Failure of postnatal ductus arteriosus closure in prostaglandin transporter-deficient mice

Abstract

Background: Prostaglandin E(2) (PGE(2)) plays a major role both in maintaining patency of the fetal ductus arteriosus and in closure of the ductus arteriosus after birth. The rate-limiting step in PGE(2) signal termination is PGE(2) uptake by the transporter PGT.

Methods and results: To determine the role of PGT in ductus arteriosus closure, we used a gene-targeting strategy to produce mice in which PGT exon 1 was flanked by loxP sites. Successful targeting was obtained because neither mice hypomorphic at the PGT allele (PGT Neo/Neo) nor global PGT knockout mice (PGT(-/-)) exhibited PGT protein expression; moreover, embryonic fibroblasts isolated from targeted mice failed to exhibit carrier-mediated PGE(2) uptake. Although born in a normal mendelian ratio, no PGT(-/-) mice survived past postnatal day 1, and no PGT Neo/Neo mice survived past postnatal day 2. Necropsy revealed patent ductus arteriosus with normal intimal thickening but dilated cardiac chambers. Both PGT Neo/Neo and PGT(-/-) mice could be rescued through the postnatal period by giving the mother indomethacin before birth. Rescued mice grew normally and had no abnormalities by gross and microscopic postmortem analyses. In accordance with the known role of PGT in metabolizing PGE(2), rescued adult PGT(-/-) mice had lower plasma PGE(2) metabolite levels and higher urinary PGE(2) excretion rates than wild-type mice.

Conclusions: PGT plays a critical role in closure of the ductus arteriosus after birth by ensuring a reduction in local and/or circulating PGE(2) concentrations.

Figure 1. Targeting strategy for knocking out the mouse PGT gene

(a) Strategy used for PGT gene targeting. Top line: endogenous locus; 2nd line: targeting vector; 3rd line: targeted locus after excision of Neo gene cassette with FLPe; Bottom line: targeted locus after excision of Exon 1 at LoxP sites with Cre recombinase. A targeting vector containing a 13 kb mouse genomic DNA segment was constructed with 3 LoxP insertions, two of them flanking a Neo gene insertion, which also includes two FLPe recombinase sites. The locations of PCR primers (AA′ or BB′) are shown by arrows. P signifies the hybridization probe for Southern blots. (b) Genotyping of wild type (+/+), heterozygote (+/−), and global KO mice (−/−). PCR products from the intact PGT Exon 1 gene and from the gene lacking Exon 1 generated 2.8 kb and 0.6 kb fragment, respectively (Primers AA′). Because of competition in the PCR reaction, both products could not be visualized in DNA from heterozygotes. Therefore, a second PCR reaction (BB′: 1.0 kb) was used to demonstrate the wild type allele (bottom gel). Global KO mice show only 0.6 kb. (c) A restriction enzyme HpaI as used for Southern blot hybridization. A 9.8 kb band in the wild type allele was replaced by an 7.9 kb band in the targeted allele.

Figure 2. Validation of PGT targeting by loss of protein expression and transport function

(a) Lung tissue from PGT Neo/Neo mice shows absence of PGT protein, whereas PGT +/+ mice express PGT (brown reaction product) in type II cells (see Supplemental Data Figure 1S for birefringence of granules of these cells in our PGT +/+ mice). (b) PGE₂ uptake by mouse embryonic fibroblasts (MEFs) derived from PGT wild type (+/+) versus PGT null (−/−) mice. ³H-PGE₂ uptake was determined alone (this represents PGT-mediated uptake plus simple diffusion) and also in the presence of excess unlabeled PGE₂ (this blocks PGT-mediated tracer uptake and reveals uptake due to simple diffusion alone). For each set of paired uptake data, we calculated the percent increase in PGE₂ uptake attributable to PGT: (uptake_tracer) ÷ (uptake_{tracer + unlabeled}) × 100. MEFs from PGT +/+ mice (n = 7) had a PGT-mediated change in uptake of +30.3% ± 8.38% (n = 7), whereas MEFs derived from PGT −/− animals had a PGT-mediated change in uptake of −1.29% ± 3.38% (n = 7). The difference was highly statistically significant (p = 0.008 by unpaired t-test; p = 0.007 by Wilcoxon two-sample test).

Figure 3. Patent DA in PGT Neo/Neo and PGT −/− mice

H&E stain of paraffin-embedded sections. (a) and (d). Low- and high-power view, respectively, of a cross-section from the torso of a PGT +/+ (wild type) mouse (representative, n = 3) eleven hours after birth. The DA has closed normally (a, arrow), and there is a normal intimal thickening (d, arrow) consisting of a loose network of cells filling and obliterating the constricted lumen. Tr, trachea; VB, vertebra; AAo, ascending aorta; DAo, descending aorta. (b) and (e). Torso of PGT Neo/Neo mouse (representative, n = 5) dying on post-natal day 2 shows PDA. An arrow marks the connection between the DAo and DA. High power view (e) reveals normal intimal thickening (arrow). (c) and (f). Torso of PGT −/− mouse (representative, n = 5) similarly shows patent DA. The pulmonary artery (PA) has dilated with reversed blood flow, and blood also fills the DAo and PA. High power (f) view reveals normal intimal thickening (arrow).

Figure 4. Endothelium and internal elastic lamina of the DA at embryonic day E19 appear normal in PGT targeted mice

(a, c, e) PGT heterozygote (+/−) examined at embryonic day E19 shows the expected patent DA in continuity with the descending aorta (DAo) (a) with a normal-appearing endothelium in H&E staining (arrow, c) and normal elastin staining of the internal elastic lamina (arrow, e). (b, d, f) PGT null mouse (PGT −/−) shows the same pattern as the PGT+/− mouse.

Figure 5. Hearts of PGT targeted mice show chamber dilatation with normal interventricular septum thickness

(a) Wild type mice (n = 3) have normal dimensions of the cardiac chambers (solid line). The interventricular septum, an indicator of intrinsic myocardial muscle development, is also normal (arrow). (b) PGT Neo/Neo mouse (n = 3) shows dilated right and left ventricular chambers (lines), but the interventricular septum, an indicator of intrinsic heart muscle development, is normal (arrow). (c) PGT −/− mouse (n = 3) shows finding similar to (b).

Figure 6. Immunolabeling for PGT shows strong expression in the normal DA compared to the kidney

(a) DA of wild type mouse (n = 3) on post-natal day 1 showing strong labeling for PGT in smooth muscle cells of the myointimal thickening (brown reaction product, arrow). (b) DA of PGT Neo/Neo mouse (n = 3) on post-natal day 1 with negative PGT labeling (arrow). (c and d) Labeling of mouse renal cortical collecting duct (c, arrow) and mouse renal papillary vasa recta endothelium (d, arrow) as positive controls (n = 10).

Figure 7

Patent ductus arteriosus resulting from perturbations of PGE₂ signaling. Genetic interruption of both COX enzymes, or genetic interruption of the EP₄ receptor, of prostaglandin dehydrogenase (PGDH), or of PGT (present work) all result in PDA in the mouse. Further detail and references are given in the Discussion.