HomeScienceA TGFβR inhibitor represses keratin-7 expression in 3D cultures of human salivary...

A TGFβR inhibitor represses keratin-7 expression in 3D cultures of human salivary gland progenitor cells


RGDSP cultures developed a mixed phenotype with increased K7 and α-amylase expression

Hydrogel precursors, including thiolated (HA-SH) and acrylated HA (HA-AES), as well as maleimide-functionalized RGDSP peptide (MI-RGDSP), were prepared as previously reported47. Cellular constructs were produced by dispersing hS/PCs in a solution containing HA-SH or HA-SH/MI-RGDSP before adding the HA-AES solution, resulting in HA or RGDSP cultures, respectively. To assess if the 3D hydrogel cultures promoted a K7-pro-ductal phenotype or reduced characteristics of an acinar phenotype, we characterized the temporal gene expression dynamics of salivary gland differentiation markers47 with qPCR (Fig. 1a). Keratins expressed by salivary gland progenitors, keratin-5 (K5, KRT5) and keratin-14 (K14, KRT14)6,48,49, presented similar temporal profiles that were unique to each hydrogel condition. There was a moderate increase in KRT5/KRT14 expression in HA cultures on day 1 that was sustained to day 14. KRT5/KRT14 expression in RGDSP cultures similarly increased on day 3; however, its expression decreased thereafter, with KRT5 expression levels decreasing significantly below that observed in HA cultures. Analysis of the gene expression profile of acinar markers, α-amylase (AMY1A) and the Na-K-Cl ion transporter (SLC12A2)1,6,49, revealed different trends. AMY1A expression was relatively stable over time in both HA and RGDSP cultures. Although AMY1A expression was elevated on day 3 in HA cultures, it returned to the basal level on day 7. HA cultures maintained SLC12A2 expression through day 7; however, on day 14, SLC12A2 was downregulated with respect to its expression level on day 1. Comparatively, loss of SLC12A2 expression occurred earlier in RGDSP cultures, with a significant reduction in expression occurring on day 7, and a further decrease on day 14.

Figure 1
figure 1

RGDSP gels promote the development of multicellular spheroids with increased K7 and α-amylase expression. (a) Temporal gene expression profile via qPCR analysis of KRT5, KRT14, SLC12A2, KRT7, TFCP2L1, FN, and LAMA1. (b) Expression of KRT18 and KRT19 on day 14 as analyzed by qPCR. (c) Western blot analysis for K5, K14, K7, and fibronectin on day 14. Blotting for fibronectin, K5, and K14 was performed sequentially on the same membrane (Fig. S7a–d). Blotting for K7 was performed separately (Fig. S7e,f). Triplicate measurements are presented from technical replicates extracted from different hydrogels. (d) Secreted α-amylase quantified by ELISA. (e) Immunocytochemistry (ICC) detailing uniform expression of α-amylase in day 14 spheroids. Single channel images are presented in Fig. S8. Quantification was conducted from 3 independent experiments, each with 3 technical repeats. Error bars represent SEM. Two-way ANOVA was conducted on data presented in (a) and (c), followed by Tukey’s multiple comparisons test. *p < 0.05 between HA and RGDSP cultures at the same time points. †,‡,§p < 0.05 from days 1, 3, and 7 measurements of the same data set, respectively for (a) and days 3, 6, 9 for (c). Two-tailed Student’s t-tests were conducted on data appearing in (b, c), *p < 0.05.

Next, we analyzed the temporal expression of KRT7 and a transcription factor involved in duct regulation, TFCP2L11,6,50. HA and RGDSP cultures stably expressed KRT7 until day 14, where RGDSP promoted a 2.2-fold higher KRT7 expression relative to the day-14 HA cultures (4.8-fold increase when normalized to the expression level of HA cultures on day 1). However, a corresponding increase in the expression of the K7 dimerization partners, keratin-18 (KRT18) and keratin-19 (KRT19)51,52, was not detected in RGDSP cultures on day 14 (Fig. 1b). TFCP2L1 expression was maintained in HA cultures until day 14, when it was downregulated relative to the normalized values of HA cultures on day 1. The expression of TFCP2L1 was maintained across the culture time, and significant variations in expression between HA and RGDSP cultures were not observed.

Finally, the gene expression dynamics of fibronectin (FN), required in salivary gland development53, and laminin, alpha-1 subunit (LAMA1), which is associated with the maintenance of the acinar phenotype54,55, was also analyzed. FN and LAMA1 were upregulated at days 3–7 under both culture conditions relative to the expression level by HA cultures on day 1. However, by day 14, FN was upregulated, and LAMA1 was downregulated in RGDSP cultures.

Protein level confirmation of these findings was made by western blot analyses of K5, K14, K7, and fibronectin (Fig. 1c). Although KRT5 mRNA expression was significantly lower in RGDSP cultures on day 14, we did not detect a loss in K5 expression on day 14 at the protein level. Furthermore, the expression of K14, the predominant heterodimeric partner of K551,52, was also retained in the RGDSP cultures at day 14. However, K7 and fibronectin levels were increased by ~ threefold (p < 0.05) in RGDSP cultures, in agreement with the corresponding mRNA expression profiles. As increased ductal K7 expression might be associated with the loss of the secretory function, amylase expression was evaluated via ELISA. RGDSP cultures produced approximately twofold more amylase per cell on day 12 than HA cultures (Fig. 1d). However, the immunostaining pattern of α-amylase indicated that α-amylase was uniformly expressed throughout both HA and RGDSP cultures and was not restricted to a specific cell population (Fig. 1e). Similarly, K7 was also uniformly expressed in RGDSP cultures (Fig. S2). Collectively, these findings indicate that RGDSP increased secretory amylase secretion while also enhancing expression of the K7.

TGF-β1 expression and nuclear SMAD 2/3 localization correspond to increased K7 expression

After confirming the development of the K7+ phenotype in RGDSP cultures, we set out to identify potential upstream regulators of K7. Literature mining indicates that members of the TGF-β family are responsible for activating K7 expression22,23,24,25,26,27,28,56. We used an immunoblot array to evaluate the production of TGF-β superfamily members in the medium collected from HA and RGDSP cultures on days 6 and day 14 (Fig. S3c–e). TGF-β1, GDF-15, and transforming growth factor-beta-induced protein (TGFBI/BIGH3) were the most highly detected TGF-β superfamily members in both HA and RGDSP cultures at both time points (Fig. S3e). TGF-β1 and GDF-15 were the only TGF-β superfamily members that increased expression from day 6 to day 14 in RGDSP cultures. Furthermore, only TGF-β1 was produced at higher levels in RGDSP cultures, relative to HA cultures, on day 14 when K7 expression was highest.

We then conducted TGF-β1 and GDF-15 ELISA to confirm the array findings and assess the temporal cytokine expression dynamics (Fig. 2a). On days 3–12, HA and RGDSP cultures increased TGF-β1 expression by 9 and 19 folds, respectively. However, RGDSP cultures maintained a higher level of TGF-β1 expression relative to HA cultures (> 2 folds) throughout the entire culture period. Notably, a 7- and a 25-fold increase in GDF-15 expression, relative to the initial day 3 HA levels, was observed at days 6 and 9 in HA cultures. However, RGDSP cultures increased 70- to 103-fold over the same period. Interestingly, GDF-15 levels continued to rise in HA cultures, and by day 12 were no longer expressed at significantly different levels in HA and RGDSP cultures.

Figure 2
figure 2

RGDSP cultures expressed high levels of TGF-β1 and GDF-15 that correlate with increased nuclear SMAD 2/3. Only TGF-β1 is required for KRT7 expression in 2D cultures. (a) ELISA experiments were conducted targeting soluble TGF-β1 and GDF-15 in cell culture medium after 3, 6, 9, and 12 days of culture (n = 9, 3 independent experiments, each with 3 technical repeats). (b) hS/PCs were cultured with TGF-β1, GDF-15, and TGF-β1 + GDF-15 for 48 h on 2D substrates, and the expression of K7 was resolved with qPCR. Data averaged from 3 independent experiments, each with 3 technical repeats. Significance determined from one-way ANOVA followed by a Dunnett’s test, * indicates p < 0.05 relative to vehicle control. (c) Fluorescent microscopy images detailing hS/PC nuclear YAP, SMAD 2/3 and nuclei on days 1, 3, 7, and 14. Volumetric fluorescent microscopy was conducted, and representative single plane images are presented. Single channel images are presented in Fig. S9. (d,e) 3D rendering was performed with Imaris 3D-4D software for the quantification of nuclear SMAD 2/3 (d) and YAP (e) signals. Filled circles represent individual nuclei, and the dashed black line indicates the mean value of each data set. Error bars represent SEM in all cases. Number of images analyzed for HA cultures: 103 (day 1), 107 (day 3), 85 (day 7) and 225 (day 14); Number of images analyzed for RGDSP cultures: 104 (day 1), 150 (day 3), 428 (day 7) and 1534 (day 14). Analysis was conducted from 3 independent experiments. Two-way ANOVA was performed on data presented in (a) and (d, e) followed by Tukey’s multiple comparisons test. * indicates p < 0.05 between HA and RGDSP at the same time points. For (a), †,‡  p < 0.05 from day 3, and day 6 measurements of the same data set, respectively. In (d, e), †,‡,§ indicates p < 0.05 from day 1, 3, and day 7 measurements of the same data set, respectively.

Since TGF-β1, GDF-15, and K7 were highly expressed in RGDSP cultures, we queried whether exogenous TGF-β1 (10 ng mL1)57,58 or GDF-15 (100 ng mL1)59,60 could induce KRT7 expression. hS/PCs were maintained on 2D substrates with each cytokine independently and in combination for 48 h before qPCR analyses was conducted (Fig. 2b). Treatment with GDF-15 alone did not alter KRT7 expression at the transcript level. In contrast, TGF-β1 induced a significant increase of KRT7 expression that was sustained in the presence of GDF-15 (GDF-15 + TGF-β1). This result suggests that TGF-β1 is capable of stimulating KRT7 expression.

To further rule out the possibility of GDF-15 in promoting K7 expression, we next inspected our cultures for signs of SASP, which is directly correlated to high levels of GDF-1536, and is associated with decreased salivary gland function38,40,61. Specifically, we compared the expression of a panel of SASP genes [IL636,40,41,61,62, (CXCL8, IL8)35,61,62, IL1063, IL1B62, (CDNK1A, P21)40,41, (CDNK2A, P16)40, (TP53, P53)32,35, GDF1536,61, SERPINE135,36,41, and MMP136,37] in HA and RGDSP cultures on day 14 to 2D cultures stimulated by TGF-β1, GDF-15, and TGF-β1 + GDF-15 (Fig. S4a,b). The expression of senescent and inflammatory factors was largely governed by TGF-β1 in 2D cultures, while the activity of GDF-15 was limited to increasing the expression of IL1B. With regard to the cell cycle inhibitors that are indicative of senescence35,36,37,40,41, TGF-β1 stimulation significantly increased the expression of cyclin-dependent kinase inhibitor 2A (CDKN2A) and tumor protein P53 (TP53) in 2D cultures. However, RGDSP cultures did not exhibit elevated levels of CDKN2A or TP53, suggesting growth arrest had not occurred. TGF-β1 significantly enhanced IL6 expression in 2D cultures; however, RGDSP cultures promoted increased CXCL8 expression without stimulating IL6 expression. These results suggest that interleukins are not responsible for KRT7 expression. Moreover, elevated expression of SERPINE1, a TGF-β1 target gene25,30, was observed in RGDSP cultures as well as TGF-β1 stimulated 2D cultures, further confirming that TGF-β1, not GDF-15 or SASP, was the dominant factor governing emergence of the K7+ phenotype.

We hypothesize that integrin-mediated adhesion via RGDSP can enhance mechanotransduction, by nuclear localization of YAP, to promote TGF-β signaling in RGDSP cultures42. To this end, we characterized TGF-β signaling and examined the role of mechanotransduction in inducing K7 expression by targeting SMAD 2/3 and YAP with double immunofluorescence (Fig. 2c). Nuclear localization of YAP or SMAD 2/3 is indicative of increased YAP and TGF-β signaling, respectively44,64,65. In HA cultures, nuclear SMAD 2/3 signal was increased at day 3, but returned to the baseline level by day 14 (Fig. 2d). Under the same condition, elevated nuclear YAP was detected on days 3 and 14. In RGDSP cultures, nuclear YAP intensity peaked on day 1, but continuously decreased thereafter (Fig. 2e).

A similar trend was observed for cytoplasmic YAP (Fig. S5a). As nuclear YAP decreased in RGDSP cultures, nuclear SMAD 2/3 increased significantly at day 7 and furthermore at day 14, resulting in significantly higher levels at day 14 than observed in HA cultures. However, cytoplasmic SMAD 2/3 levels did not significantly increase in RGDSP cultures (Fig. S5b). These findings confirm that the increased TGF-β signaling in RGDSP cultures corresponds to the rising TGF-β1 levels and increased K7 expression. Conversely, RGDSP cultures do not sustain elevated YAP expression to drive TGF-β signaling and K7 expression.

To confirm the microscopy findings, we further compared the temporal nuclear localization of SMAD 2/3 and YAP with other TGF-β and YAP targets (Fig. S5c). Irrespective of nuclear SMAD 2/3 and YAP expression, HA and RGDSP cultures were defined by an initial ADAM10/ADAM17 activation with the expression of EGFR ligands (EREG, AREG, HBEGF, TGFA), followed by a transition to MMP1 expression as the culture time proceeded. In HA cultures, the appearance of nuclear SMAD 2/3 and YAP together at day 3 dominated the transcriptional events where KRT7, FN, and ITGAV were maximally expressed. However, in RGDSP cultures, nuclear SMAD 2/3 expression was highest on day 14, corresponding to the highest KRT7, FN, and ITGAV expression. In RGDSP cultures, nuclear YAP expression clearly proceeded YAP1 and CTGF expression; however, in HA cultures, the presence of both SMAD 2/3 and YAP in the nucleus at day 3 maximized YAP1 and CTGF expression. Collectively, the temporal expression of SMAD 2/3 and YAP was confirmed both at the transcript and protein levels.

Loss of nuclear YAP stimulates GDF15 expression and downregulates TGF-β target genes

We have shown that TGF-β1 can stimulate KRT7 expression in hS/PCs (Fig. 2b), and we also know that YAP is involved in maintaining the stem/progenitor status of various tissues66,67. Thus, we questioned if loss of nuclear YAP expression, as observed in RGDSP cultures, could be sufficient to induce the expression of K766,67. To this end, a series of YAP inhibition studies were performed to investigate YAP-dependent TGF-β signaling. hS/PCs were maintained on 2D substrates with a verteporfin (VERT), an inhibitor of YAP-TEAD transcription68,69, at concentrations of 0.5, 1.0, and 2.0 µM for 24 h, in the absence of light, and the expression of nuclear SMAD 2/3 and YAP was resolved with fluorescent microscopy (Fig. 3a). As shown in Fig. 3b, VERT treatment effectively suppressed nuclear YAP expression. Compared to the vehicle controls, VERT at 0.5 μM decreased nuclear YAP expression by ~ 0.75 fold, with further reductions in YAP observed at 1.0 and 2.0 µM. Meanwhile, VERT treatment led to a significant increase in nuclear SMAD 2/3, with ~ 3.5 fold increase in response to 0.5 µM VERT (Fig. 3c). Of note, VERT treatment also increased the cytoplasmic SMAD 2/3 levels (Fig. 3d).

Figure 3
figure 3

Loss of nuclear YAP stimulates GDF15 expression and downregulates TGF-β1 target genes. (a) hS/PCs were cultured with verteporfin (VERT, 0.5, 1.0, and 2.0 µM) and the DMSO vehicle control for 24 h before SMAD 2/3 and YAP were visualized by ICC. Nuclei were counterstained with DAPI. Single channel images are presented in Fig. S10. (bd) Image quantification was performed with ImageJ to resolve nuclear YAP (b), nuclear SMAD 2/3 (c), and cytoplasmic SMAD 2/3 (d). Number of images analyzed: nDMSO = 33, n0.5 = 24, n1.0 = 19, n2.0 = 17, from 3 independent experiments, significance determined from one-way ANOVA followed by a Dunnett’s test, * indicates p < 0.05 relative to DMSO control. (eg) hS/PCs were treated with 1 μM VERT for 24 h and mRNA expression of YAP target genes: CTGF, CYR61 (e); TGF-β targets genes: JUNB, SERPINE1, and GDF15 (f); and keratins: KRT5, KRT7 (g) were resolved with qPCR (n = 9, 3 independent experiments, each with 3 technical repeats). (h) Schematic depiction detailing the role of nuclear YAP (black circle) in supporting TGF-β signaling (JUNB, SERPINE1) and repressing GDF15 expression. Error bars represent SEM in all cases. Two-tailed Student’s t-tests were conducted on data appearing in (eg). * indicates p < 0.05 in all cases.

Next, qPCR was performed targetting TGF-β and YAP pathway transcripts to characterize the effects of YAP inhibition on TGF-β signaling in cultures receiving 1.0 µM of VERT (Fig. 3e–g). As expected, CTGF and CYR61, canonical YAP target genes70, were downregulated, confirming the repression of YAP signaling with VERT treatment (Fig. 3e). The TGF-β targets, JUNB and SERPINE1, were upregulated in hS/PC cultures after treatment with TGF-β1 (Figs. S4b, S6c), indicating these genes responded to TGF-β signaling in hS/PC cultures. YAP inhibition downregulated JUNB and SERPINE1, indicating the supporting role of YAP in the expression of these transcripts (Fig. 3h). GDF15 was modestly upregulated (1.5-fold, p = 0.08) in TGF-β1 treated cultures (Fig. S4b); however, it was increased by ~ ninefold in response to VERT-mediated YAP inhibition. Additionally, the role of YAP in the regulation of keratin expression was investigated (Fig. 3g). KRT5 was downregulated with the loss of YAP, but alterations in KRT7 expression was not detected. These findings indicate that YAP represses GDF15 expression while supporting the expression of the TGF-β signaling components JUNB and SERPINE1 (Fig. 3h).

TGFβR inhibition represses TGF-β1 induced K7 expression in 2D cultures

To ascertain whether decreasing TGF-β signaling could suppress K7 expression, model 2D studies were conducted with a TGFβR inhibitor, A83-01. hS/PCs were cultured for 48 h with TGF-β1 and 2 µM of A83-01, a concentration reported not to exhibit off target signaling inhibition71. As shown in Fig. 4a, TGF-β1-treated cells were strikingly larger and extended with fewer cell–cell contacts, taking on a mesenchymal spindle shape morphology. A83-01 inhibition promoted an epithelial phenotype with F-actin localized to cell–cell contacts. When provided in combination with TGF-β1, A83-01 prevented the TGF-β1-induced morphology changes. Immunofluorescence microscopy detailed a granular expression pattern of K7, which increased in response to TGF-β1 (Fig. 4b). Furthermore, A83-01 was sufficient to resist the TGF-β1 mediated increase in K7 expression. Similarly at the mRNA level, A83-01 inhibited the TGF-β1 stimulated KRT7 expression while simultaneously decreasing KRT7 expression below that of the vehicle control (Fig. 4c). Immunofluorescence microscopy was performed to demonstrate further that KRT7/K7 was correlated with the ability of TGF-β1 to induce nuclear SMAD 2/3, and inhibition of KRT7/K7 expression via A83-01 was also accompanied by reduced nuclear SMAD 2/3 (Fig. 4d,e). Furthermore, we found that hS/PCs treated with exogenous GDF-15 did not exhibit increased nuclear SMAD 2/3 (Fig. S4c), further indicating aberrant TGF-β signaling promotes K7 expression.

Figure 4
figure 4

TGFβR inhibition represses TGF-β1 induced K7 expression. (a) hS/PCs were cultured with TGF-β1 and A83-01 for 48 h, and the expression of K7 and F-Actin was visualized with fluorescent microscopy. (b,c) A83-01 repressed TGF-β1 induced KRT7 expression at the mRNA level (b, n = 9, 3 independent experiments, each with 3 technical repeats), and at the protein level as determined by image analysis conducted with ImageJ (c, ID: integrated density; number of images analyzed: nVehicle = 55, nA83-01 = 55, nTGF-β1 = 40, nA83-01+TGF-β1 = 45 from 3 independent experiments). (d) hS/PCs were cultured with TGF-β1 and A83-01 for 48 h, and SMAD 2/3 and Ki-67 expression were investigated with ICC. White arrows indicate binuclear cells in TGF-β1 treated cultures. Single channel images of (a) and (d) panels are presented in Figs. S11 and 12, respectively. (e) ImageJ-derived analysis indicated TGF-β1 stimulated nuclear SMAD 2/3 that was inhibited by A83-01. Number of images analyzed: nVehicle = 70, nA83-01 = 57, nTGF-β1 = 41, nA83-01+TGF-β1 = 56, from 3 independent experiments. (f) Proliferation was assessed by enumeration of DAPI stained nuclei in TGF-β1, and A83-01 treated cultures after 48 h of culture. At 24 h, nVehicle = 629, nA83-01 = 637, nTGF-β1 = 308, nA83-01+ TGF-β1 = 705 from 3 independent experiments. (g,h) hS/PC expression of IGF2 (g) and KRT7-AS (h) with TGF-β1 and A83-01 treatment were investigated with qPCR. n = 9 from 3 independent experiments. Error bars represent SEM in all cases One-way ANOVAs were performed on data presented in (b), (c), and (eh) followed by Tukey’s multiple comparisons test. * indicates p < 0.05 in all cases. (i) Schematic depiction of A83-01 inhibition of TGF-β1 stimulated K7 expression.

TGF-β1 is a cytostatic factor towards epithelial cells32,33, and with TGFβR inhibition by A83-01, hS/PCs responded with rapid proliferation (Fig. 4f). In addition, competitively inhibited cultures receiving A83-01 retained higher proliferation than cultures receiving only TGF-β1 (A83-01 + TGF-β1 vs TGF-β1). The non-G0 state cell cycle marker, Ki-6772, was absent in TGF-β1 treated cultures, although the DMSO controls were stained positive (Fig. 4d). Higher expression of Ki-67 indicates that cells could proliferate, but the DMSO cultures did not proliferate to a significantly high degree than the TGF-β-treated ones over the 48-h time span of the experiment (A83-01 vs TGF-β1, Fig. 4f). On the other hand, nuclear Ki-67 was present in cultures treated with both A83-01 and TGF-β1, further demonstrating the ability of A83-01 to inhibit TGF-β signaling. Under the cytostatic effects of TGF-β1, cytokinesis failure can occur, producing binuclear cells73, which were present in TGF-β1 treated cultures (arrowheads, Fig. 4a,d). This is associated with genome instability, loss of imprinting (LOI), and led to the investigation of the expression of a long non-coding antisense (AS) RNA, KRT7-AS, reported to regulate KRT7/K7 expression73,74,75. We determined that in response to TGF-β1, IGF2, a paternally imprinted gene of insulin-like growth factor 2 (IGF-2)50,76, was upregulated ~ 20 fold, and concordantly, KRT7-AS expression was induced (Fig. 4g,h). Furthermore, TGF-β1 modestly upregulated insulin-like growth factor 1 receptor (IGF-1R, IGF1R), yet the expression of the non-imprinted insulin-like growth factor 1 (IGF-177, IGF1) was unaltered (Fig. S6a,b). A83-01 medium supplementation prevented TGF-β1 induced KRT7 and IGF2 expression (Fig. 4i), indicating that A83-01 was sufficient to inhibit the activation of genes associated with genome instability.

TGFβR inhibition represses RGDSP induced K7 expression

After associating nuclear SMAD 2/3 with increased KRT7/K7 expression and demonstrating that A83-01 can mitigate this response under 2D conditions, we investigated the effectiveness of A83-01 supplementation towards inhibiting TGF-β1 mediated K7 expression in RGDSP cultures. Spheroid formation was disrupted in A83-01-treated HA cultures (Fig. 5a). Proliferation, determined by DNA yield, in the HA + A83-01 cultures was reduced by ~ twofold, although a marginal increase in cell proliferation was observed in A83-01-treated RGDSP cultures (Fig. 5b). Although the proliferative properties of A83-01 were not maintained from 2 to 3D, the expression of KRT7 and FN were repressed in A83-01 treated RGDSP cultures to the levels of the DMSO treated HA cultures (Fig. 5c,d).

Figure 5
figure 5

TGFβR inhibition represses RGDSP-induced K7 expression. (a) hS/PCs were cultured in HA and RGDSP hydrogels with A83-01 or the vehicle control (DMSO) for 14 days and visualized with bright field microscopy. (b) dsDNA, indicative of proliferation, was resolved from simultaneous TRIzol RNA/protein extractions using the Quant-iT PicoGreen assay. (c,d,fh) hS/PCs were cultured in HA and RGDSP hydrogels for 14 days receiving A83-01 or DMSO, and expression of KRT7 (c), FN (d), IGF2 (f), KRT7-AS (g), and AMY1A (h) was assessed with qPCR. (e) hS/PCs were cultured in HA and RGDSP hydrogels for 14 days receiving A83-01 or DMSO, and western blotting was conducted to investigate K7 and fibronectin expression. Blotting for fibronectin and K7 was performed sequentially on the same membrane (Fig. S7g–i). Duplicate measurements are presented from technical replicates extracted from separate hydrogels. (i) hS/PCs were cultured for 14 days in RGDSP receiving A83-01 or DMSO, and fluorescent microscopy indicated that expression α-amylase and β-catenin was not suppressed by treatment with A83-01. Single channel images are presented in Fig. S13. Quantification was conducted from 3 independent experiments, each with 3 technical repeats. Error bars represent SEM in all cases. One-way ANOVAs were performed on data presented in (b-h) followed by Tukey’s multiple comparison test. * indicates p < 0.05 in all cases.

Western blot analysis further confirmed that the RGDSP-promoted increase in K7 and fibronectin expression was repressible by inhibition of TGF-β signaling (Fig. 5e). Although A83-01 treated HA cultures exhibited increased KRT7 and FN mRNA levels, increased protein levels of K7 and fibronectin were not detected (Fig. 5e).

On 2D substrates, we established the potential for TGF-β1 to induce IGF2 and KRT7AS expression by hS/PCs. When profiling the entire culture period, a distinct temporal event was found to occur at day 14, when IGF2 was upregulated ~ 30 fold in 3D RGDSP cultures (Fig. S6d). IGF2 expression also correlated (Pearson’s r: 0.999, p = 0.005) with KRT7 expression in RGDSP cultures. However, the expression of the IGF1 or IGF1R genes did not follow the same trend (Fig. S6d–f). As expected, A83-01 suppressed the IGF2 expression stimulated by RGDSP cultures, and IGF2 expression was not activated by inhibiting TGF-β signaling in HA cultures either (Fig. 5f). However, the enhanced expression of the KRT7-AS transcript that was observed in response to TGF-β1 was not induced in the RGDSP cultures, suggesting this mechanism is not required for the observed increased KRT7/K7 expression brought on by RGDSP (Fig. 5g).

qPCR and ICC results indicated TGF-β inhibition led to the repression of K7 and the maintenance of amylase expression (Fig. 5e–i). Furthermore, we found β-catenin remained localized to cell–cell junctions with TGF-β inhibition in RGDSP cultures (Figs. 5i, S13). Thus, we determined that the enhanced proliferation and amylase expression provided by RGDSP were maintained as K7 expression was suppressed with TGF-β signaling inhibition. Collectively we conclude that RGDSP stimulates increased TGF-β1 expression, which in turn induces K7 expression, analogous to the addition of exogenous TGF-β1 supplementation in 2D cultures (Fig. 6).

Figure 6
figure 6

Schematic depiction of TGF-β1 stimulation of K7 expression in hS/PC cultures. (a) In 3D cultures, RGDSP stimulates TGF-β1 secretion to drive K7 expression, which is repressed by the addition of A83-01. (b) In 2D cultures, exogenous TGF-β1 stimulates K7 that is inhibited by the addition of A83-01.


Source link



Please enter your comment!
Please enter your name here

Most Popular

Recent Comments

%d bloggers like this: