AAT/SERPINA1 Pi*Z variant linked to gestational length and preterm birth prevention via AAT in mice Authors: E Koivulehto, A Pasanen, A M Haapalainen, H Tiensuu, M Rämet, M Hallman Published: 2026 ABSTRACT AbstractAbout 10% of pregnancies end prematurely before 37 weeks, without effective prevention therapies. Previously, we identified rare damaging variants in SERPINA1 encoding alpha-1-antitrypsin (AAT) in families with recurrent spontaneous preterm births and detected decreased protein and transcript levels of AAT/SERPINA1 from placentas in spontaneous preterm births. Here, we investigate genetic associations between SERPINA1 variants and gestational duration and evaluate AAT supplementation as a therapeutic intervention in a mouse model of preterm birth. SERPINA1 Pi*Z variant (rs28929474-T) is associated with gestational duration (P < 5×10−8) in European-ancestry mothers with spontaneous preterm and term deliveries. We detect a nine-day decrease in gestational duration and a fourfold odds for preterm birth vs. term birth in Pi*Z homozygotes (Pi*ZZ), compared to other genotypes. In transgenic mice without endogenous AAT, exogenous Prolastin® treatment inhibits lipopolysaccharide induced preterm births (P < 0.05). Supplemented AAT is preferentially deposited in the placenta. Our findings support AAT’s protective role in spontaneous preterm births and highlight its therapeutic potential, particularly in SERPINA1 Pi*ZZ genotype carriers.Subject terms: Molecular medicine, Paediatric research, Genetic association study Introduction Annually, nearly 15 million infants are born prematurely, affecting approximately 10% of all pregnancies1. Prematurity is associated with nearly one million neonatal deaths and potentially lifelong morbidities in survivors, such as respiratory and neurodevelopmental impairments2. Notably, spontaneous premature births (SPTB) account for approximately 70% of all PTB cases3. The etiology of SPTB is intricate and multifactorial, reflecting the heterogeneity of its underlying causes and clinical presentations4,5. Emerging evidence suggests that during mid-gestation, an imbalance between pro-inflammatory and anti-inflammatory pathways may disrupt immune homeostasis, triggering SPTB6–8. The suggested key mediators include well-characterized cytokines and chemokines such as IL-1β, IL-6, IL-8, and TNF-α9, as well as a myriad of anti-inflammatory proteinases, including members of the SERPIN family such as SERPINA110,11. Alpha-1-antitrypsin (AAT) is a serine protease inhibitor encoded by the serine protease inhibitor (SERPIN) family member SERPINA112. AAT can neutralize neutrophil elastase (NE), which is essential for neutrophil chemotaxis and tissue remodeling13. Under the influence of inflammatory cytokines (e.g., IL-6 and IL-8), the concentration of AAT can increase up to fourfold14. AAT deficiency (AATD), a rare genetic disease, has been associated with various lung-related morbidities, including asthma, emphysema, and COPD 1315 as well as liver diseases such as fibrosis16. To date, over 500 SERPINA1 variants have been reported in the ClinVar database. Among these, Pi*Z (rs28929474) and Pi*S (rs17580) are the most notorious, being associated with AATD17. Case reports of patients who are either heterozygous or homozygous for these variants, i.e., carrying one or two copies of the risk allele, have demonstrated low circulating levels of AAT (<0.5 g/L) and impaired lung function17,18. Finally, the NHGRI-EBI Catalog of human genome-wide association studies (https://www.ebi.ac.uk/gwas/) reports 391 associations for the Pi*Z variant/rs28929474 T-allele, of which an association with decreased AAT-levels is the strongest. During pregnancy, mean serum levels of AAT gradually increase, with a marked rise by the third trimester, reaching 4–6-fold higher (3–4 g/L) than baseline. At delivery, AAT levels remain elevated compared to nonpregnant controls19. Alongside its inhibitory function in inflammation, AAT has proposed roles in endometrial angiogenesis and vascularization, as well as in trophoblast invasion and embryo implantation20. Low serum AAT levels (i.e., AATD) have been linked to elevated inflammatory cytokines as well as adverse pregnancy outcomes such as spontaneous abortions and PTB11,17,21,22. Decreased expression of AAT at the basal placenta was shown to be associated with SPTB, leading to the hypothesis that reduced AAT levels may result in enhanced degradation of placental fibrinoid deposits and the release of proteins into the cervix, potentially triggering SPTB11. This finding highlights AAT augmentation as a potential treatment option for addressing imminent SPTB. As AATD is often diagnosed after the onset of significant symptoms23, it is underdiagnosed in women of reproductive age22,23. Nonetheless, AAT augmentation has been used safely to relieve lung and liver morbidities during pregnancy17,21,24. Whether AATD augmentation in pregnant women with subclinical disease can reduce the risk of SPTB remains to be determined. In the present study, we investigated the genetic and molecular biological characteristics of SERPINA1/AAT in the context of SPTB and explored AAT therapy as a potential intervention. We first examined whether the AAT deficiency-associated Pi*Z variant (rs28929474) is a potential gestational modulator by analysing data from mothers who delivered preterm or at full term. To evaluate the therapeutic potential, we tested the efficacy of AAT supplementation (Prolastin®, Grifols) in an AATD transgenic mouse model of inflammation-induced preterm birth. Results Genetic associations of SERPINA1 with gestational duration As rare damaging variants in SERPINA1 have been linked to recurrent spontaneous preterm births11, we analysed whether SERPINA1 variants are associated with gestational duration or SPTB in European-ancestry mothers with spontaneous preterm and full-term births. As shown in Fig. 1 and Table 1, the SERPINA1 Pi*Z variant (rs28929474-T; missense C > T; risk allele = T) was significantly associated with decreased gestational duration (Fig. 1 and Table 1). Genetic analysis of SERPINA1 Pi*Z variant: Locuszoom plot of the rs28929474 association with gestational duration from the GWAS meta-analysis. Association P-values from linear regression (–log10 scale) are shown on the Y-axis, and the dashed line denotes the genome-wide significance threshold (P < 5 × 10e-08). On the X-axis, base-pair positions around the SERPINA1 association in chromosome 14 are shown. Each dot represents an individual variant, with rs28929474 shown as a purple diamond, and linkage disequilibrium (LD) color indicating the strength of genetic correlation between adjacent variants. Table 1Association statistics of the SERPINA1 Pi*Z variant (rs28929474) with gestational duration and spontaneous preterm birthGestational durationFreqP-valueBeta/Z- scoreSESample sizeStudy data0.03964.33e-08−5.477−219,602Meta-analysis0.01961.75e-06-0.0920.019367,615FinnGen0.03969.73e-04-0.6880.2086151,987EGG summary dataSpontaneous preterm birthFreqP-valueOdds Ratio95% CISample sizeStudy data0.01881.16e-041.1581.083–1.232344,628Meta-Analysis0.01974.25e-041.2601.131–1389111,338FinnGen0.01842.62e-021.1091.018–1.202233,290EGG summary data0.01976.16e-031.1511.050–1.1.251196,799FinnGen, additive model0.01979.43e-064.3323.684–4.981196,799FinnGen, recessive modelIn the meta-analysis of spontaneous preterm birth (SPTB), FinnGen DF5 individuals were excluded from the FinnGen dataset because they were included in the EGG SPTB dataset. Frequency (Freq) and effect estimates are reported for the T allele. For gestational duration, effect estimates are reported as beta coefficients (GWAS) and Z-scores (meta-analysis), whereas for spontaneous preterm birth, effect estimates are reported as odds ratios (OR). SE, standard error (reported for gestational duration); CI, 95% confidence interval (reported for SPTB). Association statistics of the SERPINA1 Pi*Z variant (rs28929474) with gestational duration and spontaneous preterm birth In the meta-analysis of spontaneous preterm birth (SPTB), FinnGen DF5 individuals were excluded from the FinnGen dataset because they were included in the EGG SPTB dataset. Frequency (Freq) and effect estimates are reported for the T allele. For gestational duration, effect estimates are reported as beta coefficients (GWAS) and Z-scores (meta-analysis), whereas for spontaneous preterm birth, effect estimates are reported as odds ratios (OR). SE, standard error (reported for gestational duration); CI, 95% confidence interval (reported for SPTB). Analysis of rs28929474 genotypes in the FinnGen data revealed a nine-day difference in the median gestational duration between TT and other genotypes (CC/TC) (P = 2.53e-04 [Mann–Whitney–Wilcoxon]; Fig. 2). The effect on gestational duration was more pronounced in mothers with SPTB compared to mothers with full-term deliveries (Supplementary Fig. 1). This suggests that other genetic or environmental factors may modify the association between rs28929474-T and gestational duration. In the context of liver disease, high BMI was shown to amplify the association between the PI*Z allele and markers of liver damage25. Therefore, we conducted a gene-by-environment interaction analysis of the SERPINA1 Pi*Z variant and BMI and observed a similar pattern of worsened outcome (i.e., lower gestational duration) with increasing BMI (P = 0.011; Supplementary Fig. 2). SERPINA1 Pi*Z variant (rs28929474) genotypes and gestational duration. Considering both term and preterm births, mothers with the TT genotype (Pi*ZZ) gave birth approximately nine days earlier than mothers with the CC (P = 2.39e-04) or CC/CT (P = 2.53e-04) genotypes. A two-sided Mann–Whitney-Wilcoxon test was used to compare the genotypes. Box plots show the median (center line), interquartile range (box limits corresponding to the 25th and 75th percentiles), and whiskers extending to the most extreme values within 1.5× the interquartile range. The genotype association pattern of the SERPINA1 Pi*Z variant (Fig. 2) resembled that of a recessive model, i.e., people with two copies of the Pi*Z variant were significantly more affected compared to heterozygous carriers and non-carriers. Thus, we further tested the recessive spontaneous preterm birth model in the association analysis of the FinnGen data and observed a larger effect size (Table 1). This result is suggestive that the effect on shorter gestational duration and preterm birth is more potent in individuals with the SERPINA1 PI*ZZ (rs28929474-TT) genotype. In conclusion, the SERPINA1 Pi*Z variant appears to be a significant modifier of gestational duration in mothers of European ancestry. AAT pretreatment prevents premature birth in knockout mice To assess whether AAT supplementation could be used to prevent inflammation-induced preterm birth, we utilized Serpina1a–e KO mice and wild-type mice. In the transgenic Serpina1a–e KO mouse strain, average gestation ranged from 19.5–20 days post-coitum (dpc). We used LPS to induce PTB. PTB was defined as delivery before 18.5 dpc, coinciding with the sample collection endpoint (48 h post-LPS injection). To evaluate the preventative effect of AAT, Prolastin® (AAT) or placebo was administered as a single intraperitoneal (i.p.) injection 2.5 h prior to LPS challenge (Supplementary Fig. 3A). Animals were monitored during this interval, and no adverse effects were observed before LPS administration. LPS administered i.p. at 16.5 dpc triggered dose-dependent inflammatory symptoms (panting, disorientation, lethargy, eye squinting, agitation) in both WT and KO dams. A dose of 1.25 μg was selected for further studies, as it resulted in moderate inflammatory symptoms and shortened pregnancy in most, but not all, Serpina1a–e KO and wild-type mice (Supplementary Fig. 4). As shown in Fig. 3A, AAT efficacy in preventing inflammation-induced PTB varied by genotype (Fig. 3A, B). While AAT supplementation had no apparent effect on the length of gestation upon LPS challenge in WT mice, it had 50% efficacy (6/12) in KO mice (Fig. 3A, B). Dunn’s multiple comparisons test (FDR-adjusted) showed that KO-AAT dams had significantly prolonged gestation compared to KO-Placebo (P = 0.032), WT-AAT (P = 0.047), and WT-Placebo (P = 0.043) (Fig. 3A). Birth outcomes in wildtype (WT) and Serpina1a-e knockout (KO) dams treated with placebo or AAT and challenged with lipopolysaccharide (LPS) at 16.5 days post coitum (dpc). A KO-AAT dam had a significant delay in labor onset compared to other groups (KO-Placebo: *, P = 0.032; WT-AAT: *, P = 0.047; WT-Placebo: *, P = 0.043). B Survival curve (Kaplan–Meier analysis) demonstrating treatment efficacy (%) over time (hours post LPS injection) across genotype-treatment groups. Statistical analyses were performed using Kruskal–Wallis with Dunn’s post hoc tests and Benjamini–Hochberg correction. Significance levels were indicated as: P < 0.05 (*), < 0.01 (**), < 0.001 (***), and < 0.0001 (****). Source data are provided as a Source Data-file. In a separate pilot experiment, AAT administered 2 h after LPS had no protective effect; all eight (8) dams (4 WT dams; 4 KO dams) were delivered prematurely within 20–24 h. These data indicate that pretreatment with supplemented AAT protects AAT-deficient mice from LPS-triggered SPTB. AAT supplementation improves survival rate and viability of pups Next, we evaluated the effect of AAT supplementation on fetuses. Developmental assessment using crown–rump (CR) length measurements in pups surviving beyond 48 h—corresponding to the same dams analysed in Fig. 3 (Cohort 1)—indicated that Prolastin® treatment did not impair fetal growth (Fig. 4A, B). KO-AAT litters had significantly longer CR lengths compared to both KO-placebo (P = 0.025) and WT-placebo litters (P = 0.023), indicating continued fetal growth toward term with AAT supplementation. CR lengths were consistent with expected developmental milestones and reflective of the gestational age at delivery for each litter. For pups that reached 18.5 dpc, CR growth between 16.5 and 18.5 dpc averaged approximately 0.6 cm. Litter outcomes in WT and Serpina1 KO dams treated with placebo or AAT and challenged with LPS at 16.5 dpc. A Crown–rump (CR) length measurements are presented as the mean value per litter (n = 12 litters per genotype–treatment group; 48 litters total). Plots show the median (center line), interquartile range (box limits corresponding to the 25th and 75th percentiles), and whiskers extending to the minimum and maximum values within 1.5× the interquartile range. Individual data points are shown. Pups from KO-AAT dams showed significantly increased CR length at 18.5 dpc compared to KO- and WT-placebo groups (*, P = 0.025; P = 0.023). B Schematic presentation of developmental milestones across late gestation (from 16.5–18.5 dpc). C Kaplan–Meier analysis, based on binary survival data (alive = 1, dead = 0) with pups assigned to timepoints corresponding to survival duration, indicated improved overall postnatal survival in KO-AAT pups (adjusted global log-rank P = 0.0018). D The same binary survival data were summarized by genotype–treatment group to calculate overall postnatal viability, presented as bar plots (**** P = 9.4e–08; *** P = 2.3e–05; ** P = 0.003), with KO-Placebo pups also outperforming WT-AAT (*, P = 0.045). Statistical significance was assessed via Kruskal–Wallis and Dunn’s tests (continuous data), and Pearson’s chi-square with Fisher’s exact tests (categorical data), with BH correction applied. Significance levels were indicated as: P < 0.05 (*), < 0.01 (**), < 0.001 (***), and < 0.0001 (****). Source data are provided as a Source Data-file. KO pups treated with Prolastin® showed the highest overall survival probability (Fig. 4C, D). Global survival analysis revealed significant differences among groups (adjusted log-rank P = 0.0018). Following euthanasia of the dam after the delivery of the first pup, most remaining in utero pups were viable (Supplementary Fig. 5A), showing intact placental attachment and umbilical vein engorgement. In contrast, pups delivered vaginally before 18.5 dpc were largely nonviable, regardless of treatment (Supplementary Fig. 5B). Consistent with the survival analysis, litters from KO dams treated with Prolastin® exhibited the highest viability percentages, with all phenotype comparisons reaching statistical significance (P = 9.4e–08; P = 2.3e–05; P = 0.003; Fig. 4D). These findings suggest that a selected dose of AAT is not toxic and does not compromise fetal growth. Instead, mortality in preterm-born pups likely reflects insufficient developmental immaturity at the time of delivery. Clearance of human AAT in mice is influenced by inflammation and genotype Next, we studied the kinetics of hAAT clearance after supplementation. A single 12.5 mg i.p. dose of Prolastin® elevated circulating hAAT to levels approximating the murine physiological range observed in non-gravid, unstimulated mice (~2 g/L; Fig. 5A). The end-point serum half-life averaged ~38 h post-injection. Dose of AAT was calculated based on the basal weight of young adult female mice (~20 g) and adjusted for reduced systemic absorption associated with the i.p route. Maternal weight gain during pregnancy was not factored into dose determination. Serum levels and pharmacokinetics of human AAT (hAAT) in Prolastin®-treated Serpina1a-e KO and WT mice. A Time course of hAAT levels in non-pregnant KO mice (n = 5 per timepoint) following Prolastin® administration. B hAAT levels in pregnant dams treated with Placebo and Prolastin® and challenged with LPS at 16.5 dpc (Cohort 1; n values for each time point and treatment group are indicated in the figure), measured at indicated time points. C hAAT levels in WT and KO dams at 16.5 dpc (Cohort 2; n = 3 per group) with or without LPS exposure. In non-pregnant KO mice (A), hAAT levels peaked at 6 h and declined with a terminal half-life of ~38 h. In pregnant dams, KO-AAT mice (B) showed a reduction in hAAT levels between 17.5 and 18.5 dpc (**, P = 0.0014), whereas no statistically significant difference was observed between WT-AAT and KO-AAT groups at 17.5 dpc. At 16.5 dpc (C), LPS exposure was associated with reduced hAAT levels in KO-AAT (1.49 g/L vs. 1.07 g/L, ns) but not in WT-AAT dams. Error bars represent the standard error of the mean (± SEM) of the biological replicates. Statistical comparisons were performed using Kruskal–Wallis tests with Dunn’s post hoc correction and Benjamini–Hochberg adjustment. Significance levels were indicated as: P < 0.05 (*), < 0.01 (**), < 0.001 (***), and < 0.0001 (****). Source data are provided as a Source Data-file. Prolastin®-treated gravid dams (Cohort 1), exposed to LPS, exhibited accelerated hAAT clearance compared to non-gravid, unstimulated mice from a separate control group (Fig. 5B). 24 h after administration of AAT at 16.5 dpc, mean serum hAAT concentrations were similar between KO-AAT (0.50 g/L, median: 0.63 g/L) and WT-AAT dams (0.51 g/L, median: 0.58 g/L). hAAT levels became undetectable by 48 h (18 dpc) in both groups (KO-AAT, P = 0.0014; WT-AAT, not significant), indicating rapid systemic clearance. To further assess the kinetics of hAAT clearance, serum hAAT levels were measured 4.5 h after Prolastin® administration (Cohort 2) with or without LPS (Fig. 5C). Mean hAAT levels were 1.46 g/L in WT dams and 1.49 g/L in KO dams, and 1.58 g/L (WT) versus 1.07 g/L (KO) when co-administered with LPS. These data suggest that LPS challenge reduces circulating hAAT in KO mice; however, formal statistical testing was not performed due to the low sample size of 3 per timepoint. Notably, WT dams maintained relatively high hAAT levels despite poor outcomes, suggesting that elevated circulating hAAT alone is insufficient to prevent inflammation-induced SPTB. Serum AAT levels were undetectable in pups born to Prolastin®-treated WT and KO dams, indicating that supplementary hAAT does not readily cross the placenta to enter the fetal circulation at physiologically relevant levels (Supplementary Fig. 6). LPS induces acute inflammation regardless of AAT supplementation To evaluate the immunomodulatory effects of Prolastin® (AAT) supplementation, we measured a broad panel of cytokines and growth-related mediators in Cohort 2 dams, two hours post-LPS stimulation (or in unstimulated controls) at 16.5 dpc. LPS exposure increased pro-inflammatory, anti-inflammatory, and growth-associated cytokines in both WT and KO dams (Fig. 6A–C and Supplementary Figs. 7 and 8). A single outlier in the KO-LPS group displayed low cytokine levels, possibly due to incomplete LPS exposure (Fig. 6A–C: 8). Cytokine and growth factor responses following placebo or AAT treatment with or without LPS exposure at 16.5 dpc. Heatmaps show z-score–normalized maternal serum levels of A pro-inflammatory, B anti-inflammatory, and C growth-associated cytokines and chemokines in WT and Serpina1a-e KO dams treated with placebo or AAT, with or without LPS exposure (sample number 1–12). LPS exposure induced a robust increase across all cytokine classes, with no detectable suppression by supplementary AAT. Sample group 8 (KO + LPS) exhibited low-cytokine levels and is considered an outlier, likely reflecting incomplete LPS exposure. Source data are provided as a Supplementary data file. Among pro-inflammatory cytokines, IL-6, TNF-α, IL-1α, IL-1β, and IFN-γ were significantly elevated, along with chemokines CXCL1 and CCL2 (MCP-1), suggesting neutrophil and monocyte recruitment. Increases in IL-18, CXCL10, and CCL3 (MIP-1α) further indicated leukocyte activation, consistent with TLR4 activation via MyD88 and TRIF pathways. AAT supplementation did not significantly reduce these responses relative to placebo. LPS also upregulated anti-inflammatory cytokines, including IL-10, IL-1ra, IL-4, IL-13, LIF, and IL-27 (Fig. 6B: 7–12), likely reflecting a compensatory mechanism to limit inflammation. While variability was observed across LPS-treated animals, AAT treatment did not affect the expression of these markers beyond levels seen in placebo-treated dams. In addition, angiogenic and tissue integrity markers—including VEGF, ICAM-1, GDF-15, FGF-basic, TIMP-1, LDLR, and EGF—were elevated following LPS exposure (Fig. 6C: 7–12), with ICAM-1, VEGF, TIMP-1, and GDF-15 showing the most consistent upregulation. AAT had no significant effect on the expression of these growth-associated factors. A broader cytokine and chemokine panel (Supplementary Fig. 8A) confirmed general immunomodulatory trends in dams. LPS exposure significantly increased colony-stimulating factors (M-CSF, G-CSF, GM-CSF), T cell–associated cytokines (IL-17A, IL-21), and other inflammatory mediators. Prolastin® did not have a significant modulatory effect on these values. Overall, LPS triggered a systemic inflammatory response involving both immune and tissue remodeling pathways. AAT supplementation did not significantly alter individual cytokine levels. Hence, AAT’s protective effects against LPS-induced PTB in KO dams, and its lack of efficacy in WT dams, suggest alternative mechanisms, such as placental stabilization, in the prevention of inflammation-driven preterm birth. Fetal (pup) serum cytokine profiles were also analysed across all groups (Supplementary Fig. 7B). Unlike the maternal compartment, fetal cytokine levels remained largely unchanged by genotype or treatment. Key inflammatory markers (e.g., IL-6, IL-1β, IL-10, CXCL10, VEGF) remained at baseline, indicating that acute maternal inflammation did not lead to a systemic fetal immune response. Of note, WT-mock-treated pups had the highest cytokine levels across analytes (Supplementary Fig. 9B:1), likely reflecting developmental maturation rather than inflammatory activation. This suggests that maternal inflammation is compartmentalized, with the placenta effectively shielding the fetus from systemic immune activation. These findings support the hypothesis that improved pregnancy outcomes in KO-AAT dams—and the lack of benefit in WT-AAT dams—are driven by maternal or placental mechanisms rather than fetal immune modulation. Modulation of endogenous AAT at maternal-fetal interface To assess the localization and degradation of endogenous mouse AAT (mAAT), we performed immunohistochemistry (IHC) and Western blot (WB) on Cohort 1 placentas collected at 17.5 and 18.5 dpc (Fig. 7A–C). The basal plate (BP) included decidua and junctional zone, while the chorionic plate (CP) corresponded to the labyrinth layer. AAT distribution was also evaluated in the amniotic membranes (amnion and the yolk sac; here referred to as AMS). Placental immunohistochemistry and Western blot analysis of endogenous murine AAT (mAAT) expression in WT and Serpina1a-e KO dams treated with placebo or Prolastin®, and LPS. A IHC images (scalebar 50 μm) show mAAT localization in the basal plate, chorionic plate, and amniotic sac at 17.5 days post coitum (dpc) (preterm birth; PTB) and 18.5 dpc (term birth; TB). B mAAT-positive area (% DAB staining) was quantified by region, treatment, and outcome. C Placental Western blots of mAAT and GAPDH show expression patterns across genotypes and birth outcomes. Prolastin® treatment decreased mAAT in WT-treated dams at term. Endogenous mAAT, detected by anti-mAAT IHC, showed genotype- and treatment-dependent patterns (Fig. 7A, B). In KO dams, mAAT is localized mainly to the CP and AMS, likely derived from heterozygous fetal tissues (Fig. 7A, B and Supplementary Fig. 10). Placebo-treated WT dams exhibited strong mAAT staining across CP, BP and AMS (Supplementary Fig. 9 and Fig. 7B), localized in blood vessels, extracellular matrix, and intracellular compartments. In Prolastin®-treated WT dams, mAAT staining was elevated at 17.5 dpc but decreased by 18.5 dpc, reflecting dynamic regulation in response to supplementation (Fig. 7A, B). AMS staining was higher in WT dams, with possible mAAT aggregation at 17.5 dpc, reflecting either a fetal or maternal response to LPS. WB analysis of placental lysates (Fig. 7C) confirmed these findings: WT dams showed a consistent mAAT (~54 kDa, fully glycosylated isoform) across treatments and timepoints. No partial or unglycosylated forms were detected, implying limited de novo synthesis or rapid glycosylation under inflammation. KO placentas lacked mAAT, supporting minimal fetal-to-maternal transfer. WT liver samples confirmed systemic AAT upregulation, while KO livers showed no detectable mAAT (Supplementary Fig. 11A). Overall, endogenous AAT production and retention appear highly dynamic, influenced by genotype and exogenous AAT exposure. The observed reduction in IHC staining and altered isoform abundance suggest that exogenous hAAT may transiently suppress endogenous mAAT expression and affect its post-translational modifications. This may, in part, explain why AAT fails to inhibit LPS-induced preterm birth in WT mice. Exogenous AAT is retained in the placenta following hAAT supplementation To assess the localization and quantity of exogenous human AAT (hAAT, Prolastin®), we performed anti-hAAT IHC and WB analyses corresponding to those used for endogenous murine AAT above (Fig. 8A–C). As expected, no hAAT staining was observed in placebo-treated WT or KO dams (Supplementary Fig. 10 and Fig. 8B). Prolastin®-treated groups demonstrated genotype-dependent patterns of hAAT localization at 17.5 and 18.5 dpc (Fig. 8A, B). In KO tissues, hAAT localized to vascular and extracellular regions of both the basal plate (BP) and chorionic plate (CP), remaining detectable through 18.5 dpc. Conversely, WT placentas exhibited lower hAAT signals, consistent with reduced retention or faster clearance, possibly due to functional hepatic pathways or competitive inhibition by endogenous mAAT. Placental immunohistochemistry and Western blot analysis of exogenous human AAT (hAAT) in WT and Serpina1a-e KO dams treated with placebo or Prolastin®, and LPS. A An IHC (scalebar 50 μm) showing hAAT localization in the basal plate, chorionic plate, and amniotic sac at 17.5 days dpc (preterm birth; PTB) and 18.5 dpc (term birth; TB). B quantification of hAAT-positive area (% DAB staining) by region, treatment, and birth outcome. C Placental Western blots of hAAT and GAPDH (loading control from Fig. 7C) across genotypes and birth outcomes. In combination, results show faster clearance of hAAT in WT dams and prolonged retention in KO dams, with minimal detection of hAAT in WT placental lysates at term. hAAT staining persisted longer in KO-AAT/Prolastin® dams, with a visible signal at 18.5 dpc, compared to WT dams. Detected hAAT in the amniotic membranes (AMS) likely originates from maternal uterine lining and circulation rather than pup circulation (Fig. 8A, B), differing from the fetal source of mAAT (Fig. 7A, B). This suggests that exogenous hAAT accumulates more efficiently in the absence of endogenous mAAT, indicating tight control of AAT levels in placental tissues. Western blot analysis confirmed placental retention of hAAT but showed no detectable uptake in the liver (Supplementary Fig. 11B). This is consistent with localized placental accumulation. hAAT appeared only at a molecular weight of approximately 54 kDa, indicating a fully glycosylated, mature protein. These results validate effective delivery of hAAT to the maternal-fetal interface with limited systemic distribution following intraperitoneal administration. In summary, combined IHC and Western blot analyses demonstrate that endogenous AAT expression is dynamically modulated by exogenous Prolastin®, with notable genotype-dependent differences in localization, retention, and glycosylation status. Exogenous hAAT accumulates at the placenta, highlighting its targeted bioavailability. The different retention patterns between KO and WT dams emphasize the critical role of endogenous AAT in clearance and tissue distribution of therapeutic AAT. Serpina1a–e mRNA levels are modulated by genotype and treatment To further evaluate the effects of exogenous hAAT on AAT regulation, Serpina1a–e transcript expression was assessed by RT-qPCR to determine the impact of exogenous AAT on endogenous Serpina1a–e mRNA levels during acute inflammation in liver samples from Cohorts 1–2. In Cohort 2 (16.5 dpc), LPS-treated WT dams (placebo or Prolastin®) showed broad downregulation of Serpina1a–e mRNA compared to mock-treated controls, with the lowest transcript levels observed in WT-AAT/Prolastin®+LPS animals, indicating tight regulation of AAT levels during acute inflammation. In mock-treated mice, WT livers expressed all Serpina1 paralogs, with Serpina1a showing the highest expression, whereas KO livers exhibited virtually no expression, consistent with gene deletion (Fig. 9A–E). Following Prolastin® treatment, WT livers displayed reduced Serpina1a–e expression, suggesting suppression of endogenous transcription by exogenous AAT. KO livers showed only minor Serpina1d mRNA signal, consistent with the absence of functional mAAT protein. Relative liver mRNA expression of Serpina1a–e transcripts in WT and Serpina1a–e KO dams at 16.5 dpc (Cohort 2), and 17.5 dpc and 18.5 dpc (Cohort 1). Plots A–E demonstrate individual Serpina1 paralogs a–e expression at 16.5 dpc (Cohort 2; n = 3 per group) by treatment group (1–8), normalized to Eef2. WT groups exhibited robust expression of all five paralogs, which was downregulated by Prolastin® treatment and/or lipopolysaccharide (LPS), with the lowest levels observed in the combined treatment group (group 4). KO groups showed minimal or absent transcripts, confirming knockout, with a negligible Serpina1d signal. Error bars represent the standard error of the mean (± SEM) between the three biological replicates. Plots F–J demonstrate corresponding Serpina1 paralog expression at 17.5 and 18.5 dpc (Cohort 1) under LPS challenge, normalized to Eef2. Due to limited sample availability at these time points, expression was assessed using one biological sample per group/timepoint with technical triplicate measurements. Technical replicates were averaged and are presented to illustrate expression patterns only. No statistical inference was performed based on these measurements. Error bars represent standard deviation (SD). In WT dams, Prolastin® treatment at 18.5 dpc (group 4) resulted in downregulated transcript levels, whereas WT-Placebo groups (1–2) showed a rebound upregulation following prior downregulation, suggesting feedback regulation and attenuation of inflammatory stimuli. KO dams lacked detectable functional transcripts across conditions. Source data are provided as a Source Data-file. To further investigate temporal regulation and the relative contribution of liver versus placenta to AAT production, Serpina1a–e transcript levels were analysed in liver samples at 17.5 dpc (PTB) and 18.5 dpc (TB) from Cohort 1 (Fig. 9F–J). In WT livers, Serpina1a–e transcripts were downregulated in PTB dams treated with placebo, consistent with earlier observations at 16.5 dpc. By TB, expression levels were restored. In contrast, WT dams treated with AAT/Prolastin® showed elevated transcript levels at PTB, followed by decreased expression at TB, indicating an initial transcriptional increase followed by feedback inhibition by exogenous AAT. KO dams showed no or minimal expression of Serpina1a–e (<0.13 normalized ratio; Fig. 9F–J), consistent with exon 2 deletions across all paralogs. In placentas, Serpina1a–e transcripts were undetectable across groups (Supplementary Fig. 12), confirming negligible local expression. These findings indicate that murine AAT is liver-derived and systemically delivered to the placenta, highlighting placental reliance on hepatic AAT for homeostasis. Consequently, systemic AAT deficiency may increase the vulnerability of placental structures. Discussion Our study provides evidence linking the SERPINA1 Pi*Z variant (rs28929474) to gestational duration and the risk of preterm birth (PTB) in humans and demonstrates that alpha-1-antitrypsin (AAT) supplementation inhibits LPS-induced spontaneous preterm birth (SPTB) in Serpina1a-e knockout (KO) mice. These findings emphasize AAT’s protective role in inflammation-driven pregnancy complications, although the precise mechanisms remain undefined, and underscore its therapeutic potential. Our genetic meta-analysis of the SERPINA1 PI*Z variant, using data from the FinnGen research project and the EGG consortium, identified a significant association between rs28929474 and gestational age. The variant has previously been implicated in intrahepatic cholestasis of pregnancy (ICP)26,27. Follow-up analysis in the FinnGen data revealed a large genotypic effect between the Pi*ZZ genotype and SPTB. Mothers who were homozygous for the Pi*Z allele exhibited significantly shorter gestational duration compared to heterozygous or non-carrier individuals. Among the women who gave birth preterm, the reduction in gestational duration was approximately 10 days for those with Pi*ZZ genotype, compared to women with other genotypes (242 vs. 252 days). The effect of the Pi*ZZ genotype became more pronounced with increasing BMI. Interestingly, a similar pattern of worsened outcome was previously detected in gene-environment interaction analysis in the context of the Pi*ZZ variant and liver disease. The substantial effect of SERPINA1 Pi*Z alleles on gestational duration could arise from a genetic susceptibility to inflammation-induced PTB mediated by AAT deficiency. This is in consort with our previous findings of the role of AAT/SERPINA1 at the basal placenta11. Our mouse model analyses revealed genotype- and treatment-dependent differences in AAT processing and distribution. Exogenous human AAT given intraperitoneally is localized in the placenta, rather than the liver, suggesting localized action at the maternal–fetal interface. WT dams treated with Prolastin® showed increased clearance of human AAT compared to KO dams, indicating an endogenous feedback loop. Both AAT deficiency and excess may be detrimental, supporting a threshold effect, whereby protective efficacy is achieved only within an optimal range of AAT. This is further evidenced by the downregulation of all Serpina1a–e paralogs after AAT administration and LPS challenge28. Although AAT has been shown to inhibit NF-κB signaling29, cytokine profiling revealed no detectable suppression of acute inflammatory markers within 2 h post-LPS. Placebo- and Prolastin®-treated mice exhibited similar proinflammatory cytokine profiles, indicating that the acute response to inflammatory mediators was not modulated. These data suggest that AAT’s protective effects are unlikely to be mediated through early modulation of acute inflammatory signaling in this model but may instead reflect its broader protease-inhibitory and immunoregulatory functions, including longer-term modulation of immune responses, maintenance of placental integrity, or vascular remodeling11,12,30. These findings align with the multifactorial nature of SPTB3,31,32. In placebo-treated WT-dams, LPS induced rapid downregulation of hepatic Serpina1a–e transcripts (Fig. 9), followed by restored expression by 18.5 dpc, likely a feedback mechanism restoring homeostasis. In contrast, the transcript suppression continued by 18.5 dpc in Prolastin®-treated animals, suggesting augmentation of endogenous feedback inhibition by exogenous AAT. This points to tightly regulated AAT homeostasis during pregnancy, modulated by endogenous and therapeutic factors. Future studies should utilize spatial transcriptomics or single-cell RNA-seq to provide detailed placental cell responses and employ advanced kinetic modeling or labeled AAT to distinguish exogenous vs. endogenous sources. LPS challenge elevated cytokine expression and induced labor; AAT supplementation improved pregnancy outcomes in KO dams but did not affect the expression of inflammation markers. This implies that other factors, such as maternal immune programming and maintenance of the structural integrity of the placenta, may influence the outcome33,34. AAT’s anti-protease effects may represent one facet of broader protective mechanisms. Though our focus was on preventing LPS-induced SPTB, AAT supplementation may also benefit fetal neurodevelopment. Prior work links intrauterine inflammation to neurodevelopmental impairments35. While pups were heterozygous, inflammation-induced growth restriction and neurodevelopmental risks persisted. The potential of AAT to protect fetal growth and neurodevelopment, possibly by mitigating fetal brain proinflammatory signaling or preserving placental barrier integrity, warrants further investigation. Zhang et al.36 demonstrated that AAT crosses the blood–brain barrier (BBB) in pups and alleviates hypoxia–ischemia brain injury, reducing BBB permeability, caspase-3 activation, neuronal death, and microglial activation, supporting its neuroprotective promise during gestation36. Regardless, the potential of AAT to preserve placental barrier integrity or mitigate fetal brain proinflammatory signaling warrants further investigation. We acknowledge the limitations of this study. These include the obvious interspecies differences in utero-placental physiology, anatomy, regulation of the length of pregnancy and the biology of AAT. Murine endogenous AAT levels are higher and less gestationally regulated than humans17,37. To achieve therapeutic levels, we used a single high dose (>600 mg/kg), exceeding human augmentation therapy doses (60 mg/kg) routinely applied for the treatment of AATD-associated lung disease38. The dose achieved near-physiological murine serum levels (~2 g/L; see Fig. 4A, B)39 but was rapidly cleared. Emerging evidence suggests higher human-equivalent doses (120–180 mg/kg) may enhance efficacy38,40,41, although no obstetric-specific dosing guidelines exist and safety and pharmacokinetics in human pregnancy remain to be established, emphasizing the need for dose optimization for plausible prenatal care. Nevertheless, our study establishes a clear biological rationale: AATD increases susceptibility to PTB, whereas restoration of the AAT levels confers protection, thereby supporting the investigation of augmentation strategies. Finally, we note that responses to AAT supplementation were heterogeneous in KO animals. This likely reflects threshold-dependent effects related to absent baseline AAT, local AAT bioavailability, and individual variability in the inflammatory response. We also acknowledge the limitations of our genetic findings. The FinnGen study enables well-powered analyses of genetic associations with disease endpoints derived from nationwide registries. However, the use of registry-based diagnoses may introduce misclassification, and residual confounding from comorbidities and related clinical factors cannot be fully excluded. In addition, the relatively homogeneous Finnish population and the restriction to individuals of European ancestry, while reducing potential bias from population stratification, may limit the generalizability of our findings. SPTB disproportionately affects populations that are underrepresented in European-ancestry cohorts; therefore, replication in ancestrally diverse populations is an important priority for future research. While functional evidence from our current and previous studies supports a predominantly maternal–placental mechanism of action for SERPINA1/AAT in pregnancy, we cannot exclude the possibility that fetal genetic effects may also contribute to the observed associations. In summary, our current study provides evidence linking genetic variation at SERPINA1 Pi*Z with variation in gestational duration and the risk of SPTB. Our data support a dual role for AAT in pregnancy, although the precise mechanisms remain undefined, encompassing reduced inflammation-induced SPTB in deficient individuals and potential protection of fetal growth and development. Our results suggest that exogenous AAT may cross the maternal-fetal interface (see Fig. 8), although its exact protective mechanism against imminent inflammation-induced PTB remains elusive. It is likely that AAT’s effects may involve preservation of placental integrity, including stabilization of the extracellular matrix and support of trophoblast function, as suggested previously11. The therapeutic window appears narrow, with risks posed by both deficiency and excess AAT. AAT’s benefits may extend beyond acute cytokine modulation to broader immunological and placental pathways, as also suggested by the epidemiological study of Orimoloye et al.22. Future research should clarify tissue-specific AAT regulatory networks, optimize supplementation strategies, evaluate maternal–fetal transfer of exogenous AAT, assess long-term offspring outcomes, and explore applications in other pregnancy complications, including ICP and preeclampsia13,26,28. Present data and emerging evidence underscore the importance of understanding systemic AAT deficiency as a risk factor for adverse pregnancy outcomes, with potential relevance to placental and fetal development. The established association of rs28929474 with gestational timing and adverse pregnancy outcomes, including intrahepatic ICP, suggests that SERPINA1-related pathways may contribute to multiple pregnancy complications through shared maternal biological mechanisms, although the specific pathways involved remain to be determined. In the present study, we discovered that genetic AAT deficiency is a strong risk factor for SPTB. Very premature delivery is associated with chronic neurodevelopmental and pulmonary diseases. Our transgenic mice study lays a strong foundation for translational research harnessing AAT as a therapy for inflammation-mediated pregnancy disorders. Given the clinical approval of AAT augmentation therapy for specific lung and liver diseases, exploring its potential prenatal and perinatal applications holds significant promise. Methods Genetic analysis of SERPINA1 We tested the association of SERPINA1 Pi*Z variant (rs28929474) with SPTB and gestational duration in mothers with term and preterm deliveries with data from the FinnGen research project DF12 (www.finngen.fi) and European-ancestry summary data contributed by the EGG Consortium (downloaded from www.egg-consortium.org;42). The association statistics for rs28929474 were obtained as part of a genome-wide association study (GWAS) meta-analysis; therefore, the alpha level of P < 5 × 10−8 was used to denote a significant association in the original scan. The study cohorts and meta-analysis are described below. Genetic analysis ethical statement and FinnGen ethics statement All analyses using human genetic data were conducted in accordance with the principles of the Declaration of Helsinki. Study subjects in FinnGen provided informed consent for biobank research, as mandated by the Finnish Biobank Act. Alternatively, separate research cohorts collected before the Finnish Biobank Act came into effect (September 2013) and before the initiation of FinnGen (August 2017) were collected based on study-specific consents and later transferred to Finnish biobanks following approval by Fimea (Finnish Medicines Agency) and the National Supervisory Authority for Welfare and Health (Valvira). Recruitment protocols followed biobank protocols approved by Fimea. The study protocols were approved by the Coordinating Ethics Committee of the Hospital District of Helsinki and Uusimaa (HUS; statement number HUS/990/2017) and by the Finnish Institute for Health and Welfare. Specific permit numbers are listed in Supplementary Note 1. Phenotype definitions The FinnGen Research Project (www.finngen.fi/en) integrates genome-wide genomic data with longitudinal health care information from national registries. In FinnGen, we defined preterm birth with the ICD codes O60 (ICD-10), 6440B (ICD-9), and ICD-8: 63497 (ICD-8), and full-term births as births that occurred at 39-42 weeks of gestation. We included spontaneous deliveries by excluding deliveries associated with preeclampsia, polyhydramnios, and twin or multiple gestation. In the EGG consortium summary data, preterm birth was defined as birth before 37 completed weeks of gestation or by using the ICD-10 O60 code, and controls were defined as deliveries occurring at 39–42 weeks of gestation. GWAS and meta-analysis We conducted GWAS of FinnGen data with regenie v 2.2.443 whole genome regression pipeline at FinnGen’s sandbox environment. Maternal age at the end of follow-up or death, the top 10 principal components, FinnGen chip version (1 or 2), and legacy genotyping batch were used as covariates. Logistic regression was used for SPTB (9822 cases and 186,977 controls), while linear regression was used for gestational duration (N = 67,615). Gestational duration was measured in days, and we applied a rank-based inverse normal transformation. The EGG summary data originate from a meta-analysis of European GWA studies on SPTB (15,419 cases and 217,871 controls) and gestational duration (measured in days; N = 151,987). The GWASs of the participating cohorts were run with additive models, and cohort-specific QC was applied according to EGG Consortium recommendations42. The combined effects were estimated with fixed-effects inverse-variance weighted analysis with the METAL meta-analysis tool44. We utilized METAL to conduct a meta-analysis of SPTB and gestational duration using the FinnGen and EGG data. We used fixed-effects inverse-variance weighted analysis for SPTB, and a p-value-based approach for gestational duration. The EGG data of SPTB includes samples from an earlier FinnGen release (R5); therefore, for the meta-analysis of SPTB, we excluded these samples from the FinnGen data. The sample number in the meta-analysis of SPTB was 344,628 (21,188 cases and 323,440 controls), and 219,602 in the analysis of gestational duration. Genotype analyses The SERPINA1 Pi*Z variant (rs28929474-T) genotype processing and statistical tests were conducted with PLINK45 (v1.90b7.2) using a set of unrelated DF12 individuals (available via FinnGen). Individuals were assigned to genotype groups (CC, CT, and TT), and the genotype information was linked to phenotype data (SPTB or gestational duration) using FinnGen IDs. Genotype comparisons of gestational duration between genotype groups were performed using two-sided Mann–Whitney–Wilcoxon test. Interaction analysis We performed gene-by-environment interaction (G × E) analysis using FinnGen data and generalized linear models (GLMs) in R, as follows: glm(gestational duration ~ genotype * BMI + covariates, family = gaussian). The covariates were the standard FinnGen covariates, including latent principal components 1–10, age, and the genotyping chip and batch. Animal model studies Animal work ethical declaration All procedures adhered to the Helsinki Declaration and EU Directive 86/609/EEC. Animal experiments were approved by the Finnish National Animal Experiment Board (ESAVI/28802/2022) and conducted following OULAC guidelines. Mice C57BL/6J-Serpina1em3Chmu/J knockout (KO) mice on a C57BL/6 J genetic background, lacking exon 2 from the Serpina1a–e genes (Serpina1a–e KO), were obtained from Jackson Laboratory (USA) and have been previously described by Borel et al.46. Mice were maintained under standard specific pathogen-free housing conditions with controlled environmental parameters, including standard temperature, humidity, and lighting conditions. Animals were housed in open-top cages with appropriate enrichment and had ad libitum access to food and water. Primiparous Serpina1a–e KO and wild-type (WT) females (7–15 weeks old) were time-mated with males of the opposite genotype. Vaginal plug detection marked 0.5 days post coitum (dpc). Female mice were group-housed prior to mating and maintained together until pregnancy was confirmed. We conducted two separate mice studies: Cohort 1 to study the efficacy of AAT supplementation and Cohort 2 to investigate the acute inflammatory response to LPS and supplementation. Mice cohort 1: efficacy of AAT in preventing LPS-induced SPTB At 16.5 dpc, dams were randomized by genotype and age into four treatment groups (n = 12 per group, n total = 48): 1) WT-Placebo; 2) WT-AAT; 3) KO-Placebo; 4) KO-AAT. Human alpha-1-antitrypsin (hAAT; Prolastin®, Grifols) was administered intraperitoneally (12.5 mg, 625 mg/kg). After 2.5 h, lipopolysaccharide (LPS; 1.25 µg in 100 µL) was injected i.p. to induce inflammation. Control mice received saline (NaCl) in the same volume as Prolastin®-treated dams. Dams were monitored for labor; parturition within 48 h post-LPS was defined as preterm. These dams were euthanized mid-labor for sample collection. Dams not delivering by 48 h (18.5 dpc) were euthanized, and blood was collected, and tissues were either snap-frozen or formalin-fixed. Experimental design presented in Supplementary Fig. 3A. Mice cohort 2: inflammatory response to LPS In a parallel experiment, WT and KO dams were sacrificed at 16.5 dpc (n = 3 per group, n total = 36). Groups included: mock (untreated), AAT- or placebo-treated, and AAT- or placebo + LPS-treated. AAT/placebo-only groups were euthanized 4.5 h post-injection. For AAT/placebo + LPS groups, LPS was given 2.5 h after AAT/placebo, and mice were sacrificed 2 h later (i.e., 4.5 h total post-AAT/placebo). LPS-only mice were sacrificed 2 h post-injection. Inflammatory response was analysed using serum cytokine multiplex assay as described below. Experimental design presented in Supplementary Fig. 3B. Pup assessment Pup’s biological and developmental maturity was assessed by measuring crown–rump (CR) length. Viability of vaginally delivered pups was determined based on oxygenation (skin color, diaphragm movement), touch responsiveness, vocalization, and overall activity. Pups meeting all criteria were classified as viable. For in utero pups, viability was assessed based on placental attachment, umbilical cord blood engorgement, skin color, and response to touch through the uterine wall. All viable pups, whether live-born or in utero, were sacrificed by decapitation. Blood was collected and pooled per litter for downstream analysis. Placentas and amniotic sacs were either fixed whole in neutral formalin or snap-frozen in liquid nitrogen. hAAT concentration and half-life in mice serum Serum samples from dams and pups were analysed for hAAT levels (g/L) at NordLab, University Hospital Oulu (OYS), using an immunoturbidimetric assay (cut-off: 0.17–4 g/L). Each run included internal positive (human serum) and negative (untreated WT mouse serum) controls. AAT half-life (t1/2) calculated using the first-order elimination model, based on the exponential decay equation: Where: C0 = Initial concentration at the first timepoint Ct = Concentration at the second timepoint t = Time difference between the two concentrations k = Elimination rate constant. Calculated: After calculating k, the half-life is calculated using the following formula: Calculations of half-life were initiated at 6 h, after serum concentrations had reached the elimination phase. Half-lives were determined for separate intervals to account for potential deviations from first-order kinetics, with an overall half-life (6–48 h) reported for reference. Serum cytokine multiplexing To assess inflammatory activity, we measured 45 cytokines in serum from cohort 2 dams and pups using the Mouse XL Cytokine Luminex® Performance 45-plex Fixed Panel (R&D Systems, LKTM015). Assays were performed according to the manufacturer’s instructions. Data were acquired on a Luminex MAGPIX system and analysed with xPONENT software. The WT-mock group served as the reference. Maternal plasma cytokine concentrations were visualized as a heatmap to compare expression across genotypes and treatments. Samples were grouped by condition (e.g., WT-MOCK, KO-LPS), and values were displayed on a diverging color scale (purple = low, white = mid, coral-red = high) to highlight shifts associated with genotype and inflammatory status. Histological staining and image analysis Neutral-buffered formalin-fixed tissues were embedded in paraffin and sectioned at 5 µm. Sections were deparaffinized and rehydrated before histological staining. Antigen retrieval was performed using Antigen Unmasking Solution, Tris-Based (Vector Laboratories, H-3301-250), diluted according to the manufacturer’s instructions. Immunohistochemical (IHC) staining was carried out using the ImmPRESS® HRP Horse Anti-Mouse IgG Polymer Detection Kit, Peroxidase (Vector Laboratories, MP-7402), with primary antibodies targeting either human alpha1-antitrypsin (1:1000, Sigma-Aldrich, SAB2109236) or murine alpha1-antitrypsin (1:1000, Invitrogen, PA5-79974). For each sample, negative controls omitting the primary antibody were included. Primary antibodies were incubated for 1 h at room temperature (RT), and all further steps were conducted per the detection kit protocol. Staining positivity and antibody localization were assessed based on the intensity of brown diaminobenzidine (DAB) coloration. Image resolution was preserved at acquisition, with brightness and sharpness enhanced post-acquisition for improved visualization. Quantification of staining was conducted using Fiji (ImageJ) software. DAB-stained and hematoxylin-counterstained images were deconvoluted, and the DAB signal was separated for analysis. Images were converted to 8-bit format, and thresholding was applied to calculate the percentage of DAB-positive area relative to total tissue area. Western blot Western blotting was performed on placental tissues collected at 17.5 and 18.5 dpc (cohort 1) to assess AAT isoforms and glycosylation status. Primary antibodies targeting human and murine AAT were used at a 1:2500 dilution. A single secondary antibody (goat anti-rabbit IgG Dylight 680 conjugate, Rockland 611-144-002-0.5) was applied to all samples at 1:10,000 dilution. Blots were visualized using the Odyssey Infrared Imaging System (LI-COR Biosciences). GAPDH was used as a loading control. Membranes were stripped using Restore™ PLUS Western Blot Stripping Buffer (Thermo Fisher, #46430) and re-probed with anti-GAPDH antibody (1:1000, Cell Signaling Technologies, #2118). Molecular weight markers were assigned according to the lot-specific calibration provided for the Thermo Scientific Spectra™ Multicolor Broad Range Protein Ladder (#26634), calibrated for 4–20% Tris-glycine SDS-PAGE gels. Uncropped and unprocessed western blot scans are provided in the supplements (Supplementary Fig. 13). RNA extraction and RT-qPCR Total RNA was isolated from liver and placental tissues for Serpina1a–e transcript analysis by RT-qPCR. Samples were obtained from time-mated KO and WT dams from both cohort 1 and -2. Eef2 was used as the reference gene. RNA was isolated using the RNeasy Mini Kit (Qiagen, 74104) following the manufacturer’s protocol. RNA quantity and purity were assessed before cDNA synthesis, which was carried out using the Transcriptor First Strand cDNA Synthesis Kit (Roche, # 4897030001). RT-qPCR was performed on a LightCycler 96 system (Roche). Primers and probes were obtained from Merck. Probes were labeled with FAM at the 5’ end and BHQ-1 at the 3’ end. To ensure amplification, primer and probe sequences were designed to target regions outside the knocked-out exon 2. Primer and probe sequences are as in Supplementary Data 2. Mice data statistics and reproducibility All data manipulation and statistical analyses were performed in R47 (v4.4.2). Data wrangling and transformation used the tidyverse48 (v2.0.0). Normality was tested with the Shapiro–Wilk test. Due to non-normal distributions, group differences were assessed using the Kruskal–Wallis test and Dunn’s post hoc test from FSA49 (v0.10.0), with Benjamini–Hochberg (BH) correction. Where applicable, pairwise binary comparisons employed Wilcoxon rank-sum tests, also BH-adjusted. Categorical variables (e.g., pup survival) were analysed using Chi-square or Fisher’s exact tests, with BH-adjusted pairwise comparisons. Survival analyses and Kaplan–Meier curves generated with survival50 (v3.8-3) and survminer51 (v0.5.0). Global P-values were BH-adjusted. Cytokine concentrations (pg/mL) were log₂-transformed after removing missing values and averaged across three replicates per condition. Heatmaps with hierarchical clustering were created using pheatmap52 (v1.0.13), with colors ranging from dark purple (low) to dark coral (high). IHC staining was repeated multiple times with consistent staining patterns. Representative images from different regions of a single placenta are presented. Western blot analyses were likewise repeated multiple times with consistent results. Western blot lysates were prepared from different placentas collected from the same dam (same litter), and representative blots are shown. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Reporting Summary Transparent Peer Review file Source data Source Data Acknowledgments We would like to acknowledge Grifols for their generous sponsorship of this project through their research-sponsored program. We want to acknowledge the participants and investigators of the FinnGen study. Contributors of FinnGen are listed in Supplementary Data 1. The FinnGen project is funded by two grants from Business Finland (HUS 4685/31/2016 and UH 4386/31/2016) and the following industry partners: AbbVie Inc., AstraZeneca UK Ltd, Biogen MA Inc., Bristol Myers Squibb Inc. (and Celgene Corporation & Celgene International II Sàrl), Genentech Inc., Merck Sharp & Dohme LCC, Pfizer Inc., GlaxoSmithKline Intellectual Property Development Ltd., Sanofi US Services Inc., Maze Therapeutics Inc., Johnson&Johnson Innovative Medicine Inc., Novartis AG, Boehringer Ingelheim International GmbH and Bayer AG. Following biobanks are acknowledged for delivering biobank samples to FinnGen: Auria Biobank (www.auria.fi/biopankki), THL Biobank (www.thl.fi/biobank), Helsinki Biobank (www.helsinginbiopankki.fi), Biobank Borealis of Northern Finland (https://www.ppshp.fi/Tutkimus-ja-opetus/Biopankki/Pages/Biobank-Borealis-briefly-in-English.aspx), Finnish Clinical Biobank Tampere (www.tays.fi/en-US/Research_and_development/Finnish_Clinical_Biobank_Tampere), Biobank of Eastern Finland (www.ita-suomenbiopankki.fi/en), Central Finland Biobank (www.ksshp.fi/fi-FI/Potilaalle/Biopankki), Finnish Red Cross Blood Service Biobank (www.veripalvelu.fi/verenluovutus/biopankkitoiminta), Terveystalo Biobank (www.terveystalo.com/fi/Yritystietoa/Terveystalo-Biopankki/Biopankki/) and Arctic Biobank (https://www.oulu.fi/en/university/faculties-and-units/faculty-medicine/northern-finland-birth-cohorts-and-arctic-biobank). All Finnish Biobanks are members of BBMRI.fi infrastructure (https://www.bbmri-eric.eu/national-nodes/finland/). Finnish Biobank Cooperative -FINBB (https://finbb.fi/) is the coordinator of BBMRI-ERIC operations in Finland. The Finnish biobank data can be accessed through the Fingenious® services (https://site.fingenious.fi/en/) managed by FINBB. We thank Maarit Haarala (University of Oulu, Oulu, Finland) for technical assistance. We gratefully acknowledge the staff of the Oulu Laboratory Animal Center (OULAC, Oulu, Finland) for their excellent care of the animals and their support throughout the in vivo experiments, as well as Hanna-Marja Voipio and Sakari Laaksonen for their advisory role during the license application. In this context, we would also like to acknowledge the Biocenter OULU (BCO) Transgenic and Tissue Phenotyping Core Facility (BCO-TTP) for their assistance in establishing our transgenic mouse lineage. Furthermore, CSC–IT Center for Science, Finland, is acknowledged for computational resources. Author contributions E.K. contributed to conceptualization, methodology, investigation, data curation, formal analysis, and writing – original draft, review and editing. A.P. contributed to conceptualization, methodology, formal analysis and writing – original draft, review and editing. A.M.H. contributed to conceptualization, supervision, funding acquisition, project administration and writing – review and editing. H.T. contributed to formal analysis and writing – review and editing. M.R. contributed to conceptualization, supervision, and writing – review and editing. M.H. contributed to conceptualization, resources, supervision, funding acquisition, and writing – review and editing. Peer review Peer review information Nature Communications thanks Ronald Reilkoff, Pol Solé-Navais and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Funding This project was funded through a research-sponsored program by Grifols. In addition, the study was supported by research grants from the Sigrid Juselius Foundation (M.H, A.M.H), the Foundation for Pediatric Research (M.R, A.M.H), the Suorsan Terveydenhoitosäätiö sr. Foundation (A.M.H), the Finnish Cultural Foundation (H.T), the Päivikki and Sakari Sohlberg Foundation (H.T), the Stiftelsen Alma och K. A. Snellman Foundation (A.P, H.T), the Yrjö Jahnsson Foundation (A.P), and the Medical Research Center (MRC) Oulu (H.T). Data availability Data from the mouse experiments are provided in the Supplementary Information (Source Data). Source data are provided with this paper. Competing interests This study was supported by a research-sponsored program from Grifols. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: M. Rämet, M. Hallman. Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76643-9. References Associated Data Supplementary Materials Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Reporting Summary Transparent Peer Review file Source Data Data Availability Statement Data from the mouse experiments are provided in the Supplementary Information (Source Data). Source data are provided with this paper.