Post-lanosterol inhibition profile based classification of commonly used prescription medications

Abstract

Cholesterol is an essential structural component of all cells, and the sterol biosynthetic pathway provides critical precursors for essential homeostatic molecules. Sterol biosynthesis can be disrupted by pathogenic variants in genes, as well as commonly prescribed medications. These medications disrupt post-lanosterol biosynthesis at different steps. We attempted to classify their actions based on the biochemical signatures of sterol intermediates. Our previous screening of the NIH Clinical Compound library of >1800 compounds in clinical use suggested that over 30 medications can disrupt post-lanosterol biosynthesis. Of these, we selected 11 compounds for follow up in a human dermal fibroblast model. Using LC-MS/MS we measured 13 post-lanosterol intermediates in control, DHCR7 +/- and DHCR7-/- human dermal fibroblasts exposed to cariprazine, nebivolol, rotigotine, buspirone, lurasidone, fluoxetine, hydroxyzine, amiodarone, spiroxamine, vilazodone and ziprasidone. All tested cells were exposed to 4 concentrations of each medication. We found in all fibroblasts, regardless of DHCR7 genetic makeup, sterol biosynthesis was inhibited by the tested medications. These medications could be classified in 6 groups based on the post-lanosterol profiles they produced – those that were 1) primarily DHCR7 inhibitors (cariprazine, nebivolol and rotigotine); 2) EBP inhibitors (fluoxetine); 3) combined DHCR7 and DHCR14 inhibitors (buspirone and lurasidone); 4) combined EBP and DHCR7 inhibitors (hydroxyzine); 5) combined EBP and DHCR24 inhibitors (amiodarone); and 6) multi-enzyme inhibitors (vilazodone, ziprasidone, spiroxamine). In addition, DHCR7 +/- fibroblasts responded with greater sterol profile disruptions to all medications, while DHCR7 fibroblasts from patients with Smith-Lemli-Opitz syndrome showed generally a more plateaued response. Knowing the inhibition profile-based classification of medications that have a sterol inhibiting side effect might ultimately translate into safety recommendations during pregnancy and could be critical for new drug development.

Introduction

As a structural component of all cells, cholesterol is an essential molecule of life [1]. Furthermore, the complex sterol biosynthesis pathway provides precursors for hundreds of essential homeostatic processes throughout the body [2, 3]. The demand for cholesterol is the highest during late embryonic and early postnatal life [4]. The post-lanosterol biosynthetic pathway, which proceeds through two parallel, but interconnected pathways (Kandutsch-Russell and Bloch), is described in Supplemental Material 1.

Inborn cholesterol biosynthesis disruptions give rise to complex neurodevelopmental syndromes such as Smith-Lemli-Opitz syndrome (SLOS), lathosterolosis, desmosterolosis, CDPX2, CHILD syndrome, SC4MOL deficiency, and HEM dysplasia [5, 6]. These syndromes are characterized by complex dysmorphologies and intellectual disability. Of these, SLOS is perhaps the most studied disorder, arising from two pathogenic alleles in the 7-dehydrocholesterol reductase (DHCR7) [5]. This enzyme catalyzes one of the final steps in cholesterol biosynthesis, the conversion of 7-DHC to cholesterol. It also catalyzes the conversion of 7-DHD to desmosterol, which undergoes another enzyme-mediated reaction to cholesterol. The consequence of this enzyme insufficiency results in intellectual disability, complex facial dysmorphologies, microcephaly, cleft palate, talipes equinovarus (club feet), syndactyly, and hypospadias. Notably, up to 75% of the SLOS patients meet the criteria for autism spectrum disorders (ASD) [7].

At the molecular level, the inhibition of the post-lanosterol pathway enzymes has two main biochemical consequences: cholesterol level (the end-product of the pathway) is decreased, and the sterol precursors of this biosynthetic pathway accumulate or decrease, based on the exact place of the pathway inhibition. Notably, sterol intermediates are bioactive and play a critical role in the arising pathophysiology. For example, the last precursor of cholesterol, 7-dehydrocholesterol (7-DHC) is a highly reactive molecule, spontaneously oxidizes in cells, and gives rise to 7-DHC-derived oxysterols [8–11]. These oxysterols are toxic and significantly impair cell viability and growth [12–16].

Importantly, post-lanosterol biosynthesis can be also disrupted by commonly prescribed medications [17–20]. These medications are diverse in chemical structures and primary indications, and encompass antipsychotic, antidepressant, anxiolytic, metabolic and cardiovascular medications. As a recent human population study has identified DHCR7 inhibitors as teratogens [21], it is imperative that we better understand the sterol biosynthesis inhibition mechanisms of these medications.

The exact mechanism of post-lanosterol biosynthesis interference by commonly prescribed medications remains greatly understudied. To shed light on this process, we analyzed the biochemical effects of 11 commonly prescribed medications (cariprazine, nebivolol, rotigotine, buspirone, lurasidone, fluoxetine, hydroxyzine, amiodarone, spiroxamine, vilazodone and ziprasidone) on the post-lanosterol biosynthetic pathway. These compounds were identified based on our various screening studies, shared chemical similarity and previously published literature findings [17, 20, 22, 23]. The study was carried out with human dermal fibroblasts from control, DHCR7 +/- heterozygous and SLOS (DHCR7-/-) patients and the analysis was performed using LC-MS/MS. In this analysis we had three goals: 1) understand which post-lanosterol enzymes are inhibited by each of the medications; 2) test if DHCR7 +/- heterozygosity results in a potentially increased vulnerability to these medications; and 3) evaluate if these medications (often prescribed to SLOS patients) would further disrupt the already compromised post-lanosterol biochemical profile seen in SLOS.

Materials and methods

Materials

Unless otherwise noted, all chemicals were purchased from Sigma-Aldrich Co (St. Louis, MO). HPLC grade solvents were purchased from Thermo Fisher Scientific Inc. (Waltham, MA). Hydroxyzine was obtained from Selleck Chemicals (Houston, TX). All sterol standards, natural and isotopically labeled, used in this study are available from Kerafast, Inc. (Boston, MA).

Human dermal fibroblast cultures

This is redundant statement. Fibroblast cultures were described previously [24–26]. For this study we have used 4 control, 6 DHCR7-heterozygous and 3 DHCR7-SLOS human fibroblasts. All fibroblast cultures used for experiments were between passages 8 and 15. All cells were maintained in DMEM containing 25 mM glucose and 1 mM sodium pyruvate supplemented with 2 mM L-glutamine, 10% fetal bovine serum (FBS, Thermo Scientific HyClone, Logan UT) and 2 mM antibiotic (Invitrogen) at 37 °C and 5% CO₂. At the end of the incubation period with the compounds, Hoechst dye was added to all wells to count the total number of cells per well using an ImageXpress Pico and cell counting algorithm in CellReporterX-press. After counting, wells were rinsed twice with 1X PBS, 10 μL of antioxidant mixture (BHT/TPP) was added to each well and the plates were stored at −80 °C for sterol analysis.

Ethics approval and consent to participate

All methods were performed in accordance with the relevant guidelines and regulations. Informed consent was obtained from all participants who initially donated the cells. The UNMC Office of Regulatory Affairs (ORA) has determined that this project does not constitute human subject research as defined at 45CFR46.102. The cell lines are de-identified; the research was not FDA-regulated, and it does not involve any interventions or interactions with living individuals.

Compound exposure

All cells were cultured in DMEM with 10% delipidated FBS in 96-well plates for 6 days with fresh medium changed every 48 h. The summary of compound exposures, concentrations and human doses are presented in Supplemental Material 2.

Post-lanosterol LC-MS/MS measurements

Sterols were extracted by adding to each well of the 96-well plate the deuterated sterol standards mixture (10 µL) and MeOH (100 µL) [27]. The plate was agitated on a shaker for 30 min and then allowed to rest to settle cell debris. The MeOH with extracted sterols was transferred to an analysis plate and dried under vacuum. The derivatizing reagent was freshly prepared with 2-methyl-6-nitrobenzoic anhydride (20 mg), N,N-dimethylglycine (14 mg), DMAP (6 mg), and Et₃N (0.1 mL) in anhydrous CHCl₃ (0.9 mL). Derivatizing reagent (100 µL) was added to each sample and allowed to react at room temperature for 30 min. The samples were dried under vacuum and subsequently dissolved in MeOH (100 µL) for LC-MS/MS analysis. Samples were analyzed on an Acquity UPLC system equipped with ANSI-compliant well plate holder. The sterols (10 µL injection) were analyzed on an Agilent Poroshell EC-C18 column (10 cm × 2.1 mm, 1.9 µm) with CH₃CN:MeOH:H₂O, 70:25:5 (0.01% (v) formic acid, 1 mM NH₄OAc) mobile phase at a column temperature of 40 °C. The flow rate was 400 µL/min for 11.5 min, then ramped to 600 µL/min at 11.6 min with a total run time of 16 min. A TSQ Quantum tandem mass spectrometer (ThermoFisher) was used for MS detections, and data were acquired with a Xcalibur software package. Selected reaction monitoring (SRM) of the DMG derivatives was acquired in the positive ion mode using electrospray ionization (ESI). Samples were excluded from analysis only for technical measurement failures. The excluded samples represented less than 1% of total measurements performed. The investigators were blinded to the sample type during the preparation and data extraction phase of the experiments.

Statistical analyses

Sample size estimates are based on previously observed effect sizes and power calculations. Statistical analyses were performed using Graphpad Prism 10 for Windows and MS-Excel. Unpaired two-tailed t-tests were performed for individual comparisons between two groups. ANOVA and regression analyses were performed separately for wild-type, DHCR7 +/- and SLOS fibroblasts across the tested range of concentrations. Benjamini-Hochberg corrections were used to correct for multiple testing in our analyses. The Welch’s correction was employed when the variance between the two groups was significantly different. The p values for statistically significant differences in t-test in comparison to vehicle-treated cultures are highlighted in figure legends: *0.001–0.01; **0.0001–0.001, ***<0.0001.

Results

Post-lanosterol biosynthesis occurs through two parallel, interconnected paths – the Kandutsch-Russell and Bloch pathways (Supplemental Material 1). After analyzing the post-lanosterol biochemical profile from all the fibroblasts in our study, we were able to classify the tested compounds into 6 groups.

Primary DHCR7 inhibitors: cariprazine, nebivolol, and rotigotine

Cariprazine (CAR) is an antipsychotic [28–30], nebivolol (NEB) is an antihypertensive agent [31, 32], while rotigotine (ROT) is utilized for treatment of Parkinson’s disease, restless legs syndrome, and depression [33, 34]. They do not share an obvious chemical structure, yet all three caused an increase in levels of 7-DHC, indicating that they are potent inhibitors of DHCR7 (Fig. 1). Of this group, CAR is a prototypic and most potent inhibitor, highly elevating 7-DHC levels at concentrations of 1–5 nM. Control and DHCR7 +/- heterozygous fibroblasts show a strong dose response to escalating doses of CAR. In our in vitro system, CAR (1–5 nM) was the strongest DHCR7 inhibitor, followed by NEB (10–50 nM), while ROT (25–100 nM) had the weakest effect on sterol biosynthesis. Cell responses to ROT and NEB can be found in Supplemental Material 3-4. None of these three compounds had an effect on the already greatly elevated baseline levels of 7-DHC in SLOS fibroblasts. The highest tested concentrations of CAR, NEB and ROT approached 7-DHC levels seen in fibroblasts of SLOS patients.

Fig. 1. Post-lanosterol biochemical profiles of primary DHCR7 inhibitors.

Measurements represent the mean responses of multiple biological replicates originating from different patients (DHCR7 + / + CNT n = 4; DHCR7 +/- HET n = 6; DHCR7-/- SLOS n = 3). Chemical structure is denoted in the top left corners of the panels. Y axis in Log₁₀ scale denotes sterol intermediate levels in nmol/million cells, while X axis indicates the measured post-lanosterol analytes. Medication concentrations are depicted below each panel. Bars represent a summary of multiple biological and technical replicates for each treatment. Statistical significance: *** p < 0.0001. Note the sharp, dose dependent increase in DHCR7 + /+ (CNT) and DHCR7 +/- (HET) human dermal fibroblast in 7-DHC levels that reach ranges seen in SLOS patient fibroblasts.

Combined DHCR7 and DHCR14 inhibitors: buspirone and lurasidone

Buspirone (BUS) is used as an antidepressant [35–37] and lurasidone (LUR) is an antipsychotic for treatment of schizophrenia and bipolar depression [38, 39]. Both compounds have similarities in chemical structure (piperazine ring). BUS and LUR inhibit both DHCR7 and DHCR14, altering levels of 14d-zymosterol, 14d-zymostenol, 7-DHC, and desmosterol in a dose-dependent manner (Fig. 2 and Supplemental Material 3-4). In this group, based on the sterol profile disruption, BUS (50–250 nM) is the prototype pharmacological agent. BUS showed a more robust impact on sterol biosynthesis than LUR (250–1000 nM). While both compounds increased 7-DHC in control and DHCR7-carrier cells, there was no effect on 7-DHC levels in SLOS cells, likely due to already plateaued 7-DHC levels in the DHCR7-/- genotype. Interestingly, the effect on 14d-zymosterol, 14d-zymostenol, and desmosterol was similarly dose-dependent in all fibroblasts regardless of genotype.

Fig. 2. Post-lanosterol biochemical profiles of combined DHCR7 and DHCR14 inhibitors.

Measurements represent the mean responses of multiple biological replicates originating from different patients (DHCR7 + / + CNT n = 4; DHCR7 +/- HET n = 6; DHCR7-/- SLOS n = 3). Chemical structure is denoted in the top left corners of the panels. Y axis in Log₁₀ scale denotes sterol intermediate levels in nmol/million cells, while X axis indicates the measured post-lanosterol analytes. Medication concentrations are depicted below each panel. Bars represent a summary of multiple biological and technical replicates for each treatment. Statistical significance: *** p < 0.0001. Note the sharp, dose dependent increase in DHCR7 + /+ (CNT) and DHCR7 +/- (HET) human dermal fibroblast in 14d-zymosterol, 14d-zymostenol and desmosterol in all fibroblasts regardless of DHCR7 genetic makeup.

Primary EBP inhibitor: fluoxetine

Fluoxetine (FLU) is a selective serotonin reuptake inhibitor, and it is prescribed as an antidepressant [40, 41]. Fluoxetine increased levels of zymosterol, zymostenol, and 8-DHC in a dose-dependent manner in control and DHCR7 pathogenic variant carrier (heterozygous) human fibroblasts (Fig. 3 and Supplemental Material 3-4). Again, the SLOS cells reacted differently to FLU exposure, only with increased levels of zymosterol and zymostenol, but not 8-DHC.

Fig. 3. Post-lanosterol biochemical profiles of a primary EBP inhibitor.

Measurements represent the mean responses of multiple biological replicates originating from different patients (DHCR7 + / + CNT n = 4; DHCR7 +/- HET n = 6; DHCR7-/- SLOS n = 3). Chemical structure is denoted in the top left corner of the panel. Y axis in Log₁₀ scale denotes sterol intermediate levels in nmol/million cells, while X axis indicates the measured post-lanosterol analytes. Medication concentrations are depicted below the panel. Bars represent a summary of multiple biological and technical replicates for each concentration. Statistical significance: * p < 0.01 ** p < 0.001 *** p < 0.0001. Note the dose dependent increase in zymosterol and zymostenol in all fibroblasts regardless of DHCR7 genetic makeup.

Combined EBP and DHCR7 inhibitor: hydroxyzine

Hydroxyzine (HYD) is an antihistaminic agent, also used as anxiolytic [42, 43]. Based on the biochemical sterol altering profile, HYD inhibits both the EBP and DHCR7 enzymes. This resulted in dose-dependent elevation of zymostenol, zymosterol, and 8-DHC, and reduction in DHL, 7-DHD, and 7-DHC levels (Fig. 4 and Supplemental Material 3-4). This was observed in all fibroblasts, regardless of genetic makeup. One of the most noteworthy findings is the reduction of toxic 7-DHC levels in the SLOS fibroblasts, potentially suggesting that, after further validation, HYD might be considered a potential therapeutic treatment for SLOS patients. However, the simultaneous 8-DHC rise in the SLOS fibroblasts raises doubts in the effectiveness of such treatment.

Fig. 4. Post-lanosterol biochemical profiles of a combined EBP and DHCR7 inhibitor.

Measurements represent the mean responses of multiple biological replicates originating from different patients (DHCR7 + / + CNT n = 4; DHCR7 +/- HET n = 6; DHCR7-/- SLOS n = 3). Chemical structure is denoted in the top left corner of the panel. Y axis in Log10 scale denotes sterol intermediate levels in nmol/million cells, while X axis indicates the measured post-lanosterol analytes. Medication concentrations are depicted below the panel. Bars represent a summary of multiple biological and technical replicates for each concentration. Statistical significance: ** p < 0.001 *** p < 0.0001. Note the dose dependent increase in zymosterol, zymostenol and 8-DHC in all fibroblasts regardless of DHCR7 genetic makeup, and a decrease of 7-DHC, 7-DHD and DHL levels.

Combined EBP and DHCR24 inhibitor: amiodarone

Amiodarone (AMI) is used to prevent and treat irregular heartbeats [44, 45]. By inhibiting both EBP and DHCR24, it elevates levels of zymosterol, 7-DHD, and desmosterol (Fig. 5 and Supplemental Material 3-4), without much effect on other sterol intermediates. Fibroblasts of all genotypes showed a similar, dose-dependent response to AMI in rise of zymosterol and 7-DHD, but SLOS fibroblasts had only a small and non-significant increase in desmosterol levels compared to the untreated baseline.

Fig. 5. Post-lanosterol biochemical profiles of a combined EBP and DHCR24 inhibitor.

Measurements represent the mean responses of multiple biological replicates originating from different patients (DHCR7 + / + CNT n = 4; DHCR7 +/- HET n = 6; DHCR7-/- SLOS n = 3). Chemical structure is denoted in the top left corner of the panel. Y axis in Log10 scale denotes sterol intermediate levels in nmol/million cells, while X axis indicates the measured post-lanosterol analytes. Medication concentrations are depicted below the panel. Bars represent a summary of multiple biological and technical replicates for each concentration. Statistical significance: *** p < 0.0001. Note the dose dependent increase in zymosterol, desmosterol and 7-DHD in all fibroblasts regardless of DHCR7 genetic makeup.

Multi-enzyme inhibitors (DHCR7, DHCR14, DHCR24 and EBP): vilazodone, ziprasidone, spiroxamine

Vilazodone (VIL) is a selective serotonin reuptake inhibitor and is used to treat depression [46]. Ziprasidone (ZIP) is an antipsychotic used to treat schizophrenia, mania, and bipolar disorder [47, 48]. Spiroxamine (SPI) prevents sterol synthesis in fungi and it is used to treat mildew in cereals and control fungal disease in grapes, wheat, rye, barley and oats [49]. It is very toxic to aquatic life and it is harmful if swallowed, inhaled or comes in contact with the skin. Spiroxamine is not used in humans. The common feature for these three compounds is their similar sterol profile in human fibroblast cultures. They appear to be multi-enzyme inhibitors of the post-lanosterol biosynthetic pathway, including DHCR7, DHCR14, DHCR24 and EBP. All three of the compounds strongly influence 14d-zymosterol, 14d-zymostenol and desmosterol levels with less pronounced and consistent effects on 7-DHC, 8-DHC, lathosterol, and DHL. In this group, we believe that VIL is the prototype multi-inhibitor (Fig. 6). ZIP and SPI profiles and data can be found in Supplemental Material 3-4. Notably, the findings were once again mostly consistent across the three tested fibroblast genotypes.

Fig. 6. Post-lanosterol biochemical profiles of multi-enzyme inhibitors.

Measurements represent the mean responses of multiple biological replicates originating from different patients (DHCR7 + / + CNT n = 4; DHCR7 +/- HET n = 6; DHCR7-/- SLOS n = 3). Chemical structure is denoted in the top left corner of each panel. Y axis in Log₁₀ scale denotes sterol intermediate levels in nmol/million cells, while X axis indicates the measured post-lanosterol analytes. Medication concentrations are depicted below the panel. Bars represent a summary of multiple biological and technical replicates for each concentration. Statistical significance: *** p < 0.0001. Note the complex sterol disruptions in all fibroblasts regardless of DHCR7 genetic makeup, which include both increases and decreases.

A summary of sterol biosynthesis inhibitory side effects of all tested medications can be found in the Graphical Abstract and in the table in Supplemental Material 5. Examples of regression and ANOVA data can be found in Supplemental Material 6.

Discussion

The current study expands and validates the list of the prescription medications with off-target sterol biosynthesis inhibition. Furthermore, it provides a classification of their effects based on the post-lanosterol biochemical profiles they produce (see Graphical Abstract). Our studies revealed that 1) all these medications have significant post-lanosterol biosynthesis altering effects; 2) the tested medications can be classified into different subgroups based on their post-lanosterol biosynthesis inhibition profiles; 3) cholesterol decreases are a less than ideal readout of sterol inhibition, as cholesterol changes are very small compared to intermediate changes; 4) DHCR7 +/- heterozygous, SLOS and control human dermal fibroblast react with a similar sterol profile change to these medications, with the DHCR7 +/- cells generally showing a stronger sterol disruption that often approach levels seen in SLOS.

The sterol biosynthesis inhibiting side effects of prescription medications were first described in 2013 by Hall and colleagues [50]. They noticed that aripiprazole and trazodone caused elevations of 7-dehydrocholesterol in the absence of SLOS pathogenic variants. This finding was later verified and further advanced in various follow-up studies [51, 52]. In vitro [26, 27], in vivo, pregnant transgenic mouse models [53–56], human biomaterials from psychiatric patients [57], pregnant women blood samples [58] and postmortem studies [59] all yielded concordant results – these two medications were DHCR7 enzyme inhibitors in addition to their main mechanism of action. These studies also paved the way for identification of more than a dozen prescription medications that showed similar sterol inhibiting side effects. Many of these medications are highly prescribed by physicians around the world, with an estimated >200 million yearly prescriptions in the US alone [60, 61].

Notably, many of these medications are routinely prescribed to pregnant women [62, 63]. This is of particular concern knowing the critical role of sterol biosynthesis during intrauterine and early postnatal life. Furthermore, a 2016 study by Bolland and Tatonetti raised concerns about sterol inhibition during pregnancy [21]. They described a functional environment around DHCR7 inhibition in an investigation of fetal outcomes following prenatal exposure to DHCR7 modulators. They found that first-trimester exposure to DHCR7 inhibitors resulted in outcomes similar to those of known teratogens and suggested that DHCR7 activity should be considered during drug development and prenatal toxicity assessment. This concern is further magnified by the increasing polypharmacy use, where multiple medications with sterol-inhibiting side effects might be prescribed to pregnant women [63]. In this context, animal studies revealed that such sterol-inhibiting polypharmacy might have synergistic effects during development, which could put the developing baby at further risk for adverse outcomes.

The inclusion of DHCR7 +/- fibroblasts (from parents of SLOS patients) in our study was essential for understanding the risk of potential heterozygosity and genetic factors on the disturbances of the sterol biochemical profiles – and potentially human health during the development of organ systems. Single-allele heterozygous DHCR7 pathogenic variant carriers make up about 3% of the human population [64]. While these DHCR7 +/- genetic variants alone have not been associated with adverse outcomes, these carriers have mildly elevated 7-DHC levels at baseline. Our current study, in alignment with previous studies [26], suggests that the DHCR7 +/- fibroblasts show a more robust sterol disruption response to these medications. This would suggest that the most likely adverse effects in the human population would arise when maternal single allele DHCR7 +/- heterozygosity, developing fetus heterozygosity, and sterol biosynthesis inhibiting polypharmacy would be all present. However, it is also notable that we have very limited knowledge of the effects of pathogenic variants in the other enzymes in the sterol biosynthetic pathway.

We believe that these effects do not merely represent findings in cell lines. The sterol biosynthesis pathway is highly conserved across tissues and vertebrate species [65]. Previously reported sterol biosynthesis disruptions by aripiprazole and trazodone were similar across Neuro2a cells, mouse models, human blood samples from pregnant women, sera from psychiatric patients and postmortem brain samples [17, 26, 57, 59]. This is suggestive that our findings could have a potential impact on human health, too. This leads us to a second important question: do the benefits of these medications for patients outweigh the potential risk? We believe that this depends on several critical factors: genotype of the individuals, organ system and stage of life. Namely, cholesterol is a stable molecule with a turnover rate measured in years in the adult brain, and it is unlikely that these precursor changes would have a major effect in adulthood. However, based on our previous studies we believe that these medications could have a deleterious impact on intrauterine brain development when the unborn child has single-allele pathogenic variant in a sterol gene. We propose this based on multiple converging scientific evidence. First, DHCR7 inhibitors can be considered teratogens in human populational studies [21]. Second, in every in vivo and in vitro system studies to date have shown that sterol inhibition is stronger in DHCR7 +/- heterozygosity. Third, the brain is only 2% of body weight, but contains 25% of sterols in the human body [4]. Fourth, brain development is the period of the strongest sterol biosynthesis [4]. Finally, we know that some of these sterol intermediates and the oxysterols they generate are toxic for both developing neurons and glial cells [12, 14].

These studies, while very informative about the biochemical sterol disruptions caused by medications, brought to the surface several additional questions. First, we do not know if this is a direct inhibition of the sterol biosynthesis enzyme(s), or a result of a complex molecular cascade inside the cells. This will have to be answered by follow-up gene expression and proteomic studies. Second, the increased and decreased level of each sterol precursor, in complex combinations, will likely have a different effect on human health. This will depend on the dose of medication, level of analyte changes and the chemical property/toxicity of the sterol intermediate. Third, the benefit or the risk of treatment of SLOS patients with these medications remains a very complex question. For example, CAR is a very potent inhibitor of DHCR7 in both heterozygous carriers and controls, but it has a small effect on 7-DHC levels in SLOS patient fibroblasts, perhaps due to already reaching plateaued levels. Furthermore, while ZIP and HYD both strongly reduce 7-DHC levels in all three fibroblast models (which should be theoretically beneficial for SLOS patients), they strongly affect multiple other sterol intermediates, with unknown consequences on the health during development.

The inhibition profile-based classification of medications that have a sterol inhibiting side effects we developed might ultimately translate into safety recommendations during pregnancy and could be critical for guidance of new drug development. It is also foundational data for precision medicine. The medications we tested are often necessary or even lifesaving for expectant women. Knowing which fluctuations in the sterol intermediate levels are safe, and which can be detrimental for the developing baby, especially in the context of the genetic makeup, will be essential for choosing the right and safe medication for a patient during pregnancy.

Author contributions

Conceptualization: ZK,NAP,KM. Formal analysis: ACA,KAT,ZK. Investigation: KAT,ACA,ZK. Writing-original draft: ZK, KM. Writing-review and editing: KAT,ACA,KM,NAP,ZK. Supervision: ZK,NAP,KM. Project administration: ZK, KM. Funding acquisition: ZK, KM.

Data availability

Data generated in this series of experiments will be deposited to the NCBI database at the time of acceptance to publication. Sterol measurements are also available in the Supplementary material spreadsheet.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41398-025-03785-7.