Characterization and functional evaluation of goat PDX1 regulatory modules through comparative analysis of conserved interspecies homologs

Abstract

PDX1 is a crucial transcription factor in pancreas development and mature β-cell function. However, the regulation of PDX1 expression in larger animals mirroring human pancreas morphogenesis and endocrine maturation remains poorly understood. Therefore, we conducted a comparative analysis to characterize regulatory regions of goat PDX1 gene and assessed their transcriptional activity by transient transfection of several transgenic EGFP constructs in β- and non-β cell lines. We recognized several highly conserved regions encompassing the promoter and cis-regulatory elements (Area I-IV) at 5’ flanking sequence of the genes. Within the promoter, we identified that a key E-box and nearby CAAT element synergistically drive transcription, constituting the basal promoter of goat PDX1 gene. Furthermore, each recognized regulatory area separately enhances this basal promoter activity in β-cells compared to non-β cells; however, cooperatively, they exhibit a bifunctional regulatory effect on transcription. Additionally, the intact ~ 3 kb upstream region (Area I-III) functions as the most efficient reporter transgene in vitro and shows islet-specific expression in native rat pancreas. Together, our findings suggest that the regulation of goat PDX1 gene is governed by conserved regions similar to other mammals, while both structurally and functionally, these regions exhibit a closer resemblance to those found in humans.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-77614-0.

Introduction

Pancreas organogenesis has been extensively studied in several vertebrate species such as rodents¹, humans^2,3, and small ruminants (sheep and goat)^4–6. In the process of endoderm patterning, inductive signals play a crucial role in subdividing the developing gut into organ-specific areas and influencing the expression of different transcription factors (TFs). These TFs take part in the specification of foregut endodermal epithelium to dorsal (DP) and ventral (VP) pancreatic buds, leading to pancreas formation^7,8. Even though the signaling pathway through which TFs regulate pancreas organogenesis is highly conserved, the timeline of cell lineage specification and functional maturation are different between rodents and large animals⁹. For instance, unlike mice, which undergo three major transition phases during pancreas organogenesis, humans do not have a distinct secondary transition for endocrine differentiation. Instead, human pancreatic development progresses more continuously from the early stages. This smoother progression contrasts with the sharp developmental transitions observed in rodents, highlighting evolutionary differences in pancreatic development³. In human and sheep, functional maturation of insulin-secreting cells mostly occurs before birth, while in rodents it proceeds postnatally^5,10–14. Islet formation in sheep and humans involves epithelial-mesenchymal transition (EMT), marked by the co-expression of the mesenchymal marker vimentin and insulin during pancreatogenesis¹⁵. This co-expression persists into adulthood in a subset of islet β-cells, unlike in rodents where vimentin expression is absent in adult β-cells. This difference suggests a more extended and sustained islet growth and maturation process in large mammals compared to rodents¹⁵. Pancreas development in human, sheep, and mice share conserved molecular mechanisms, however, the expression patterns of key factors show differences across these species. Pancreatic and duodenal homeobox factor 1 (PDX1), a master regulator of pancreas generation, is expressed between 29 and 31 days of gestational age (dGA) in humans and 24–29 dGA in sheep, but it is observed in an earlier embryonic day (E8.5) in mice¹⁵. PDX1 plays a key role in early pancreatic specification, as well as in the maturation and function of islet β-cells. Initially, PDX1 is highly expressed in epithelial progenitor cells throughout the pancreatic epithelium, surrounding gut tube, and central nervous system. However, during the EMT phase, its expression decreases as these progenitor cells transition to a mesenchymal state, allowing them to migrate and cluster together to form islets. In adults, PDX1 expression increases again and becomes primarily restricted to β-cells and some δ-cells¹⁶. Although the high levels of Pdx1 expression induce β-cell specification, reduced expression of this TF is required for acinar cell differentiation and maturation as well¹⁷. Pdx1 knock-out mouse embryos with pancreas deficiency do not stay alive, however, a small dorsal bud with a few glucagon-expressing cells survive without insulin expression^18,19.

Developmental studies show that different levels of Pdx1 expression during pancreas specification and in various cell types are regulated by a conserved proximal promoter and several distal enhancers in the 5’ flanking region²⁰. Pdx1 transcription is controlled by a TATA-less promoter, leading to the utilization of multiple transcription start sites, as reported for humans, mice, and rats. Analysis of reporter constructs driven by the Pdx1 promoter in transgenic mice and cultured β-cells revealed that approximately 6.5 kb upstream of the mouse Pdx1 transcriptional start site contains four cis-regulatory regions, Area I-IV, essential for its regulation²¹. Proximal enhancers including Areas I, II, and III are located between − 2700 and − 1800 bp and control Pdx1 expression during pancreatic development²⁰. Deletion of Area II resulted in serious developmental deficiency and massive failure of endocrine progenitor development. However, the remaining Areas (I, III, and IV) were enough to drive early pancreatic multipotent progenitor specification and complete acinar differentiation²². A complementary study also revealed that the predominant role of Area II in β-cell development is potentiated by Area I, while Area III influences acinar development through PTF1A binding²³. Distal Area IV is located between − 6200 to -5670 bp and is only active in postnatal β-cells. In contrast to mutations in neighboring Areas I-III, Area IV–specific mice knockouts did not disturb embryonic or early postnatal β-cell development, although neurogenin3 progenitors and subsequent insulin and somatostatin expressing cells were reduced²⁴. In addition, targeted demethylation of Area IV significantly decreased the function of Pdx1 enhancer as well as the expression of its target genes²⁵. Identical regulatory regions were also found in the human genome between − 2810 and − 1670 of the PDX1 gene, referred to as PH1, PH2, and PH3²⁶ but ~ 0.5 kb of human PDX1 Area IV element is located farther between − 8656 to -8155 bp²⁷. These regulatory areas are highly conserved in Pdx1-expressing species, but the PH2/Area II domain is absent in chickens²⁸.

DNase I footprinting analysis showed that several trans-activators, including SP1/3, HNF-3b, HNF-1a, MAFA/B, PAX6, NKX2.2, and PDX1, bind to and modulate the transcriptional activity of four conserved Areas. Additionally, potential cis-acting binding sites for various TFs such as bHLH/bHLH-ZIP proteins (E-box), CTF/NF-1 (CAAT), C/EBP, and upstream stimulatory factor (USF) were identified around the proximal promoter sequences. Mutations and deletions in these sequences severely reduced Pdx1 promoter function, while point mutations within the conserved E-box motif abolished specific transcription in β-cells²¹. These findings highlight PDX1 as the master TF for pancreas formation and β-cell function.

To date, a detailed analysis of the literature has revealed a noticeable lack of experimental investigations clarifying the function of PDX1 regulatory regions in large animal models. The improved understanding of the function of these regions in goat, sheep and pig whose fetal physiology, organ size, gestational period and especially β-cell proliferation and differentiation resemble the human^15,29, raise the possibility of using the animals as research models for pancreas biological studies and even appropriate hosts for generation of human organs³⁰.

In the present study, we identified the chromosomal location and the sequence of goat PDX1 promoter as well as its regulatory elements, through a comparative genomic analysis. We also identified five highly conserved regions between human and goat including the promoter sequence and four cis-regulatory areas. Our findings demonstrate that the proximal sequence could efficiently express the EGFP reporter gene in β-cell lines. The basal activity of the promoter is driven by a functional E-box II (-192) and a nearby CAAT box (-213) within this sequence. Moreover, we discuss changes in the transcriptional activity of this basal promoter when cis-regulatory regions are introduced upstream. Although each of these regulatory areas enhanced the transcriptional activity of the promoter, their bi-functional regulatory role was also observed cooperatively.

Results

Comparative genomic analysis of identified goat PDX1 upstream region

As mentioned previously, critical cis-acting transcription regulatory elements of mouse Pdx1 span around the 6.2 kb of the 5’-flanking sequence of its genomic region^20,21. This conserved regulatory region was also characterized in ~ 6.5 and ~ 8.6 kb of the 5’-flanking rat and human Pdx1/PDX1 genes, respectively²⁷. To identify this noncoding regulatory region within the goat genome, approximately 9 kb of PDX1 sequence was obtained from the whole genome shotgun sequence of the Capra hircus (NCBI Reference Sequence: NC_030819.1). This sequence encompasses 8.9 kb of the 5’-flanking region upstream of the transcription start site and continues with the 5’UTR and 0.1 kb related to the ATG start codon (chr12: 54590100–54599100). Comparative analysis of this sequence was conducted using mVISTA with a default identity level of 70% over a 100 bp length and the MLAGAN alignment program. This analysis identified twelve conserved regions within the coordinates of human sequence (Supplementary Table S3). Among them, five sequences were highly similar across rodents (Fig. 1A and Fig. S1). Based on these multiple sequence alignments, VISTA also visualized the genetic diversity between rodents and large animals in the case of PDX1 5’ upstream region by the phylogeny diagram. Rank VISTA analysis indicated that these five noncoding conserved sequences were statistically significant homologous in the alignment (P value < 0.05) and perfectly matched with four identified regulatory areas (Area I-IV) plus the core promoter region of PDX1 (Fig. 1A and Supplementary Table S4).

Fig. 1.

Comparative analysis of goat PDX1 upstream regulatory regions. (A) Schematic diagrams indicate the goat PDX1 chromosome position (upper panel) and comparative analysis of ~ 9 kb upstream regulatory regions among human, mice, and rat (middle panel) by mVISTA. Phylogenic diversity among the species is also displayed in the left panel. (B) Sequence alignment analysis of goat PDX1 upstream regulatory region by UCSC shows several evolutionary conserved sequences among goat, human, and a hundred vertebrates, which are additionally analyzed by some track hubs including EPD, CAGE, and GeneHancer. (C) Transcription initiation region of human PDX1 aligned with several other species and the multiple start sites are reported by FANTOM5 and CAGE (upper panel). The sequencing result of goat PDX1 transcription initiation site also depicts in the bottom panel. (D) Graphical maps of CpG islands within ~ 3.5 kb PDX1 upstream regulatory region of goat, human, mouse, and rat, which are analyzed by MethPrimer browser (right panel) and schematic representation of CpG islands and their interference with regulatory regions (left panel). (E) MatInspector analysis of promoter elements within ~ 1.5 kb (˗1.4 to + 0.1 from the start codon ATG) of goat PDX1 region by Genomatix. Lavender color boxes; four regulatory areas, tan color box; proximal promoter, light blue boxes; CGIs (CpG islands), orange boxes; CEBPs, dark blue boxes; E-boxes, green box; CAAT box, pink boxes; CG-boxes.

Detailed analysis of this sequence by the UCSC Genome Browser revealed a considerable identity among a hundred vertebrate species in the upstream region of their Pdx1 gene. Among studied species, the identity of the goat PDX1 regulatory region with that of human, mouse, and rat was 86.6%, 84.4%, and 83.0%, respectively. A significantly higher score was observed over a longer stretch of the sequence in human BLAT analysis (GRCh37/hg19, Score: 1804, Span: 9704) compared to the best hits in mouse (Score: 557, Span: 4777) and rat (Score: 831, Span: 6151). In accordance with previously reported regulatory sequences of the PDX1 gene, the GeneHancer track set in UCSC analysis shows the interaction between proximal evolutionary conserved enhancers (Area I-III) and PDX1 promoter (Fig. 1B). However, a recent study has revealed a physical interaction between regulatory regions and promoters in islet endocrine cells. This study found that a distal regulatory area (Area IV) also interacts with PDX1 promoter both in β and δ-cells³¹. Hence, these highly homologous sequences were considered potential candidates for goat PDX1 cis-regulatory elements (CREs), which were followed by sequencing analysis of respective PCR products of goat genomic regions and functional analysis. The absence of a consensus TATA box as a transcription initiator sequence in the 5’-flanking region of the goat PDX1 genomic sequence, similar to that seen in humans and rodents, prompted us to analyze the transcription initiation sites for this gene. RefTSS identifies a transcription initiation peak for human PDX1 located 140 bp upstream of the translation start site. This peak has been validated as a major transcription start site through the analysis of mapped reads using FANTOM5 and CAGE. Additionally, several alternative minor transcription start sites upstream of the ATG start codon were identified (Fig. 1C). Although the transcription start sites identified in this study have nucleotide locations that differ from those previously described³², they are all located in the same DNA region and exhibit highly conserved nucleotides across species (Fig. 1C). In this study, mRNAs were collected from a goat fetal pancreatic sample at 30 dGA and reversely transcribed using PDX1-specific primers (Supplementary Table S1). The sequencing results indicate that goat PDX1 mRNAs are transcribed from an initiation site located 209 bp upstream of the translation start codon, with an 89.2% identity to the corresponding human region, which is higher than that observed in rodents. The NCBI BLAST result confirmed that the obtained sequence was highly similar to several predicted PDX1 mRNA sequences and completely matched the reported Ovis aries PDX1 partial coding sequence (Sequence ID: KP059125.1)³³.

CpG islands (CGIs) are typically considered as the sites of transcription initiation and are linked to gene promoters³⁴. MethPrimer browser was used to analyze regions up to 3.5 kb upstream of the PDX1 gene in various species, such as goat, human, mouse, and rat. The borders of CGIs were depicted as light blue boxes in peaks and valleys graphs. We have illustrated these graphical maps in schematics to better illustrate their interaction with regulatory regions. As shown in Fig. 1D, six CpG-rich clusters are distributed throughout the analyzed PDX1 sequences of goat and human, indicating a higher CpG dinucleotide content in the DNA of large animals compared to rodents. Although rats have four CGIs in upstream regulatory region relative to Pdx1 translation start site, the CpG dinucleotides content, particularly those around the promoter are significantly lower. In fact, the increase in the number of CGIs during mammalian evolution is attributed to the rise in chromosome number and recombination rate³⁵, explaining the presence of a higher number of CGIs in the analyzed goat genome in comparison to other species. Subsequent analysis for promoter region identification was performed using Genomatix (Gene2Promoter) to determine the conservation of goat PDX1 sequence with some orthologous promoters via comparative genomic analysis. In addition, MatInspector was utilized to find general core promoter elements such as CEBP, CAAT, GC, and E-boxes. According to Fig. 1E, three CEBPs, one CAAT, two GC-boxes, and six E-boxes were recognized within this promoter region. Some of these elements were consistent with the sequences reported for human and rodent PDX1/Pdx1 promoter sequences²¹. Therefore, about 1.5 kb fragment (from − 1.4 to + 0.1 kb relative to the translation initiation site) was cloned, sequenced, and functionally characterized by transfection into β and non-β-cell lines.

Sequences necessary for β-cell-specific expression of the goat PDX1 promoter

To experimentally delineate the functional sequences in the 5’ flanking promoter region of goat PDX1, a fragment extending about 1.5 kb was cloned and linked to the EGFP reporter gene. Subsequently, a series of 5’ and 3’ deletion constructs were developed as illustrated in Fig. 2. Seventeen constructs were obtained and transiently transfected into BTC3 and MIN6 cells to evaluate their transcriptional activity using flow cytometry (Fig. 3A, B). The results showed that deletion of up to 608 bp from the full-length promoter (PDX-FR2) did not significantly affect EGFP expression in the PDX-FR1 and PDX-V0 constructs (P > 0.05 for PDX-FR1 and P > 0.9 for PDX1-V0 in both cell lines). The AseI-SbfI restriction deletion shortened this sequence from − 735 to -533 bp that contained the CEBP III and II. This deletion led to a substantial enhancement in the reporter gene activity of PDX-V1 construct in BTC3 cells (P < 0.0001), while it had no effect on reporter gene activity in MIN6 cells (P = 0.6149).

Fig. 2.

Schematic representation of goat PDX1 regulatory elements and their cloning process. The position of primers and restriction enzymes used to construct regulatory fragments is shown relative to the ATG start codon.

Fig. 3.

Functional analysis of recognized motifs in goat PDX1 promoter region. (A) A series of 5’ and 3’ goat PDX1 promoter deletion plasmids (left panel) were constructed and transfected to β-cells including MIN6 and BTC3 (right panel). The activity of each construct was measured by flow cytometry 48 h after transfection. Data are presented as mean ± SEM for three independent transfection experiments. ^*P < 0.0021 and ^****P < 0.0001 for BTC3 cells, ^##P = 0.0013, ^###P = 0.0002 and ^####P < 0.0001 for MIN6 cells, ^$$$$P ≤ 0.0001 for regulatory constructs and ns; no significance. (B) Transfected β-cells were imaged using a fluorescence microscope (upper panels) 48 h after transfection and their flow cytometry analysis revealed the percentage and intensity of EGFP expression of each construct (bottom panels). pCMV-EGFP; positive control. PL; Promoter-less construct. Scale bar; 200 μm.

Additional deletion of the CEBP I sequence reduced the promoter activity of PDX-V2 constructs by about 40% and 75% in MIN6 and BTC3 cells, respectively (P < 0.0001). In the following, the removal of E-BOX III did not further reduce the promoter activity of PDX-V3 (P > 0.9 for BTC3 and P = 0.001 for MIN6). However, deletion of CAAT element led to a severe reduction (~ 3-fold decrease) in reporter activity of both BTC3 (P = 0.027) and MIN6 (P < 0.0001). Despite the known importance of E-Box II as a crucial proximal element in PDX1 promoter function²¹, PDX-V4 exhibited minimal reporter expression. Similarly, the transcriptional activity was not affected by additional 5’ deletion of two other tandem elements (GC-box in PDX-V5 and E-BOX I in PDX-V6 constructs), especially when compared to the promoter-less construct with negligible EGFP expression.

In addition to the 5’ deletion analysis, a series of 3’ deletion constructs were generated to investigate which of the examined promoter elements could restore the transcriptional activity of the PDX1 promoter. The PDX-ΔV1, PDX-ΔV2, and PDX-ΔV3 constructs containing the CEBP I-III and E-BOX III elements could not drive the transcription of the PDX1 promoter. In contrast to our previous expectation regarding the CAAT elimination, adding this element did not improve the promoter activity of PDX-V4. Interestingly, this activity was restored when the E-BOX II was located adjacent to CAAT in the PDX-ΔV5 construct. This 10-fold increase (P < 0.0001 for both cell lines) remained unchanged by the promoter sequence expansion up to 50 bp where the GC-box was present in the promoter. The contradictory effect of CAAT and E-BOX II on transcriptional activity was also analyzed by the elimination of these sequences from the full-length promoter. Both individual deletions (PDX-ΔC and PDX-ΔE2) reduced the promoter strength in driving EGFP expression by approximately 70–50% in β-cell lines (P ≤ 0.0001), compared to PDX-FR2, the intact control construct. This data confirms that both elements are required simultaneously for effective regulation of PDX1 expression.

β-cell-specific expression of goat PDX1 through enhancer element sequences

Regulatory areas of the Pdx1 gene have been previously defined as critical cis-regulatory elements for PDX1 expression and thus maturation and maintenance of β-cells²³. These four areas within the goat PDX1 upstream region were initially amplified using specific primers as depicted in Fig. 2. Sequencing of these areas showed a significant sequence identity to proximal and distal enhancer areas (Area I-IV) of the human, mouse, and rat PDX1/Pdx1 genes (Fig. S1). To determine their stimulatory properties, these areas were subcloned directly upstream of the obtained V3 goat PDX1 basal promoter (V3) to generate “pA” plasmids. The β-cell-specific expression of these constructs was evaluated by transfection assay in β (BTC3, MIN6) and non-β (NIH3T3, MCFF) cells (Fig. S2) and the results are presented as the expression ratio of β- relative to non-β-cells (Fig. 4 and Fig. S3).

Fig. 4.

Regulatory properties of goat PDX1 areas. (A) Regulatory areas of goat PDX1 were cloned upstream of core promoter pV3 plasmid (left panel) and transfected to β-cells (MIN6 and BTC3) and non β-cells (MCFF) (right panel). The intact ~ 3 kb upstream region was also cloned and analyzed in combination with and without Area IV. The activity of each construct was evaluated by flow cytometry 48 h after transfection. To normalize the activity of the reporter, relative EGFP expression is calculated by dividing the expression of each construct to pV3, and the β-cell specificity results are presented as the relative activity of each construct in β-cells to MCFF. Data are presented as mean ± SEM for three independent transfection experiments. ^****P < 0.0001 for BTC3 cells, ^####P < 0.0001 for MIN6 cells and ^$$$$P < 0.0001 for BTC3 vs. MIN6 cells, ⁺⁺⁺⁺P < 0.0001 for regulatory constructs and ns; no significance. (B) Images of transfected reporter activity of β-cells using a fluorescence microscope (upper panels) 48 h after transfection and their flow cytometry analysis (bottom panels). Scale bar; 200 μm.

Among four conserved regulatory areas, Area II could drive the highest specific EGFP expression independently in both β-cell lines (P < 0.0001). Area I showed higher activity in β-cells relative to MCFF than Area IV (P < 0.0001) (Fig. 4), while owing to their higher level of expression in NIH3T3 cells, both exhibited equal lower activity (Fig. S3). Area III element directed EGFP expression at a higher level in BTC3 compared to MIN6 cells (P < 0.0001), consistent with earlier findings of human Area III activity in these β-cells²⁸.

In addition to independent analysis of regulatory areas, we assessed their cooperative activity by their different combinations upstream of the V3 core promoter. According to previous studies, Area I/II showed an additive effect in the expression of the reporter gene in the islet β-cells³⁶. In our experiment as shown in Fig. 4, EGFP expression was also effectively driven by pA12V3 compared to other combinations. However, this construct showed lower activity compared to Area II when the expression level was measured relative to non-β-cells (Fig. S2). The addition of either Area III or Area IV to pA12 resulted in pA123 and pA412 constructs, respectively, and led to a decrease in the rate of EGFP expression in MIN6 cells (P < 0.0001 for both construct) (Fig. 4). Nevertheless, combining both of these areas in the pA4123V3 construct restored the initial expression activity of pA12V3 (P > 0.9). To determine whether the physical proximity of Area IV with the intact ~ 3 kb upstream region of goat PDX1 promoter (AR) also contributes to diminishing the promoter activity, this area was cloned upstream of AR fragment (pA4AR). As a result, the addition of Area IV noticeably reduced the EGFP expression level in both transfected β-cell lines (P < 0.0001).

Islet-specific expression of goat PDX1 areas in pancreatic tissue

Previous studies indicated that pancreatic intraductal hydrodynamic injection can be effectively employed for the in vivo evaluation of plasmids³⁷. Thus, we aimed to investigate the expression pattern of pAR plasmid in native pancreatic islets. Initially, the common-bile duct exposure, cannulation, and infusion procedure were optimized using Trypan Blue dye, which gave rise to a uniform distribution of the injected dye in pancreas (Fig. S4A). Moreover, Hematoxylin and Eosin (H&E) staining analysis was performed to examine the impact of intraductal injection on pancreatic tissue damage. Our findings revealed that up to 24 h after injection, the transient expansion of intercellular space returned to the normal size of around 20 μm (Fig. S4B). Next, different injection parameters using pCMV-EGFP plasmid were compared. Isolated islets revealed that an injection volume of approximately 3% of body weight containing about 9 µg plasmid with a flow rate of 1 ml/s was the optimal injection condition with the most islet-targeted reporter expression (Fig. S4C). Employing this injection condition, pAR was used to validate the pancreatic islet-specific EGFP expression. 48 h post-injection, immunofluorescence (IHF) staining of the dissected pancreas verified the expression of EGFP in the islets (Fig. 5A). In addition, EGFP mRNA level was determined in isolated islets by RT-qPCR and revealed that EGFP expression was restricted to the islets of the pancreas but not surrounding tissues (Fig. 5B).

Fig. 5.

Islet-specific expression of goat PDX1 regulatory areas. (A) The pAR-EGFP construct was injected into the male Wistar rat’s pancreas duct. 48 h post-injection the pancreas sections were stained by insulin, EGFP, and DAPI. Scale bars; 200 μm (upper panels) and 20 μm (bottom panels). (B) The relative mRNA levels of EGFP in islets of the pancreas, liver, duodenum, and spleen were analyzed by RT-qPCR in which pAR-EGFP was used as test plasmid and pCMV-EGFP was used as positive control). 18s rRNA was employed as an internal control. Data are presented as mean ± SEM for three independent experiments.

Discussion

In pancreatic development, the coordinated expression of several TFs orchestrates the acquisition of exocrine and endocrine compartments. Mice are frequently employed as models for uncovering the molecular mechanisms behind pancreas formation, however primate islet anatomy and function noticeably differ from rodents⁹. Small ruminants, such as goat and sheep, have advantages due to their resemblance to humans in pancreas morphogenesis^6,15,38,39, and the timing of structural and functional maturation of fetal endocrine cells^29,40. In human and sheep the formation of glucose-sensitive cells and insulin secretion occurs before birth, while in rat’s islet maturation proceeds postnatally²⁹. Similarly, β-cell regulatory factors like PDX1 are expressed later in human and sheep than in rodents^15,41. Therefore, a better understanding of the molecular basis of pancreas organogenesis in comparative animal models seems essential. Lacking information about PDX1 regulatory regions in large animal models, we decided to identify the goat PDX1 regulatory areas within the 5’ upstream region of this gene.

By comparative genome analysis we identified five closely matched non-coding regions with human PDX1, including the promoter and four regulatory areas (Areas I–IV). The level of identity between goat PDX1 regulatory regions and their homologs in human was 90.3%, 94.6%, 94.6%, 88.8%, and 93.4% for the promoter and Areas I to IV, respectively, which was higher compared to rodents²⁷. Furthermore, within these regulatory regions, we observed a significant overlap of CpG islands between human and goat, indicating more similar tissue-specific methylation dynamics. Research has found that the PDX1 promoter has low methylation levels in beta cells whereas in alpha cells, it is highly methylated⁴². This promoter-associated methylation status is increased in the islets of diabetic patients⁴³ and pancreatic neuroendocrine tumors⁴⁴. In addition, in an aberrant metabolic intrauterine environment, Pdx1 is methylated at several CpG sites and permanently silenced, a state that persists after birth and leads to the onset of diabetes in adulthood⁴⁵. Despite different epigenetic regulation in rodents due to lower CGI density, human and small ruminants likely undergo similar intrauterine growth retardation and epigenetic changes in the PDX1 promoter during development⁴⁶.

As shown in Fig. 1E, we observed that the goat PDX1 promoter, similar to other TATA-less promoters, contains multiple G/A and G/C sequences along with cis-acting elements such as CEBP, CAAT, GC-, and E-boxes^{32, 47–49}. To investigate the significance of these elements in promoter activity, we generated a series of deletion constructs. Removal of identified elements from the 5’ end affected promoter activity, but deletion of the CAAT box led to a substantial decrease in EGFP expression in β-cells. Further deletion of adjacent E-BOX II did not cause an additional reduction in reporter expression. These observations differ from previous studies that suggested the deletion of E-BOX II as a key element resulted in a significant reduction in mouse Pdx1 promoter activity²¹. In our analysis of 3’ deletion constructs, the CAAT box alone was insufficient to drive transcription, but the addition of E-BOX II restored promoter activity. Moreover, elimination of each sequences from the full-length promoter resulted in a significant decrease in reporter expression, indicating that both CAAT and E-BOX II are essential for the basal activity of the goat PDX1 promoter. In TATA-less gene regulation, the CAAT box serves as a binding site for nuclear factor Y^50–52. NFY can either activate or suppress transcription depending on the promoter context⁵³. The stimulatory mechanism of NFY may rely on its suppressive association with USF-1, which binds to the E-box sequence⁵⁴. Alternatively, the USF1/PDX1 complex auto-regulates the PDX1 promoter in β-cells, but this feedback loop can be disrupted by SREBP-1c and protein kinase CK2^48,49. These findings suggest that competition between activators and repressors at PDX1 promoters regulates its expression.

In addition to promoter regulations, transactivation of conserved 5′ cis-regulatory areas is essential for Pdx1 expression, driven by islet-enriched TFs^55–57. Therefore, we investigated how the goat PDX1 regulatory areas affect gene expression in β-cell lines. It has been shown that enhancers respond differently to various types of promoters⁵⁸. Accordingly, goat PDX1 areas were sub-cloned directly upstream of the acquired basal promoter (V3). The results from transient transfection of β-cell lines have demonstrated that all these areas independently enhanced the transcriptional activity of the basal promoter (Fig S2). Among them, Area II exhibited the highest level of expression, which aligns with its known pivotal role in driving Pdx1 transcription²². We also observed subsequent levels of expression in Areas I, emphasizing the functional interactions between Areas I and II in inducing robust Pdx1 expression in β-cells³⁶. In our study, modest reporter expression observed in goat Area III and IV, highlighting their similar regulatory properties in mature β-cells. These regions harbor analogous binding sites for factors that contribute to pancreatic fate determination and β-cell lineage specifications such as FOXA⁵⁹, PDX1⁶⁰, PTF1A^28,61, and EGR1⁶². Area III predominantly regulates Pdx1 expression throughout the pancreatic bud during early development via PTF1A⁶¹. Stage-restricted behavior of Area IV was verified by an enhancer-reporter assay during the definitive to pancreatic endoderm transition. This enhancer is primed by acquiring FOXA2, chromatin remodeling, and H3K4me1 deposition, sequentially⁶³. Subsequent activation of this enhancer occurs through the recruitment of PDX1 and histone acetylation during the primitive gut tube stage⁶⁴. Additionally, experiments showed that low Pdx1 expression directed by Area IV is sufficient for acinar cell development and gut/stomach cell specification independent of other areas⁶⁵. Area IV-specific knockout had a minor effect on pancreatic development but reduced Ngn3 + cells, leading to a smaller islet β-cell area²⁴. Therefore, the spatiotemporal activity of Areas III and IV seems to be influenced by their epigenetic signature and the stage of differentiation^61–64. To understand the interactive behavior of these regulatory modules, we placed them upstream of the Area I-II segment (A12), as the most transcriptionally active enhancer complex in β-cell lines. Addition of Area III or Area IV significantly decreased the β-specific activity of the A12 construct, suggesting their involvement in the dose-dependent regulation of PDX1 as they harbor motifs for both lineage and signal-dependent transcription factors⁶⁶. This is supported by studies showing that some regulatory elements can function as both enhancers and silencers depending on the cellular context⁶⁷. Accordingly, we integrated the distant Area IV upstream of the ~ 3 kb goat PDX1 sequence, resulting in a marked reduction of β-cell selective enhancer activity compared to non-β cells. This proximity may eliminate several regulatory regions within the native genomic context, the 3D architecture of enhancer clusters, and the structure of the adjacent islet-specific LncRNA (PLUTO)⁶⁸. Notably, in a study conducted by Chiou et al., physical interactions between regulatory regions and promoters in primary islets were revealed. Their data showed that the PDX1 promoter in alpha cells links with Areas I-III, while in beta and delta cells, it is controlled by Areas IV and I-III³¹.

The present study confronts certain limitations. Goats are important species both in agricultural and biomedical studies⁶⁹, yet their current reference genome is incomplete⁷⁰ and RNA sequencing is limited to a few numbers of tissues that have been annotated using automated computational analysis. The inadequate comprehensive bioinformatic studies and motif discovery tools has restricted our ability to predict TF binding sites within the PDX1 regulatory region. Future research could address this limitation by creating binding-defective mutants for each TF site and analyzing their regulatory function in detail. The enriched transcription factors in β-cells also influence the specific activity of PDX1 regulatory regions⁶⁰. Despite sharing a similar T-antigen origin, MIN6 and BTC3 cells exhibit noticeably distinct G1/S expression patterns⁷¹, different levels of glucose-stimulated insulin secretion⁷², diverse ability to form pseudo islets⁷³, and different expression of β-cell specific TFs. Nevertheless, MIN6 cells widely used in pancreatic research⁷³, provided more reliable regulatory expression data in our experiments. On the other hand, the NIH3T3 cell line, often considered as a non-beta cell line, displayed a similar goat PDX1 expression pattern to the applied β-cells (Fig. S2, S3). Likewise, mouse Pdx1 regulatory constructs exhibited a high level of activity in NIH3T3 cells (Fig. S5). As a consequence, the ability of NIH3T3 cells to express PDX1 (74) could inaccurately alter the β-cell specificity of regulatory fragments. Hence, it seems that the functional activity of PDX1 regulatory areas could be more precisely discriminated in MIN6/MCFF cells. In future research, this challenge could be tackled by utilizing primary goat β-cells along with transgenic studies which are the ideal method to unravel how these enhancers operate in a multi-cellular context.

Conclusion

Taken together, our study demonstrates that varying thresholds of PDX1 activity, particularly in relation to the regulatory regions, are crucial for optimal β-cell function. By identifying and analyzing the promoter and non-coding regions of the goat PDX1 gene, we have emphasized the vital role of cis-regulatory elements and their complex interactions in driving β-cell-specific expression. Notably, the conserved nature of these regulatory areas and their functional similarities to human counterparts suggest that goats serve as a valuable model for understanding pancreatic organogenesis and diabetes pathogenesis. However, future studies employing primary β-cells and transgenic models will be crucial in elucidating the precise regulatory dynamics of PDX1.

Materials and methods

Bioinformatics studies in characterization of 5’ flanking regulatory regions of goat PDX1 gene

The sequence of the characterized goat PDX1 gene including the 5’ flanking region was obtained from NCBI’s reference sequence database (NC_030819.1) and blasted with the UCSC genome browser on human (GRCh37/hg19). To locate candidate orthologous regulatory regions, the obtained sequence was also aligned with genomes of related species by VISTA⁷⁵ (Table S3, S4). Pdx1 core promoter sequences from other species were achieved by the Gene2Promoter database and compared with the goat sequence. Further analysis of this sequence using MatInspector⁷⁶ depicts core binding sites in promoter elements. MethPrimer browser⁷⁷ was also employed to identify the CpG islands.

Amplification of 5’ flanking regions of goat PDX1 gene

In this experiment, different sets of primers were designed and used to amplify the characterized PDX1 5’ flanking regulatory regions (Table S1) from goat genomic DNA isolated by DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany). As depicted in schematic Fig. 2, the 2.8 kb fragment encompassing Area I, II, and III; spanning the sequences from − 2779 to + 102 relative to the ATG start codon and proximal core promoter region was amplified using PDX-F and A1-R primers. Within this fragment, the ATG start codon mutation was induced by site-directed mutagenesis (ATG/A-F, ATG/A-R). This fragment was utilized as a PCR template as well as a digestion sample to develop other promoter regions. Similarly, proximal regulatory areas (Area I, Area II, and Area III), in addition to distal Area IV, were also amplified using specific primers. The integrity of all fragments was validated by sequencing and restriction digestion using appropriate enzymes (Thermo Scientific, Waltham, MA). Pfu DNA polymerase (Thermo Scientific) was utilized for all PCR reactions and acquired products were then cloned into pTZ57R/T vector (Takara, Kusatso, Japan) after treatment with Ex Taq DNA polymerase (Takara, Kusatso, Japan).

Construction of reporter plasmid

The Luciferase reporter gene in pGL4.10 plasmid (Promega, Madison, WI, USA) was removed and replaced with EGFP fragment excised from pEGFP-C1 (Clontech Laboratories, California, United States) to construct a basic plasmid (pGL4-EGFP). PDX1 promoter fragments were ligated upstream of the reporter gene in pGL4-EGFP plasmid (PL: Promoter-less construct), and a series of deletion constructs were generated. Regulatory areas were also cloned into pDXV3, constructing regulatory plasmids. In parallel, pGL4-CMV-EGFP (pCMV-EGFP) was constructed as a positive control plasmid.

Cell culture

Mouse insulinoma β-cells (MIN6 and BTC3) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 4.5 g/L glucose, supplemented with 15% horse serum, 2.5% heat-inactivated bovine serum, 100 U/ml penicillin-streptomycin, and 2 mM L-Glutamine. NIH3T3 and primary male caprine fetal fibroblasts (MCFF) were cultured in DMEM with 4.5 g/L glucose supplemented with 10% FBS and 100 U/ml penicillin-streptomycin. All cell lines were provided by Royan Institute (Esfahan, Iran) and culture medium and other reagents were purchased from Gibco (Waltham, MA).

Cell transfection

When the cells reached approximately 80% confluence in 24-well plates, they were transfected using 1.5 µl Lipofectamine LTX (Thermo Scientific) according to the manufacturer’s recommendation with 500 ng of pDNAs in 100 ul Opti-MEM I reduced-serum medium (Gibco). 48 h after transfection, EGFP expression was evaluated using an inverted fluorescent microscope (IX71; Olympus, Tokyo, Japan), and the images were acquired with a DP72 digital camera (Olympus).

Flow cytometry

48 h post-transfection, about 5 × 10⁵ transfected cells were detached and resuspended to 500 µl PBS and analyzed by flow cytometer. The mean fluorescence intensity (MFI) was measured by a BD FACSCalibur and data were analyzed with Cell Quest Pro software (Becton-Dickinson, Mountain View, USA).

Animals

10 male Wistar rats (10–12 weeks old with an approximate weight of 300 ± 20 g) were provided by Royan Institute (Esfahan, Iran). All methods were conducted in accordance with the relevant guidelines and regulations. Furthermore, all animal protocols were approved by the Institutional Ethical Committee of the Royan Institute (IR.ACECR.ROYAN.REC.1399.107), which are in compliance with the ARRIVE guidelines.

Intraductal injection

Rats were euthanized under the combined administration of xylazine (5 mg/kg body weight per rat) (Bioveta, Czech Republic) and ketamine (75 mg/kg body weight per rat) (Rotexmedica, Germany). Following euthanasia, a preliminary step was taken to ensure proper positioning and stabilization of the animal prior to the surgical intervention. An incision was made in the abdomen to expose the common bile duct (CBD). The highest part of the CBD was clamped to prevent the leak of DNA solution injected into the gallbladder and liver. A 22-gauge angio-catheter was inserted into CBD through the sphincter of Oddi and another clamp was placed on the injection site to restrict the flow of the injected solution into the duodenum. After the injection, clamps were removed and the injection site of the duodenum was sealed with a medical tissue adhesive (Epiglu, Meyer-Haake GmbH, Germany) and the incision site was sutured. As a primary test, 5% trypan blue saline solution was injected to detect the spreading area of the injected solution. Also, to investigate the optimal injection parameters including dose, injection flow rate, and time for tissue harvest, different amounts of pCMV-EGFP ranging from 1 to 20 µg/ml were injected for 5 and 10 s and animals were studied at 24 and 48 h after injection.

Islet isolation

48 h post-injection, rats were sedated by treatment with CO₂, and a laparotomy was performed. The pancreata were exposed and perfused with 10 ml collagenase IV solutions (Sigma-Aldrich, St. Louis, MO) through the common bile duct. The distended pancreata were removed and incubated in a water bath for 15 min at 37 °C with shaking. Once digested, tissues were immediately chilled on ice and 10 ml ice-cold STOP solution was added to each sample. Sand-like disrupted tissues were poured through the 500 μm mesh and centrifuged in a swinging-bucket centrifuge at 250 g for 2 min. Pellets were re-suspended in 5 ml Histopaque-1077 (Sigma-Aldrich) and centrifuged at 1750 g for 20 min. The islets were handpicked and washed 2 times with HBSS (Gibco).

Reverse transcription quantitative PCR (RT-qPCR)

Total RNA was isolated from 10 mg of harvested tissues using TRIzol reagent (Sigma-Aldrich). 1 µg of the total RNAs were treated with DNaseI (Thermo Scientific) and then cDNA synthesis was performed by PrimeScript 1st strand cDNA Synthesis Kit (TaKaRa), both according to manufacturer’s protocols. RT-qPCR was conducted by Applied Biosystems StepOnePlus Real-Time PCR System (Thermo Scientific) in 10 µl reactions using the SYBR Green Master Mix (Takara) and rEGFP, rPDX1, and 18s rRNA primers (Table S2), designed by Beacon Designer 7.9 software (Premier Biosoft International, Palo Alto, CA, USA). All RT-qPCR reactions were performed in triplicate. Relative fold changes in gene expression were normalized to 18s rRNA as internal control and calculated using the 2^−ΔΔCt method.

Histological analysis

Pancreatic tissue samples isolated from injected animals were fixed in 4% (v/v) paraformaldehyde, paraffin-embedded, and cut to 6 mm in length. Prior to staining, sections were deparaffinized and rehydrated. Hematoxylin and Eosin (H&E) staining was accomplished following routine protocol. The pancreatic cell distances were measured through Image J software (National Institutes of Health, Bethesda, MD). For IHF staining, sections were permeabilized with 0.2% Triton X-100 for 5 min, blocked in 10% host serum in PBS⁻ for 1 h, and then incubated with primary antibodies (rabbit anti-GFP, 1:1000, ab290, Abcam and goat anti-insulin, 1:200, sc-7839, Santa Cruz) at 4 °C overnight. Conjugated secondary antibodies were used for fluorescent detection (goat anti-rabbit FITC, 1:2000, F1262, Sigma-Aldrich and donkey anti-goat Alexafluor 594, 1:1000 A11057, Invitrogen). DAPI dye solution with 1 µg/ml concentration (D-8417, Sigma-Aldrich) was also employed to detect nuclei in immunofluorescent images, and all of them were collected on a fluorescence microscope (BX51, Olympus) equipped with a DP70 digital camera (Olympus).

Statistical analysis

Statistical significance was determined by Student’s t-test and two-way ANOVA using GraphPad Prism (version 8.0.0 for Windows, GraphPad Software, San Diego, California USA), in which P < 0.032 was considered significant, respectively. All data are presented as mean ± standard error of the mean (SEM) for three independent experiments.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

CAGE

Cap analysis of gene expression

CBD

Common bile duct

CDS

Coding sequence

C/EBPs

CCAAT/enhancer binding proteins

CGIs

CpG islands

CREs

Cis-regulatory elements

dGA

Days of gestational age

DP

dorsal pancreatic bud

E-box

Enhancer box

EGR1

Early growth response factor 1

EMT

Epithelial-mesenchymal transition

FANTOM5

Functional Annotation of the Mammalian Genome

FOXA

Forkhead Box A

H&E

Hematoxylin and Eosin

HNF

Hepatocyte Nuclear Factor

IHF

Immunofluorescence

IUGR

Intrauterine growth retardation

MAF

Musculoaponeurotic fibrosarcoma

MCFF

Male caprine fetal fibroblast

MFI

Mean fluorescence intensity

NF-Y

Nuclear factor Y

NGN3

Neurogenin-3

Nkx

Homeobox family of NK genes

PanNETs

Pancreatic neuroendocrine tumors

PAX

Paired box

PDX1

Pancreatic and duodenal homeobox factor 1

PTF1a

Pancreas associated transcription factor 1a

SP

Specificity protein

SREBP-1c

Sterol regulatory element-binding protein-1c

UCSC

University of California, Santa Cruz

USF

upstream stimulatory factor

VP

Ventral pancreatic bud

Author contributions

“N.R. Conceptualization, Methodology, Formal analysis, Investigation, Validation, Visualization, Software, Writing-original draft; K.D. Project administration, Resources, Supervision, Validation, Writing-review & editing; A.K.E. Methodology, Validation; S.M. Methodology, Validation; M.R. Methodology, Validation; F.J. Investigation, Validation; M.H.N.E. Funding acquisition, Investigation, Supervision, Writing-review & editing. “.

Data availability

All data generated or analyzed during this study are included in this published article, its supplementary information files, and publicly available repositories. The datasets generated during the current study are also available in the GenBank repository with the accession numbers of PQ066235, PQ066236 and PQ066237.

Declarations

Competing interests

The authors declare no competing interests.

Ethics statement

The animal study protocol was approved by the Institutional Ethical Committee of the Royan Institute (IR.ACECR.ROYAN.REC.1399.107), which is in compliance with the ARRIVE guidelines. Consent to participate is not applicable.