom panel) of enhancement (n = 8–16). d Isolectin B4 (IB4) staining showed poor development of fetal capillaries in the labyrinth of placentae of Lgals3-/- KO mice as revealed by AngioTool analysis (Total vessel length and vessel area; left bar = 500 µm and right = 150 µm, n = 25–34). In all figures, data are expressed as mean ± SEM. *P < 0.05, †P < 0.01, and ‡P < 0.001 using two-tailed t test.

Fig 3: Gal-3 deficiency leads to FGR.a Macroscopic appearance of the implantation sites on embryonic day (E)13. The pictures show the normal phenotype of Lgals3+/+ (WT) compared to Lgals3-/- (KO) dams, which exhibit resorbed fetuses (left panel, bar = 1 cm). Abortion rate (%) was calculated on E13 as follows: abortion rate = (fetal resorptions x 100)/total number of implantations (right panel, n = 19–24). b Fetal body weight (in grams) of E13 and E17 fetuses carried by Lgals3+/+ (WT) or Lgals3-/- (KO) dams (left panel, n = 13–21). Theiler stage developmental analysis on E13, upper arrows show that pinna is not turned forward and lower arrows show absence of separated fingers in KO fetuses (bar = 0.25 cm). On E17, upper arrows display fewer skin wrinkles, smaller whiskers, and lower arrows show that eyes are clearly visible through the eyelids in gal-3 KO fetuses (bar = 0.5 cm) (right panel). c Mean body weights of pups from Lgals3+/+ and Lgals3-/- dams. Reduced neonatal body weight was observed in Lgals3 KO offspring, which persisted until postnatal day (P)14 (n = 55–161). d E17 asymmetrical FGR as identified by fetal brain-to-liver weight ratio (n = 15–30). e Placental weight (mg) observed on E13 and E17 (left panel, n = 13–21). H&E histological analysis of E10 and E13 implantation sites shows an enlarged decidua in Lgals3-/- mice (middle panel, bar = 500 µm). Isolectin B4 (IB4) analysis of placental depth related to the decidua was significantly lower in placentas of Lgals3 KO mice (right panel, n = 7–8). f Accumulation of mature NK cells (PAS+DBA+) were significantly elevated at E13 (left panel, bar = 500 µm, n = 15) and their perforin granules (arrows) were released in gal-3 deficient pregnancy (right panel, bar = 50 µm, n = 8–11). g Progesterone levels measured in E13 serum by ELISA (n = 5). h Quantitative PCR (qPCR) analysis of decidual IFN-?, TNF-a, TGF-ß, IL-10, and IL-17 gene expression in Lgals3+/+ (WT) and Lgals3-/- (KO) dams at E7 (n = 4–9). In all figures, data are expressed as mean ± SEM. *P < 0.05, †P < 0.01, and ‡P < 0.001 using two-tailed t test.

Fig 4: Down-regulation of galectin-3 by siRNA in HTR-8/SVneo cells (abbreviated gal-3 in the figure). Silencing efficiency was analysed by qPCR (a) and Western blot (b) in control (L, open bars), scramble-transfected (S, light gray bars) and galectin-3 transfected (sigal-3, dark grey bars). LGALS3 was significantly reduced in galectin-3 silenced cells, while it had no effect on LGALS1 (a middle) or LGALS8 expression (a right), as evidenced by qPCR. One-way analysis of variance (ANOVA) with Tukey post-hoc test (a = 0.05) was used when data passed Shapiro-Wilk normality test (LGALS1 and LGALS8). When data didn’t pass normality test (LGALS3), non-parametric Kruskal-Wallis test was used followed by Dunn’s Multiple Range test (a = 0.05) for post-hoc comparisons. (b) Western blot analysis of galectin-3 in control, scramble-transfected and galectin-3 transfected cells, using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as the loading control, showed significant reduction of galectin-3 compared to the control (S). (b) The graph gives the average of three independent experiments. Data passed Kolmogorov-Smirnov normality test and were analysed by one-way analysis of variance (ANOVA) with Tukey post-hoc test (a = 0.05). (c) Galectin-3 in cellular compartments in galectin-3 transfected HTR-8/SVneo cells. The abbreviations for subcellular fractions are: C – cytoplasmic, M – membrane, Ns – nuclear soluble, Nc – nuclear chromatin. Differences were significant at ***p < 0.001.

Fig 5: Maternal gal-3 is necessary for normal placental and fetal development.a Circulating maternal gal-3 evaluated by ELISA in PRINCE cohort of patients (n = 65) during first, second, and third trimester, where some women subsequently development FGR (n = 32). b Western blot analysis of gal-3 expression in placental tissue from normal, healthy human pregnancies (NP) (n = 27) and pregnancies complicated by FGR (n = 19). Densitometry was performed to quantify changes in protein expression using ImageJ. c Representative images of gal-3 and cytokeratin (CK) 7 immunohistochemistry in NP and FGR placentae (n = 14, bar = 100 µm; left panel). The percent staining area of gal-3 and CK7 slides was quantified, and a ratio was calculated by dividing the percent area of gal-3 to CK7 staining in slides from the same placental tissue sample (right panel). d Schematic diagram showing the lack of paternal (mating of Lgals3-/- (KO) male with Lgals3+/+ (WT) female, pKO) and maternal (mating of Lgals3+/+ (WT) male with Lgals3-/- (KO) female, mKO) gal-3. The offspring is heterogeneous (HT) for gal-3 in both mating combinations (left panel). The abortion rate was calculated on E13 as follows: abortion rate = (fetal resorptions x 100)/total number of implantations (middle panel, n = 6–8). Pictures display whole implantation sites on E13. Arrows point to fetal resorptions (right panel, bar = 1 cm). e The embryo body weight was determined on E13 to identify potential fetal growth restriction (FGR) (n = 6). To assess the fetal development, Theiler stage (TS) analysis was performed. Normally developed embryos on E13 are in TS22 (pKO). FGR embryos are in a lower stage (e.g., TS21 or 21.5) and are characterized by a pinna that is not turned forward and non-separated fingers (arrows; mKO) (bar = 0.25 cm). f Placental efficiency was calculated as fetal/placental weight ratio on E13 (n = 14–21). g Glycogen cells in the spongiotrophoblast of the placenta were stained with periodic acid Schiff (PAS) and counted in 3 squares (1 × 1 mm) per mouse on E13 (n = 14–21). The number of mature uterine natural killer (uNK) cells in the decidua basalis were counted in 3 to 4 squares (1 × 1 mm) per mouse (n = 19–26). In all figures, data are expressed as mean ± SEM. *P < 0.05, †P < 0.01, and ‡P < 0.001 using two-tailed t test.