INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IN HUMAN PLACENTAL ARTERIES and veins from women with gestational diabetes and those with normal pregnancies, reoxygenation after a period of hypoxia induces transient contractions that are significantly larger in vessels from placentas affected by gestational diabetes (6). This contraction is mediated by cyclooxygenase products that do not seem to originate from the endothelium in normal and gestational diabetic placental vessels, since it is eliminated by indomethacin and not altered by removal of the endothelium. Micromolar concentrations of H₂O₂ also produce an endothelium-independent prostaglandin-mediated contraction that is larger in placental vessels from women with gestational diabetes (6). We previously hypothesized that reoxygenation after exposure to hypoxia results in a prostaglandin-mediated vascular contraction through a transient generation of H₂O₂ (6, 17).

We have also shown that lactate causes relaxation in human placental vessels of normal pregnancies by an endothelium-independent, O₂-dependent process that appears to be mediated by the generation of H₂O₂ and stimulation of cGMP production (19). Bovine pulmonary arteries relax in an endothelium-independent manner when exposed to lactate and H₂O₂ through a mechanism that is thought to involve the stimulation of soluble guanylate cyclase by an increase in the metabolism of H₂O₂ by catalase (1, 20). These responses to lactate and H₂O₂ are inhibited by prior exposure of the pulmonary arteries to an elevated physiological level of nitric oxide, through a mechanism that appears to be dependent on the formation of peroxynitrite (13, 24). In these studies it was demonstrated that a 2-min exposure to 50 nM nitric oxide caused a prolonged inhibition of peroxide metabolism by catalase (13, 24), a process that appears to be an essential component of the mechanism through which H₂O₂ causes cGMP-mediated pulmonary arterial relaxation (1). Although the response of human placental arteries and veins to H₂O₂ is dominated by a somewhat transient prostaglandin-mediated contraction (6, 17), a prolonged relaxation response to H₂O₂ is observed in placental veins in the presence of an inhibitor of prostaglandin synthesis (19). Because lactate seems to activate only the relaxation component of the response to H₂O₂, lactate may cause H₂O₂ generation in placental vessels at a cellular site that is accessible to the stimulation of guanylate cyclase and not to the activation of arachidonic acid release and its metabolism by cyclooxygenase to contractile prostaglandins (19). The relaxation to lactate is altered in placental vessels from women with severe preeclampsia (5). There is evidence for an increase in nitric oxide production (11) and peroxynitrite formation (15) in the placental circulation of patients with preeclampsia. Because diabetic pregnancies may have pathophysiological mechanisms similar to preeclamptic pregnancies (23), changes in the metabolism of nitric oxide could be a contributing factor to altered placental vascular H₂O₂-associated responses in preeclampsia and diabetes.

In our present study we investigated whether the enhanced contractile response to exogenous H₂O₂ previously observed (6) in placental vessels from women with gestational diabetes was related to a selective alteration in the relaxation response of these vessels to H₂O₂. We compared responses to relaxing agents observed in placental vessels of women with normal pregnancies with responses in similar vessels of women with gestational diabetes to identify vasodilator mechanisms that are altered. Lactate was used to cause responses through the endogenous generation of H₂O₂, and exogenous H₂O₂ was studied in the presence of the cyclooxygenase inhibitor indomethacin. Also, we examined whether rates of tissue H₂O₂ metabolism by catalase and the production of a decomposition product of nitric oxide (nitrite) are altered in placental vessels from women with gestational diabetes.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient characteristics. Placentas obtained as pathological discards from 12 women with gestational diabetes and 47 women with uncomplicated full-term pregnancies were selected for the vascular reactivity experiments reported in this study. The diabetic women were identified by the criteria of Carpenter and Coustan (2) for the glucose tolerance test and classified as having type III (gestational carbohydrate intolerance or gestational) diabetes according to Hollingsworth's (9) adaptation from the National Diabetes Data Group (16). Glycosylated Hb levels were obtained from 10 of the diabetic women, and at the time of delivery, 9 women demonstrated good glycemic control on the basis of a glycosylated Hb level of <8%. The one exception was a woman who had a glycosylated Hb of 8.8%. The mean glycosylated Hb for this group was 7.2% (range 6-8.8%). The women in the uncomplicated group were given 50-g oral glucose challenge tests at 24-28 wk of gestation. The test results were considered negative if plasma glucose values were <140 mg/dl at 1 h after the glucose load. The patient characteristics are summarized in Table 1.

Experimental techniques. Placentas from the women with gestational diabetes and the women with uncomplicated full-term pregnancies were obtained immediately after vaginal or cesarean delivery and placed in ice-cold saline solution. Placental arteries and veins were then dissected from the periphery of the fetal side of the chorionic plate within 1 h of delivery and placed in ice-cold Krebs bicarbonate buffer (pH 7.4) containing (in mM) 118 sodium chloride, 4.7 potassium chloride, 1.5 calcium chloride, 25 sodium bicarbonate, 1.1 magnesium sulfate, 1.2 monobasic potassium phosphate, and 5.6 glucose. Endothelium-intact rings, 1-2 mm diameter and 2 mm wide, were prepared and mounted on wire hooks attached to force displacement transducers (type FT03, Grass Instruments, Quincy, MA) for measurement of changes in isometric force. Vessels were incubated in 10-ml baths (Metro Scientific, Farmingdale, NY) in Krebs buffer gassed with 5% O₂-5% CO₂-90% N₂ at 37°C. During the equilibration period, tension was continuously adjusted to 5 g, which we found to be the optimal baseline force for observation of maximal contraction to 25 mM potassium in all groups. Stretching the vessels to this level of baseline force produced ~1 g of spontaneous force generation on the basis of the maximal relaxation elicited by exposure to nitroglycerin. Changes in force were recorded on a polygraph (model 7, Grass Instruments) and are reported as the increase in force above the 5 g of combined passive and stretch-induced force. After a 2-h equilibration at the optimal passive tension, the vessels were depolarized with Krebs bicarbonate solution containing potassium chloride (123 mM) in place of sodium chloride. This treatment produces maximal contraction of ~7 g and enhances the reproducibility of subsequent contractions. The vessels were then reequilibrated with Krebs bicarbonate solution for 30 min before experiments were conducted.

Exposure of placental vessels to lactate and other vascular relaxants. Placental arteries and veins were initially precontracted with 1 µM prostaglandin F₂ (PGF₂). Arteries or veins that did not generate 0.5 g of force were given a second dose of 1 µM PGF₂, which adjusted the average of the observed level of force generation to >1.5 g in all groups. Once the contraction achieved a steady-state level, increasing cumulative doses of lactate (1-10 mM) were added to the organ baths, with 5 min allowed between doses. Lactate solution was made in distilled deionized water, and pH was adjusted with 1 M sodium hydroxide to 7.4 to keep a constant pH in the bath and to eliminate the possibility of any responses that might be produced by the change in pH. Lactate was added to the 10-ml baths in 10- to 100-µl aliquots to produce final concentrations of 1-10 mM. Experiments examining the relaxation to nitroglycerin [a stimulator of soluble guanylate cyclase via nitric oxide generation and cGMP-mediated relaxation (10)], forskolin [a stimulator of adenylate cyclase and cAMP-mediated relaxation (21)], and arachidonic acid [which appears to cause relaxation of placental vessels through a mechanism that may involve its metabolism by a cytochrome P-450-mediated pathway (18)] used a protocol similar to that described for lactate in different vessel rings from the same placentas. Nitroglycerin was added in cumulative doses of 1 nM-1 µM. Forskolin and arachidonic acid were added in cumulative doses of 0.01-10 µM. Forskolin was dissolved in ethyl alcohol to a final concentration of 10 mM before dilution with distilled deionized water. Dilutions of PGF₂ and arachidonic acid were made in distilled deionized water. In all the above studies the relaxation is expressed as percentage of the PGF₂-induced contraction. Preliminary experiments indicated that the response to all relaxing agents was independent of the level of submaximal force when evaluated as percent relaxation. Data for some responses described for vessels from normal patients were derived from previously reported studies (5, 18, 19).

Exposure of placental vessels to H₂O₂. Placental arteries and veins were initially precontracted as described above. The vessels were treated with 10 µM indomethacin for 15 min. Cumulative doses of H₂O₂ (1 µM-1 mM) were added to the organ baths when a steady-state level of tone was reached. A 2-min period was allowed between doses. Dilutions of H₂O₂ were made in distilled deionized water, and 10-µl aliquots were added to the 10-ml baths as indicated. Indomethacin was dissolved in absolute ethyl alcohol to produce a 10 mM solution, and 10 µl from the solution were added to the 10-ml bath. Data for responses described for veins from normal patients were derived from a previously reported study (19).

Measurement of the rates of metabolism of endogenously formed H₂O₂ by catalase in placental vessels. The rates of oxidation of methanol to formaldehyde by human placental arteries and veins were employed to quantitate intracellular H₂O₂ metabolism by catalase, as previously described (13). The formaldehyde generated was measured by the Nash reaction with acetylacetone and ammonium acetate, and the resulting 3,5-diacetyl-1,4-dihydrolutidine product was quantified by its absorbance at 412 nm with use of an extinction coefficient of 17.8 cm¹ · mM¹. Human placental vessels were incubated in the tissue baths employed for tone studies in Krebs bicarbonate buffer (~150 mg placental vessel/ml Krebs buffer) in the presence or absence (for blanks) of 0.5 M methanol, a concentration that previously was found to optimize the sensitivity of the assay. Aliquots of 0.5 ml were removed immediately after the addition of methanol (for a time 0 point) and 5 min later for assay of the amount of formaldehyde formed. To these aliquots, 0.5 ml of Nash reagent (150 g ammonium acetate and 2 ml acetylacetone per liter, adjusted to pH 6.0 with acetic acid) was added. As previously reported (13), the release of formaldehyde was found to be linear with time for ~10 min, and incubation of arteries for 10 min with known concentrations of formaldehyde that were typically produced under the experimental conditions resulted in an ~85% recovery of the added formaldehyde.

Measurement of rates of basal nitric oxide production in placental vessels. Nitric oxide was measured by quantitation of the amount of its stable decomposition product nitrite (25) that accumulated from endothelium-intact placental vessels that were incubated under conditions described for studies on vascular reactivity for periods of 10 min. As previously described (25), the diazotinization of sulfanilic acid colorimetric assay for nitrite [in the presence of N-(1-naphthyl)-ethylenediamine at acidic pH] was used to detect the formation of nitrite. The assay is standardized with known amounts of sodium nitrite under the conditions in which endogenous production of nitrite was measured. Data are evaluated as nanomoles per minute per gram blotted weight of tissue. The vessels employed for measurement of nitrite production were derived from a larger patient group, which had characteristics similar to those used in the studies on vascular reactivity.

Materials. Acetylacetone, ammonium acetate, arachidonic acid, forskolin, H₂O₂, lactic acid, N-(1-naphthyl)-ethylenediamine, sodium nitrite, and sulfanilic acid were purchased from Sigma Chemical (St. Louis, MO). Nitroglycerin solutions were prepared by dissolving 0.4-mg sublingual tablets (Parke-Davis) in distilled water. PGF₂ was purchased from Cayman Chemical (Ann Arbor, MI). Stock solutions of arachidonic acid (10 mM) were prepared in equimolar sodium carbonate and handled under nitrogen atmosphere until use.

Statistical analysis. A nonpaired Student's t-test was used to compare responses between two groups. The accepted level of significance was P < 0.05. The number of experimental determinations (n) in all cases is equal to the number of placentas from which a vessel ring was used for a study or control group. Only one artery or vein from a single placenta was included in each analysis. Values are means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of gestational diabetes on the relaxation of placental vessels to lactate. As shown in Fig. 1, the addition of 1, 5, and 10 mM lactate to placental arteries of pregnancies with gestational diabetes, precontracted with 1 µM PGF₂, caused relaxation responses that were 85, 40, and 54%, respectively, less than (P < 0.05) the relaxation of arteries derived from normal patients to these same doses of lactate. Placental veins of pregnancies with gestational diabetes showed relaxation responses when exposed to 1, 5, and 10 mM lactate that were reduced (P < 0.05) by 80, 71, and 66%, respectively, compared with the response of normal veins treated with these same doses of lactate (Fig. 1).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Summary data showing effect of 1-10 mM lactate on endothelium-intact human placental arteries (A) and veins (B) from normal [arteries (n = 8) and veins (n = 15)] and gestational diabetic [arteries (n = 12) and veins (n = 11)] pregnancies. GDM, gestational diabetes mellitus. Data from normal vessels were previously published (19). * P < 0.05 vs. response in vessels from normal patients.

Effect of gestational diabetes on the relaxation of placental vessels to arachidonic acid. As shown in Fig. 2, there was no significant difference in the observed relaxation to arachidonic acid at any of the doses (10 nM-10 µM) examined between placental vessels of normal pregnancies and placental vessels of pregnancies with gestational diabetes.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Summary data showing effect of 0.01-10 µM arachidonic acid on endothelium-intact human placental arteries (A) and veins (B) from normal [arteries (n = 13) and veins (n = 12)] and gestational diabetic [arteries (n = 8) and veins (n = 6)] pregnancies. [AA], arachidonic acid concentration. Data from normal vessels were previously published (18).

Effect of gestational diabetes on the relaxation of placental vessels to forskolin. As shown in Fig. 3, placental arteries and veins of pregnancies with gestational diabetes relaxed in a manner similar to the vessels derived from normal pregnancies after the addition of 0.01-10 µM forskolin.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Summary data showing effect of 0.01-10 µM forskolin on endothelium-intact human placental arteries (A) and veins (B) from normal [arteries (n = 7) and veins (n = 9)] and gestational diabetic [arteries (n = 9) and veins (n = 8)] pregnancies. [Forsk], forskolin concentration. Data from normal vessels were previously published (5).

Effect of gestational diabetes on the relaxation of placental vessels to nitroglycerin. As shown in Fig. 4, 1 nM-1 µM nitroglycerin caused a concentration-dependent relaxation of arteries and veins from normal pregnancies that was similar to the relaxation response to nitroglycerin observed in arteries and veins derived from pregnancies with gestational diabetes.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Summary data showing effect of 1 nM-1 µM nitroglycerin on endothelium-intact human placental arteries (A) and veins (B) from normal [arteries (n = 12) and veins (n = 7)] and gestational diabetic [arteries (n = 4) and veins (n = 6)] pregnancies. [NTG], nitroglycerin concentration. Data from normal vessels were previously published (5).

Effect of gestational diabetes on the relaxation of placental vessels to H₂O₂ in the presence of indomethacin. In the presence of indomethacin, 1 µM-1 mM H₂O₂ caused only a minimal relaxation of placental arteries from normal and diabetic pregnancies (Fig. 5). The treatment of placental veins from normal pregnancies with H₂O₂ in the presence of indomethacin caused a prominent relaxation at doses >10 µM. As shown in Fig. 5, the addition of 0.1 and 1 mM H₂O₂ to placental veins of pregnancies with gestational diabetes, precontracted with 1 µM PGF₂, caused relaxation responses that were 80 and 94% less than (P < 0.05) the relaxation of veins derived from normal patients to these same doses of H₂O₂. Thus the relaxation to H₂O₂ was essentially eliminated (P <0.01) in veins from placentas of women with gestational diabetes.

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Summary data showing effect of 1 µM-1 mM H2O2 in presence of 10 µM indomethacin on endothelium-intact human placental arteries (A) and veins (B) from normal [arteries (n = 6) and veins (n = 5)] and gestational diabetic [arteries (n = 6) and veins (n = 5)] pregnancies. [H2O2], H2O2 concentration. * P < 0.05 vs. response in vessels from normal patients.

Effect of gestational diabetes on the rates of metabolism of endogenously formed H₂O₂ by catalase in placental vessels. As shown in Fig. 6A, arteries and veins derived from pregnancies with gestational diabetes showed a rate of metabolism of endogenously formed H₂O₂ by catalase that was 61 and 44% less than the rates of metabolism by catalase observed in placental arteries and veins obtained from normal patients.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   A: summary data showing amount of conversion of 0.5 M methanol to formaldehyde in endothelium-intact human placental arteries and veins from normal [arteries (n = 12) and veins (n = 10)] and gestational diabetic [arteries (n = 5) and veins (n = 5)] pregnancies. This measurement quantifies amount of endogenous H2O2 metabolism by catalase. * P < 0.05 vs. response in vessels from normal patients. B: summary data showing level of endogenous nitrite production by endothelium-intact human placental arteries and veins from normal [arteries (n = 67) and veins (n = 62)] and gestational diabetic [arteries (n = 33) and veins (n = 33)] pregnancies. * P < 0.05 vs. response in vessels from normal patients.

Effect of gestational diabetes on nitrite production by placental vessels. As shown in Fig. 6B, arteries and veins derived from pregnancies with gestational diabetes showed a rate of nitrite production that was 195 and 113% greater than the level of nitrite production observed in placental arteries and veins obtained from normal patients. An inhibitor of the biosynthesis of nitric oxide from L-arginine, nitro-L-arginine (0.1 µM), markedly inhibited the formation of nitrite in each experimental group. The rate of nitrite production was significantly reduced (P < 0.05) in the presence of 0.1 mM nitro-L-arginine by 66% to 0.051 ± 0.007 nmol · min¹ · g¹ (n = 18) and by 67% to 0.147 ± 0.040 nmol · min¹ · g¹ (n = 9) in placental arteries from normal and GDM patients, respectively, and by 55% to 0.058 ± 0.013 nmol · min¹ · g¹ (n = 19) and by 47% to 0.129 ± 0.015 nmol · min¹ · g¹ (n = 10) in veins from normal and GDM patients, respectively.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Human placental arteries and veins from women with gestational diabetes exhibited a relaxation to lactate that was significantly less than the response seen in vessels from uncomplicated full-term pregnancies. A cGMP-associated relaxation to H₂O₂ observed in the presence of indomethacin (19) was eliminated in placental veins of women with gestational diabetes compared with normal veins. An observation in placental arteries that is difficult to explain is that H₂O₂ causes a minimal prostaglandin-independent relaxation. This observation is somewhat surprising, because gestational diabetes enhances the prostaglandin-mediated contractile response to H₂O₂ (6) and attenuates the relaxation to lactate in placental arteries (Fig. 1). In contrast, responses to other relaxing agents, such as arachidonic acid, forskolin, and nitroglycerin, were similar in placental arteries and veins from normal and gestational diabetic pregnancies. The hypothesized mechanism through which lactate and H₂O₂ elicit responses in human placental arteries and veins, shown in Fig. 7, suggests that a defect in the mechanism of relaxation to H₂O₂ may be responsible for the altered responses seen in vessels from patients with gestational diabetes described in the present study and in a previous investigation that identified an enhanced prostaglandin-mediated contractile response to H₂O₂ (6).

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Hypothesis for origin of alterations in vascular responses to posthypoxic reoxygenation, H2O2, and lactate (Lac) observed in placental vessels derived from GDM patients. In these placental vessels, cyclooxygenase (COX)-derived prostaglandin (PG)-mediated contraction to posthypoxic reoxygenation and to H2O2 is presumably increased as a result of an attenuation of catalase (Cat) activity, which appears to be essential for simultaneously activated relaxation response to H2O2 that involves an increase in cGMP as a result of stimulation of soluble form of guanylate cyclase. Relaxation response to Lac is thought to involve generation of cytosolic NADH formed during its metabolism to pyruvate (Pyr) and an increase in H2O2 production by NADH oxidase. Inhibition of Cat in GDM patients is likely to also mediate attenuation of relaxation to Lac. Because placental vessels from GDM patients show evidence of increased nitric oxide (NO) production and because exposure of Cat to NO causes a prolonged inhibition of its activity, inhibition of Cat by NO may participate in alterations in responses observed in placental vessels from GDM patients.

The observation that gestational diabetes did not influence the responses to several of the vasodilators examined suggests that only certain mechanisms are altered in the placental vessels. Nitroglycerin, a stimulator of soluble guanylate cyclase, is thought to produce a nitric oxide-elicited cGMP-mediated relaxation (10). Forskolin, a well-established direct stimulator of adenylate cyclase (21), is thought to function through activation of a cAMP-mediated relaxation. Human placental vessels possess an endothelium-independent mechanism of relaxation to arachidonic acid that seems to be mediated by metabolites formed by a cytochrome P-450-dependent metabolic pathway (18). Although the actual mechanism through which arachidonic acid produces relaxation in placental vessels is not known, endothelium-derived cytochrome P-450 metabolites of arachidonic acid seem to produce relaxation as a result of opening potassium channels, which causes hyperpolarization (7). Thus the various components of the signal transduction mechanisms for each of these three relaxing agents do not seem to be altered in placental vessels from pregnancies with gestational diabetes, which is similar to our previously reported findings in placental vessels from pregnancies with severe preeclampsia (5).

An alteration in prostaglandin formation has been observed in placentas of women with gestational diabetes. It was shown that although there was no difference in thromboxane production, there was a significant decrease in prostaglandin I₂ production in these placentas compared with placentas of nondiabetic women (23). In our previous study it was speculated that a decrease in the production of prostaglandin I₂ could influence the prostaglandin-mediated contractile responses to H₂O₂ that are altered in gestational diabetic-derived placental vessels. The smaller relaxation to lactate seen in the placental vessels of pregnancies with gestational diabetes is not likely to be related to an alteration in prostaglandin production, because we have shown that the relaxation to lactate does not involve prostaglandin mediators (19). On the basis of these observations and the absence of a detectable change in the reactivity of diabetic placental vessels to contraction with prostaglandins (6), cyclooxygenase-derived mediators do not appear to contribute to the alterations in responses to H₂O₂ and lactate that are examined in this study.

Endothelium-derived nitric oxide originating from the metabolism of L-arginine (10) is known to be formed in the human placental circulation (3, 22). We have shown that the mechanism of relaxation to lactate does not involve mediators produced from the endothelium, because there was no difference in relaxation to lactate between endothelium-intact and endothelium-denuded vessels (19). The mechanism of relaxation to lactate did not involve mediators derived from the metabolism of arginine, because pretreatment with nitro-L-arginine, an inhibitor of the production of arginine-derived nitric oxide, did not affect the relaxation to lactate (19). Similarly, the cGMP-associated relaxation to H₂O₂ observed in placental veins in the presence of indomethacin is also not influenced by removal of the endothelium or the presence of nitro-L-arginine (17, 19). On the basis of the absence of detectable changes in the responses to nitroglycerin, gestational diabetes does not appear to alter placental vascular relaxation to nitric oxide generated by this vasodilator drug or the ability of nitric oxide to stimulate soluble guanylate cyclase. Therefore, the observed alterations in relaxation to lactate and H₂O₂ are probably not related to modification of a relaxation mediated by the endothelium or arginine-derived nitric oxide or as a result of changes in the ability of soluble guanylate cyclase to produce a cGMP-mediated relaxation.

In placental vessels of normal pregnancies, lactate produces a relaxation response through a mechanism (Fig. 7) that appears to incorporate an O₂-dependent cGMP-mediated relaxation (19). Our previous studies suggest that the intracellular formation of H₂O₂, produced by lactate increasing endogenous NADH oxidase activity through an elevation of cytosolic NADH, contributes to the relaxation caused by lactate in several different vascular preparations (4, 12, 14, 20). The data in the present study indicate that a process in the mechanism of relaxation to lactate is altered in placental vessels of pregnancies with gestational diabetes.

The nature of the observed alterations in responses to H₂O₂ in placental vessels from gestational diabetic patients suggests that a defect could be present in the mechanism of the relaxation response to H₂O₂ and that this defect could also be contributing to the altered response to lactate. Our previous work in isolated bovine pulmonary arteries has demonstrated that H₂O₂ metabolism by catalase mediates the stimulation of guanylate cyclase activity (1). We have also demonstrated that exposure to 50 nM nitric oxide for 2 min can selectively cause a prolonged attenuation of the relaxation of bovine pulmonary arteries to H₂O₂ and lactate by inhibiting vascular catalase activity (13, 24). In the present study, isolated human placental arteries and veins from women with gestational diabetes were observed to have a reduced level of metabolism of endogenously generated H₂O₂ by catalase and an increased production of nitrite, a stable decomposition product of nitric oxide. It could be hypothesized that the observed reduced level of H₂O₂ metabolism by catalase might originate from several different processes in addition to an inhibition of catalase, including a decrease in the formation of endogenous H₂O₂ or an increased rate of metabolism of H₂O₂ by other systems such as glutathione peroxidase. However, these latter two mechanisms do not appear to explain the observed behavior of placental vessels from diabetic patients that exhibit an increased expression of a prostaglandin-mediated contractile response to exogenous H₂O₂ or endogenously formed H₂O₂ that is thought to occur during posthypoxic reoxygenation (6). This is because peroxide metabolism by cyclooxygenase appears to turn on the production of prostaglandins (8), and a reduced level of H₂O₂ generation or increased metabolism of peroxide by other systems would be expected to reduce the responses that are observed. Thus these observations are consistent with a primary role for an impairment of catalase activity in the responses of placental vessels to H₂O₂ and lactate that were observed to be altered in women with gestational diabetes (Fig. 7). In addition, the increased level of nitric oxide production observed in placental arteries and veins from women with gestational diabetes may be the origin of the impaired catalase activity.

The decreased relaxation to H₂O₂ and lactate observed in placental vessels of pregnancies with gestational diabetes may help explain aspects of the alteration in responses to H₂O₂ and changes in PO₂ previously observed in these placental vessels (6). H₂O₂ and posthypoxic reoxygenation cause a markedly enhanced prostaglandin-mediated contraction of placental vessels obtained from women with gestational diabetes (6). In the present study the data are consistent with an impairment of the cGMP-mediated relaxation to exogenous and endogenously generated H₂O₂ in placental vessels from women with gestational diabetes. On the basis of the hypothesized model shown in Fig. 7, the previously reported (6) enhanced contractile responses to exogenous H₂O₂ and peroxide generated during posthypoxic reoxygenation could originate from a markedly reduced expression of the simultaneously activated relaxation response to H₂O₂ in these vessels. The potential inhibitory effect of an increase in the production of nitric oxide on catalase in placental vessels from patients with preeclampsia (11, 15) and gestational diabetes could be a contributing factor in the changes in H₂O₂-associated responses that are observed in these complications of pregnancy. Because relaxation to lactate could be a potential defense mechanism present in the placental circulation designed to protect against fetal hypoxia (19), it is possible that the loss of this mechanism in gestational diabetes and preeclampsia could contribute to the complications of pregnancy that are observed. As previously discussed (6), the enhanced prostaglandin-mediated contractile response to posthypoxic reoxygenation observed in diabetic vessels could potentially promote vasospasm in the placental circulation. However, additional studies are needed to better define the role of alterations in responses to lactate and endogenous H₂O₂ formation in the complications of gestational diabetes.