Prognostic and mechanistic potential of progesterone sulfates in intrahepatic cholestasis of pregnancy and pruritus gravidarum

Abstract

A challenge in obstetrics is to distinguish pathological symptoms from those associated with normal changes of pregnancy, typified by the need to differentiate whether gestational pruritus of the skin is an early symptom of intrahepatic cholestasis of pregnancy (ICP) or due to benign pruritus gravidarum. ICP is characterized by raised serum bile acids and complicated by spontaneous preterm labor and stillbirth. A biomarker for ICP would be invaluable for early diagnosis and treatment and to enable its differentiation from other maternal diseases. Three progesterone sulfate compounds, whose concentrations have not previously been studied, were newly synthesized and assayed in the serum of three groups of ICP patients and found to be significantly higher in ICP at 9-15 weeks of gestation and prior to symptom onset (group 1 cases/samples: ICP n = 35/80, uncomplicated pregnancy = 29/100), demonstrating that all three progesterone sulfates are prognostic for ICP. Concentrations of progesterone sulfates were associated with itch severity and, in combination with autotaxin, distinguished pregnant women with itch that would subsequently develop ICP from pruritus gravidarum (group 2: ICP n = 41, pruritus gravidarum n = 14). In a third group of first-trimester samples all progesterone sulfates were significantly elevated in serum from low-risk asymptomatic women who subsequently developed ICP (ICP/uncomplicated pregnancy n = 54/51). Finally, we show mechanistically that progesterone sulfates mediate itch by evoking a Tgr5-dependent scratch response in mice.

Conclusion: Our discovery that sulfated progesterone metabolites are a prognostic indicator for ICP will help predict onset of ICP and distinguish it from benign pruritus gravidarum, enabling targeted obstetric care to a high-risk population. Delineation of a progesterone sulfate-TGR5 pruritus axis identifies a therapeutic target for itch management in ICP.

Figure 1

Gestational serum profiles of PM2DiS, PM3S, and PM3DiS in group 1. Panels A, B, and C show the mean concentrations of PM2DiS, PM3S, and PM3DiS, respectively, for serum samples obtained at different gestational time points from women with uncomplicated pregnancies (control, closed squares), untreated ICP (closed circles), and UDCA‐treated ICP (closed triangles). Error bars represent ± standard error of the mean. P values for gestational week category comparison of untreated ICP versus controls were determined by Student t test.

Figure 2

Progesterone sulfates and autotaxin can predict subsequent onset of ICP in pregnant women with pruritus. The receiver operating curves (A) improved toward an optimal area under the curve of 1.0 when biomarkers were evaluated in combination: PM2DiS + PM3DiS (complete line), autotaxin (dashed line), and PM2DiS + PM3DiS + autotaxin (dotted and dashed line). (B) A combined predictive score (PM2DiS + PM3DiS + autotaxin) of greater than 0.25 for individual samples plotted against the gestational day of sampling reliably predicted all ICP cases. Women who developed ICP (n = 14, closed circles) and PG (n = 14, open triangles) were reliably distinguished by this score; dashed line represents demarcation between the two groups. Abbreviation: AUC, area under the curve.

Figure 3

Progesterone metabolites can activate TGR5 and elicit a Tgr5‐mediated itch response in mice. cAMP formation was monitored over time in HEK293 cells expressing TGR5 or control cells that were treated with 10 μmol/L of forskolin or increasing concentrations of PM3S, PM3DiS, or PM2DiS. Data are presented as a dose response for all three compounds in both cell types (A) or time course for PM3S in TGR5‐expressing cells (B). Values represent mean ± standard deviation of n = 3. (C) Wild‐type or Tgr5‐KO mice were intradermally injected with vehicle or 20 μL of 100 μmol/L PM3S, and scratching events were counted for the indicated time periods. Values represent mean ± standard error of the mean of n ≥ 4. *P < 0.05 for vehicle versus PM3S‐administered mice scratch comparison; #P < 0.05 for PM3S WT versus PM3S Tgr5‐KO scratch comparison as determined by one‐way analysis of variance. Abbreviation: BRET, bioluminescence resonance energy transfer.