LIN28B regulates androgen receptor (AR) in human trophoblast cells through Let-7c

Abstract

LIN28B is an RNA-binding protein necessary for maintaining pluripotency in stem cells and plays an important role in trophoblast cell differentiation. LIN28B action on target gene function often involves the Let-7 miRNA family. Previous work in cancer cells revealed that LIN28 through Let-7 miRNA regulates expression of androgen receptor (AR). Considering the similarities between cancer and trophoblast cells, we hypothesize that LIN28B also is necessary for the presence of AR in human trophoblast cells. The human first trimester trophoblast cell line, ACH-3P was used to evaluate the regulation of AR by LIN28B, and a LIN28B knockdown cell line was constructed using lentiviral-based vectors. LIN28B knockdown in ACH-3P cells resulted in significant decreased levels of AR, and increased levels of Let-7 miRNAs. Moreover, treatment of ACH-3P cells with Let-7c mimic, but not Let-7e or Let-7f, resulted in significant reduction in LIN28B and AR. Finally, forskolin-induced syncytialization and Let-7c treatment both resulted in increased expression of syncytiotrophoblast marker ERVW-1 and a significant decrease in AR in ACH-3P. These data reveal that LIN28B regulates AR levels in trophoblast cells likely through its inhibitory actions on let-7c, which may be necessary for trophoblast cell differentiation into the syncytiotrophoblast.

Introduction

In humans, impaired trophoblast cell differentiation and invasion into the maternal spiral arteries is thought to be an underlying cause of placental disorders including preeclampsia (PE) and intrauterine growth restriction (IUGR) [Regnault et al 2002]. Placental dysfunction is a significant health problem as it not only can adversely affect maternal and fetal well-being but also has long-term health effects that include hypertension, cardiac disease, and obesity in offspring [Simmons 2012; Sarr el al 2012; Barker et al 1990]. Studies evaluating placentas in women with PE consistently show maternal and fetal vascular abnormalities, which may be related to trophoblast cell differentiation [Kovo et al 2013]. Therefore, understanding the regulation of human trophoblast differentiation and molecular events that control placental growth and development is crucial in understanding underlying causes of placental disorders.

Our previous research demonstrated a role for LIN28 in mouse and human trophoblast cell differentiation (Seabrook et al., 2013; West et al., 2018). LIN28 is important in maintaining pluripotency in human stem cells, and knockdown of LIN28 in human and mouse trophoblast cells leads to their differentiation [Seabrook et al 2013; West et al., 2018]. LIN28 is an RNA-binding protein that controls cell function directly through binding of target mRNAs (e.g., [Wilbert et al., 2012; Hafner et al., 2013]), or indirectly by regulating Let-7 miRNAs [Viswanathan et al., 2008]. LIN28A and B are two homologs of the heterochronic gene, LIN28 [Zhang et al 2016]. Both LIN28A and B are important in regulating the let-7 family of miRNAs by inhibiting let-7 processing at pri- and pre-miRNA stages, keeping cells in an undifferentiated state [Zhang et al 2016]. Let-7s themselves target mRNA for degradation or repression [Boyerinas et al 2010]. In general, as cells differentiate LIN28 levels decrease, which is associated with an increase in let-7 miRNAs.

Contrary to mice, human trophoblast cells primarily express LIN28B [West et al., 2018; Canfield et al., 2018], and in this study we explored the possibility that androgen receptor (AR) is a potential target for LIN28B. AR localizes to villous stromal cells, cytotrophoblast and syncytiotrophoblast in term placenta [Iwamura et al., 1994; Horie et al., 1992], and previous work in our lab suggests AR is involved in regulating placental angiogenesis through its interaction with VEGFA during pregnancy [Cleys et al., 2014]. Considering AR is a known regulator of cell proliferation and migration in cancer cells, we recently postulated this also may be true in trophoblast cells during early placental development [McWhorter et al., 2018]. Moreover, studies in prostate cancer demonstrated that LIN28B regulates AR expression through Let-7c miRNA [Tummala et al., 2013].

Therefore, the main objective of this study was to determine if LIN28B can regulate AR expression in trophoblast cells. First trimester human trophoblast cells ACH-3P were used to develop a LIN28B knock-down cell line and determine if LIN28B regulates AR. In addition, we explored the possibility that Let-7’s are involved in trophoblast AR expression.

Materials and Methods

Cell lines and treatments

The commonly used human first trimester trophoblast cell line ACH-3P cells (a gift from Ursula Hiden, Medical University of Graz, Austria) were used in this study. These cells were immortalized by fusion of primary first trimester human trophoblast cells with human choriocarcinoma cell line, AC1–1 15 [Hiden et al., 2007]. ACH-3P cells were authenticated by demonstrating the presence of trophoblast marker cytokeratin-7 and absence of vimentin as originally done [Hiden et al., 2007] (Supplemental Figure 1). Cells were cultured in DMEM F-12 medium (HyClone), 10% FBS (HyClone), and 1% PSA (Corning Life Sciences). Syncytialization of ACH-3P cells was induced with 40 µM forskolin (Sigma-Aldrich) treatment for 48 h. Control cells were treated with 0.1% DMSO. All experiments were run in triplicate with three biological replicates per assay.

Human first trimester placental samples

Human first trimester (11.5 week) placental samples were obtained from elective terminations from anonymous, non-smoking, non-drug using patients, in accordance with the Colorado State University Institutional Review Board. Samples were stored in sterile PBS upon collection and were transferred to ice cold 4% paraformaldehyde (Fischer Scientific) upon receipt. Samples were stored overnight at 4°C in PFA, then transferred to 70% ethanol at 4°C until embedded in paraffin blocks. For immunohistolocalization, first trimester (n=2) tissue samples were cut at 5 µm thickness taken from the center of paraffin blocks.

Immunostaining

To determine cellular localization of AR in human placental samples, 5 µm sections of embedded tissue were mounted onto charged glass microscope slides (Premiere). Sections were deparaffinized and rehydrated in successive 4 minute baths of Citrasolv (Decon Labs), 100% ethanol, 90% ethanol, 70% ethanol, 50% ethanol, and distilled water. Sections then underwent antigen retrieval using 10 mM sodium citrate buffer (10 mM Sodium citrate, 0.05% Tween 20, pH 6.0) and microwaved for 10 minutes after being brought to a boil. Sections were rinsed 3 times with PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH 7.4) then blocked at room temperature using 6% goat serum in PBS.

Sections were then incubated in a humidity chamber overnight at 4°C in polyclonal antibodies against AR (Abcam 74272) and hCG (NeoMarkers) at 1:100 for immunofluorescence. After washing 3 times in PBS, sections were incubated for 1 hour at room temperature in 1:1000 goat anti-rabbit secondary antibody (Abcam ab6721) for immunohistochemistry, or 1:1000 goat anti-rabbit AlexaFluor 488 (Abcam, ab150077) and goat anti-mouse AlexaFluor 594 (Abcam, ab150092) for immunofluorescence. After 1 hour, slides were washed in PBS then dehydrated by successive baths of 50%, 70%, 90%, and 100% ethanol. Diaminobenzedine (DAB) was used as the final chromogen for immunohistochemistry (product no. SK-4100; Vector Laboratories). Slides used for immunofluorescence were mounted with Prolong Gold containing DAPI (4’,6-diamidino-2-phenylindole) (Life Technologies). Secondary antibody only staining was used as a negative control. Slides were visualized using Olympus D73 camera on Nikon Eclipse E800 microscope and CellSense 1.3 software.

LIN28B Knockdown (KD)

LIN28B KD and non-target scramble control cell lines were previously created using commercially available MISSON shRNA Lentiviral Transduction Particles, lentiviral-based pLKO.1-puro vectors (Sigma-Aldrich TRCN0000122599) with a puromycin resistance gene for selection downstream of a human phosphoglycerate kinase eukaryotic promoter [West et al., 2018]. Real time PCR and Western blot analysis revealed a 79% knockdown of LIN28B mRNA [West et al., 2018] and approximately 85% knockdown of LIN28B protein (Suppl. Figure 1), resp.

Western blot

Western blot analysis was used to determine the presence AR protein in ACH-3P cells. Cells were lysed in RIPA buffer (20 mM Tris, 137 mM NaCl, 10% glycerol, 1% nonidet P-40, 3.5 mM SDS, 1.2 mM sodium deoxycholate, 1.6 mM EDTA, pH 8) with 10% protease/phosphate inhibitor cocktail (Sigma-Aldrich) and 1 mM phenymethanesulfonyl fluoride. The BCA protein assay kit (Pierce) was used to determine protein concentration. Absorbance was measured at λ 595 nm using a Biotek Synergy 2 Microplate Reader (Biotek). Protein was electrophoresed in 4–15% Mini-PROTEAN TGX Stain-Free precast gels (Biorad) and transferred to 0.2 μm pore nitrocellulose membrane (Biorad) at 110 volts at 4°C for 1 hour.

Membranes were then blocked in 5% milk-TBST (50 mM Tris, 150 mM NaCl, 0.05% Tween 20, pH 7.6) for 1 hour at room temperature then incubated in a goat polyclonal antibody to AR (Abcam ab19066) or rabbit polyclonal antibody to LIN28B (Bethyl A303–588A) diluted 1:1000. Blots were incubated at 4°C overnight. Membranes were then washed 3 times for ten minutes for a total of 30 minutes in 1x TBST, incubated with a horseradish peroxidase-conjugated secondary antibody rabbit anti-goat (Abcam ab6741) and goat anti-rabbit (Pierce 1858415) for 1 hour at room temperature, diluted 1:2000. Rabbit polyclonal antibodies; GAPDH (paired with AR) and αTubulin (paired with LIN28B), were used as loading controls. Membranes were developed using ECL Western Blotting Detection Reagent chemiluminescent kit (Thermo Fisher Scientific) and imaged using a ChemiDoc XRS+ chemiluminescence system (BioRad). Densitometry was measured using ImageJ software (NIH) to compare relative density of normalized protein samples and student’s t-test was used to compare relative intensity between ACH-3P non-target controls and LIN28B KD cells. P values ≤ 0.05 were considered to be statistically significant. All experiments were run in triplicate with three biological replicates per assay.

Real-Time RT-PCR

Messenger RNA was extracted from cell pellets using mRNeasy Mini Kit (Qiagen) while small RNAs were extracted from cell pellets using miRNA Mini Kit (Qiagen) according to the manufacturer’s directions. Complementary DNA (cDNA) was generated from 1µg of total RNA using qScript cDNA Supermix (Quanta Biosciences) and quantitative real-time RT-PCR (qPCR) of mRNA was performed. Each 20ul qPCR reaction consisted of 10μl SsoAdvanced Universal Probes Supermix (Biorad), 1μl of 150 nM TaqMan Gene Expression Assay (Applied Biosystems; (Table 1), and 9μl of cDNA template diluted to 11ng/uL. Quantitative PCR was performed using the LightCycler480 thermal cycler (Roche) with the following parameters: 10 min pre-incubation at 95°C, 40 cycles of amplification, which included denaturation at 95°C for 10 seconds, annealing at 60°C for 30 seconds and extension at 72°C for 30 seconds, followed by a final cooling cycle at 40°C for 5 min. Normalization of mRNA in human cells was accomplished using GAPDH.

Table 1:

Accession numbers for probes used in all real-time RT-PCR assays

Gene	Accession Number
AR	NM_001011645.2
LIN28B	NM_001004317.3
ERVW-1	NM_001130925.1
LGALS13	NM_013268.2
GAPDH	NM_001256799.2

For miRNA analysis, cDNA was generated from 500 ng of total cellular RNA using miScript II RT Kit (Qiagen). Each 10μl miRNA qPCR reaction consisted of 5μl LightCycler 480 SYBR Green I Master mix (Roche), 1μl of forward primer, 1μl universal reverse primer (Qiagen miScript for human let-7 miRNA primers), and 8μl of cDNA template diluted to 1ng/uL. MicroRNA qPCR was performed using the LightCycler480 thermal cycler (Roche) with the following cycling parameters: 15min enzyme activation step at 95°C, followed by 45 cycles of amplification, which included denaturation at 94°C for 15 seconds, annealing at 55°C for 30 seconds, and extension at 70°C for 30 seconds, followed by a melting curve analysis to confirm amplification quality and specificity. Normalization of let-7 miRNA levels in human cells was calculated using the geometric mean of small nucleolar RNA, C/D box 41 (SNORD41), and small nucleolar RNA, C/D box 44 (SNORD44) miRNAs. For all qRT-PCR experiments, normalized data was analyzed and plotted as 2^-ΔCp. Student’s T-test was used to compare relative mRNA levels in non-target controls and LIN28B KD ACH-3P cells. P values ≤ 0.05 were considered to be statistically significant. For accession numbers of Taqman probes, see table 1. All experiments were run in triplicate with three biological replicates per assay.

Let-7 mimic transfection

ACH-3P WT cells were seeded at 4×10⁴ cells per well in 500ul DMEM F-12 on a 24 well plate. Once adhered, cells were transfected with hs-let-7c-5p, hs-let-7e-5p, or hs-let-7f-5p (Qiagen) to mimic human mature miRNA (5’UGAGGUAGUAGGUUGUAUGGUU, 5’UGAGGUAGGAGGUUGUAUAGUU, and 5’UGAGGUAGUAGAUUGUAUAGUU respectively) and 3uL HiPerFect Transfection Reagent (Qiagen) to a final concentration on 5nM of mimic. Control cells were transfected with Mission siRNA Universal Negative Control (sic-002) and 3uL HiPerFect Transfection Reagent at the same concentration. Cell pellets were collected at 48 hours for total RNA and protein. All mimic experiments were performed with three bioreplicates per treatment. Student’s t-test was used to compare let-7 levels between Universal siRNA controls and mimic treated ACH-3P cells. All experiments were run in triplicate with three biological replicates per assay.

Statistics

GraphPad Prism 7^© software was used to perform unpaired t-tests using qRT-PCR 2^-ΔCp values and Western blot band density comparing non-target scramble controls and LIN28B KD cells. A p-value of less than or equal to 0.05 was considered statistically significant.

Results

AR protein in the human placenta

Previous studies in our lab demonstrated the presence of LIN28A in mouse placenta and human ACH-3P cells, and revealed a role for LIN28A in trophoblast cell differentiation. Subsequent studies uncovered that human first trimester placenta primarily contain LIN28B [West et al., 2018; Canfield et al., 2018], which localized to the cytotrophoblast. The goal of this study was to determine if AR is regulated by LIN28B in trophoblast cells. In human first trimester placenta, AR appears to localize to trophoblast cells right underneath the syncytiotrophoblast layer, which is positive for hCG, as well as in trophoblast stromal cells (figures 1A and B). Furthermore, Western blot analysis revealed a single band of ~79kDa for AR detected in ACH-3P cell lysate (Suppl Figure 2).

Figure 1. AR localization in human first trimester placental tissue.

Localization according to immunohistochemistry (A, 20X) and immunofluorescence (B, 40X). Immunofluorescent co-labeling of AR (green) and hCG (red) suggests localization of AR to cytotrophoblast (white arrow) and stromal cells (blue arrow). Inserts in B indicate DAPI stained, no primary antibody negative control. Scale indicates 20µm.

Effects of ACH-3P LIN28B KD on AR

ACH-3P cells treated with lentiviral-based pLKO.1-puro vectors (Sigma-Aldrich) designed to specifically target LIN28B were used to examine the effect on AR transcript and protein using qRT-PCR and Western Blot, respectively. AR mRNA and protein levels both were significantly decreased in ACH-3P LIN28B KD cells versus non-target scramble controls (figure 2).

Figure 2. Relative levels of AR mRNA (A) and protein (B) in ACH-3P scramble control and LIN28B knockdown (KD) cells.

AR mRNA and protein levels are both significantly lower in ACH-3P LIN28B KD cells compared to scramble controls. * indicates p ≤ 0.05, ** indicates p ≤ 0.01. All experiments were run in triplicate with three biological replicates per assay.

Effects of LIN28B KD on let-7 miRNA

LIN28 is a well-known regulator of Let-7 miRNAs. To determine which Let-7 miRNA is regulated in ACH-3P cells by LIN28B, and as such could play a role in regulation of AR levels, relative miRNA levels were analyzed for Let-7a, b, c, d, e, f, g, and i using qRT-PCR. Let-7 family members; let-7c, let-7e, let-7f, let-7g and let-7i were significantly higher in ACH-3P LIN28B KD cells compared to non-target scramble controls (figure 3).

Figure 3. Relative levels of let-7 miRNA in LIN28B KD cells compared to non-target scramble controls.

Let-7c, e, f, g, and i miRNA levels were significantly increased in the LIN28B KD cells compared to non-target scramble controls. * indicates p ≤ 0.05, ** indicates p ≤ 0.01. All experiments were run in triplicate with three biological replicates per assay.

Effects of Let-7 overexpression in ACH-3P cells.

To determine if AR is regulated by specific let-7s, ACH-3P cells were treated with let-7e, f, or c mimic or siRNA universal negative control. Let-7e and f were chosen based on their high and significantly increased levels in LIN28B KD ACH-3P cells, whereas let-7c levels also were significantly higher, and previously shown to regulate AR levels in prostate cancer cells. As expected, let-7c, e, and f miRNA was significantly increased in the mimic transfected ACH-3P WT cells versus controls (supplemental figure 4). Let-7c overexpression led to significantly decreased LIN28B and AR levels, and significantly decreased AR protein (figure 4). Cells treated with let-7e and f mimics did not yield changes in LIN28B or AR levels (figure 4).

Figure 4. Effect of Let-7c, e, and f overexpression in ACH-3P cells on AR and LIN28B.

A 24-hour treatment of ACH-3P cells with let-7c mimic significantly decreased LIN28B (A) and AR levels (B) compared to siRNA Universal controls. Let-7e and f mimic treatment did not alter LIN28B or AR levels compared to siRNA Universal controls. Let-7c mimic significantly decreased AR protein compared to siRNA universal controls (C). * indicates p ≤ 0.05. All experiments were run in triplicate with three biological replicates per assay.

Effects of let7-c and forskolin treatment on markers of trophoblast cell differentiation

ERVW-1 and LGALS-13 both have been used as markers of trophoblast cell differentiation and are expressed the syncytiotrophoblast (ERVW-1, LGALS-13) or extravillous trophoblast (LGALS-13). Our previous work indicated that LIN28B KD leads to increased ERVW-1 levels and increased secretion of hCG in culture media, but decreased LGALS13 levels [West et al., 2018]. To determine if let-7c impacts trophoblast cell differentiation, ACH-3P cells were treated with let-7c mimic or siRNA universal negative control. Let-7c overexpression also significantly increased ERVW-1 levels and significantly decreased LGALS13 levels in ACH-3P cells compared to controls (figure 5). Finally, because let-7c decreased AR and increases syncytiotrophoblast marker ERVW-1, the possibility was explored that ACH-3P syncytialization is accompanied by decreased AR. ACH-3P WT cells were treated with 40µM forskolin for 48h which resulted in significantly increased ERVW-1 levels (figure 6A) and significantly decreased AR (figure 6B) levels compared to DMSO treated controls. These data suggest LIN28B knockdown leads to a decrease in AR, which possibly involves let-7c. Furthermore trophoblast cell fusion into syncytiotrophoblast is accompanied by a decrease in AR.

Figure 5. Effect of Let-7c overexpression in ACH-3P cells on ERVW-1 and LGALS13.

A 24-hour treatment of ACH-3P cells with let-7c mimic significantly increased ERVW-1 (A) and significantly decreased LGALS13 levels (B) compared to siRNA Universal controls. * indicates p ≤ 0.05, ** indicates p ≤ 0.01. All experiments were run in triplicate with three biological replicates per assay.

Figure 6. Relative levels of ERVW-1 (A) and AR (B) in forskolin treated ACH-3P cells compared to DMSO controls.

Relative levels of ERVW-1 were significantly higher (A) whereas AR levels were significantly lower in forskolin treated cells compared to controls. * indicates p ≤ 0.05, ** indicates p ≤ 0.01. All experiments were run in triplicate with three biological replicates per assay.

Discussion

The goal of this study was to determine if AR is a target for LIN28 and regulated by LIN28 in trophoblast cells. We previously demonstrated that LIN28 is present in the placenta and plays a role in trophoblast differentiation [Seabrook et al., 2013; West et al., 2018]. Specifically, knockdown of LIN28A in mouse trophoblast stem cells leads to their differentiation into trophoblast giant cells, whereas LIN28A knockdown in ACH-3P cells leads to increased syncytialization [Seabrook et al., 2013]. Since then, we found that LIN28B in fact is the predominant LIN28 homologue expressed in human first trimester placenta, localized to the cytotrophoblast cells, and present in commonly used human trophoblast cell lines [West et al 2018]. This was also recently confirmed by Canfield and colleagues [Canfield et al., 2018]. Similar to LIN28B, AR is not present in the differentiated syncytiotrophoblast layer, which is positive for hCG, but appears to localize to trophoblast cells right beneath the syncytiotrophoblast layer. These possibly are cytotrophoblast cells, based on a similar localization pattern as LIN28B [West et al., 2018]. Furthermore, AR is localized to the stromal cells confirming previous reports describing AR localization to the decidua and villous stromal cells [Horie et al 1992; Yoshida et al 2016]. Androgen and AR and their role in cancer have been extensively studied where it regulates cell differentiation, proliferation, migration and tissue vascularization [Yoshida et al 2016]. These are also events necessary for proper placental development and function.

Knockdown of LIN28B in ACH-3P cells leads to differentiation towards syncytiotrophoblast, with increased levels of ERVW-1 and secretion of hCG [West et al., 2018]. This study revealed that knockdown of LIN28B leads to significantly lower levels of AR mRNA and protein, suggesting LIN28B either directly or through let-7 miRNAs regulates AR abundance in trophoblast cells. Furthermore, knockdown of LIN28B resulted in significant increased levels of let-7c, e, f, and g. This is similar to West et al. [2018], showing let-7c was higher in LIN28B KD cells, although it was not statistically significant. To determine if AR possibly is regulated by let-7’s, ACH-3P cells were transfected with synthetic let-7 mimics. Mimics for let-7e and f were selected because of their high abundance in LIN28 KD cells, and let-7c because, in prostate cancer cells, AR is regulated by let-7c. Interestingly, significantly decreased levels of both LIN28B and AR were observed only in let-7c mimic transfected cells. Cells transfected with let-7e or let-7f did not significantly alter LIN28B or AR levels. These data indicate that similar to what has been observed in prostate cancer cells, [Nadiminty et al 2012] AR is a target of let-7c, and possibly AR levels in trophoblast cells are maintained by LIN28B inhibiting maturation of let-7c. Furthermore, these data point to a role for AR in possibly preventing trophoblast differentiation towards syncytiotrophoblast layer, and possibly promoting trophoblast cell proliferation. Because LIN28B is a RNA binding protein, it also remains possible AR is a direct target of LIN28B.

LIN28 binding to let-7 miRNA precursors involves its cold shock domain (CSD) and zinc knuckle domain (ZKD), the latter recognizing a specific GGAG-like motif in the precursor stem loop [e.g., Piskounova et al., 2008]. Recently, Ustianenko and colleagues [Ustianenko et al 2018] identified the binding site for CSD in let-7 precursors and demonstrated that let-7 miRNAs fall in two separate catagories; precursors recognized by both CSD and ZKD and precursors recognized only by ZKD. The authors indicate this possibly explains why some let-7’s are suppressed more strongly than others, and also suggest that despite identical seed regions, let-7 miRNA targets are unique but not mutually exclusive gene sets. Moreover, it is unclear how this is further impacted by the observation that ACH-3P do contain LIN28A, albeit at lower levels, and its levels are increased in LIN28B KD cells.

Trophoblast cell differentiation was characterized by determining expression of the syncytiotrophoblast marker ERVW-1 as well as forskolin-induced syncytialization of ACH-3P cells for 48h [Hiden et al., 2007]. Knockdown of LIN28B results in increased levels of ERVW-1, as well as secretion of hCG in ACH-3P cells, characteristic of differentiated syncytiotrophoblast cells [West et al, 2018]. ERVW-1 is an envelope glycoprotein that encodes the fusion gene Syncytin-1 and is necessary for the fusion and differentiation of cytotrophoblast to form the syncytiotrophoblast layer of the placenta [Reubner et al 2013]. In this study we demonstrate that let-7c overexpression also leads to a significant increase in ERVW-1 suggesting both decreased LIN28B and increased let-7c leads towards differentiation of ACH-3P cells to a syncytiotrophoblast fate. Alternatively, in JEG3 human choriocarcinoma cells LIN28B knockdown leads to a small but significant reduction of ERVW-1 [Canfield et al., 2018]. It is unclear if this is due to compensatory up-regulation of LIN28A (as seen in ACH-3P cells) in these particular cells.

The similarity between LIN28B knockdown and let-7c overexpression was extended with the observation that LGALS13 levels were decreased in let-7c mimic treated cells compared to siRNA Universal controls. LGALS13 encodes Galectin-3, otherwise known as Placental Protein-13 (PP13) and is highly expressed in the syncytiotrophoblast and to a lesser extend in extravillous trophoblast [Than et al 2014]. This is contrary to our previous observation when LIN28A was knocked down in ACH-3P cells, which revealed an increase in LGALS13 [Seabrook et al., 2013]. However, complete abolishment of both LIN28A and LIN28B in ACH-3P cells also leads to a decrease in LGALS13 [West et al., 2018], suggesting regulation of this gene is more complex, which is not surprising considering its localization in not only syncytiotrophoblast but also extravillous trophoblast.

Finally, both LIN28B and AR are not present in the syncytiotrophoblast layer, and we demonstrate that syncytialization of ACH-3P cells is accompanied by a decrease in AR abundance, like LIN28B. In vitro forskolin treatment leads to fusion of trophoblast cells, similar to naturally syncytializing cells [Wice et al 1990]. As expected, treatment of ACH-3P cells with forskolin also led to significantly increased ERVW-1. Therefore, we conclude that LIN28B regulates AR in trophoblast cells, and this likely involves let-7c. Furthermore, as with LIN28B knockdown, let-7c overexpression decreases AR, and leads to differentiation towards syncytiotrophoblast layer.