Significance of Notch1 Signaling Pathway in Human Pancreatic Development and Carcinogenesis

Abstract

In animal studies, Notch1 signaling pathway plays an important role in the pancreatic embryogenesis by promoting pancreatic progenitor cell self-renewal and exocrine linage development. The persistent activation of Notch pathway could arrest the organ development and keep cells at an undifferentiated stage. Studies have shown that Notch1 signaling pathway is up-regulated in invasive pancreatic ductal adenocarcinoma (PDAC). Here we examined the expression pattern of Notch1 and Hes1 in human fetal pancreas tissues to elucidate the role of Notch1 in human pancreatic embryonic development. We also compared Notch1 expression in tissues from PDAC, chronic pancreatitis and pancreatic intraepithelial neoplasia (PanIN). Our data show that Notch1/Hes1 signaling pathway is activated during early pancreatic embryogenesis and reaches the highest at birth. After pancreas is fully developed, Notch1/Hes1 pathway is inactivated even though Notch1 protein cell-surface expression is up-regulated. We also showed that the expression of both Notch1 and Hes1 are present in 50% (33/66) of PDACs, but not in PanINs. These findings indicate that Notch1 activation is only apparent in late stage of pancreatic carcinogenesis, suggesting that treatment with Notch signaling inhibitors including γ-secretase should be selectively used for PDACs with confirmed Notch1 signaling activation.

Introduction

Notch signaling is an important signaling pathway that has been shown to regulate cell-fate determination during pancreatic embryonic development. The mammalian Notch receptor family has four members, Notch 1–4. Notch is a single-pass transmembrane heterodimeric receptor. In mammals, the canonical Notch activation is initiated by binding interactions between the extracellular domain of a Notch family member on a signal-receiving cell and a Notch ligand from the Delta/Serrate/LAG-2 family on a signal-sending cell. Ligand binding initiates a series of proteolytic events that culminates in the release of the intracellular domain of Notch (NICD) and the formation of a large transcriptional activation complex leading to the activation of downstream target genes^1–4. In mice, activation of Notch signaling pathway is required for embryonic development of pancreatic acini, but the pathway must be turned off at a certain point for the embryonic differentiation of endocrine cell lineage⁵. Persistent activation of Notch1 in embryonic pancreas could arrest both endocrine and exocrine development and trap both cell lineages in an undifferentiated state³. Notch1 is the earliest Notch receptor expressed in the mouse pancreatic epithelium while Notch 2 is expressed relatively later and mainly in branched pancreatic ducts⁶. Several studies have shown that Notch1 induces the activation of downstream target genes including Hes1 and Hey1^{7, 8}. Hes1 is a basic helix-loop-helix (bHLH) transcriptional repressor and functionally inhibits cell differentiation and keeps cells at the precursor stage or proliferative stage. The lack of Hes1 in mice results in the depletion of pancreatic epithelial precursors and excessive differentiation of endocrine cells⁹.

In humans, Notch pathway activation has been shown occur early during pancreatic carcinogenesis¹⁰. Both human studies and mouse model show that Notch pathway components and Notch targets are upregulated in pancreatic ductal adenocarcinomas (PDACs) and their precursors¹⁰. Gamma-secretase is one of the proteases that mediate Notch cleavage and hence the release of NICD and Notch activation. Recently, blockage of Notch activation using γ-secretase inhibitor (GSI) has been explored as a new therapeutic strategy for pancreatic cancer treatment^{11, 12}. Currently, several phase I clinical trials using GSI and gemcitabine have been tested in patients with the advanced stage of pancreatic cancer. The GSIs being tested to treat pancreatic cancers include MK0752 (Merck) and RO04929097 (Roche/Genentech). Recent literature also indicates that both Notch1 and Notch 2 are important members involved in pancreatic carcinogenesis^{11, 13–15}.

Although it is suggested that the activation of the Notch signaling pathway is important in human pancreatic development, most research data relies on the studies of mouse models and pancreatic cells lines. This is the first study that examined the expression of Notch1 and Hes1 in human fetal pancreas harvested from autopsy to study Notch1 signaling pathway in human pancreatic embryonic development. We also re-examined Notch1 and Hes1 expression in human PDACs and paired normal pancreas tissue. In addition, we compared Notch1 and Hes1 expression in chronic pancreatitis, and in various stages of pancreatic cancer development including acinar to ductal metaplasia, pancreatic intraepithelial neoplasm (PanIN) 1, 2 and 3, in order to elucidate the role of Notch1 pathway in the pancreatic ductal carcinogenesis.

Materials and Methods

Cases

Forty-three autopsy cases from 2002–2009, ranging from gestational age 18 weeks to 18 years after birth, were retrieved from the Department of Pathology Archives at University Hospitals Case Medical Center. Formalin fixed, paraffin embedded pancreatic tissue were included in this study: 15 pancreata at gestational age between 17–18 weeks; 15 pancreata at gestational age between 21–24 weeks; and 13 pancreata from 0–3 months old after birth. For the study of PDAC, 66 of formalin fixed, paraffin-embedded blocks were retrieved from patients who had surgical resection for PDAC including 29 females and 37 males, with patient age ranging from 44 to 83 years old. Clinical and pathological data were obtained from detailed chart review, including age, gender, tumor size, tumor location, lymph node metastasis status, and histological subtype of the invasive carcinoma. The H&E-stained slides from each case were examined by light microscopy, and representative sections were selected for the immunostaining. Virtually all PDAC cases are stage II cancers (12 IIB and 54 IIB), based on the American Joint Committee on Cancer (AJCC) manual (http://www.cancerstaging.org/staging/posters/pancreas8.5x11.pdf). Tissues from patients with chronic pancreatitis (n=20) and associated PanIN-1 (n=15) and 2 (n=15) were also selected. PanIN-3 (n=15) were selected from the above PDAC cases.

Immunohistochemical study

Immunohistochemical stains were performed by the diagnostic Immunohistochemistry Laboratory of the University Hospitals. Briefly, unstained 4 μm-sections were prepared from the paraffin blocks and baked for 30 minutes at 60° C in a Boekel Lab oven. The slides were then processed using a Bond Automated Immunostainer (Leica). The slides were deparaffinized with Bond Dewax solution (Leica), and the antigen retrieval was achieved by Bond Epitope Retrieval Solution 1 (Leica), a citrate buffer based (pH 6.0) solution for 20 minutes at 100° C. The slides were incubated in the appropriate primary antibody dilutions for 15 min at room temperature. The Bond Polymer Refine Detection kit (Leica) was used to visualize the slides and subsequently counterstained with hematoxylin. The primary antibodies that were used include rabbit anti-Notch1 monoclonal antibody (clone D1E11, Cell Signaling, #3608, used in 1:100 dilution) and rabbit anti-Hes1 antibody monoclonal antibody (clone EPR4226, Epitomics, # 2922-1, used in 1:200 dilution). For the Notch1 and Hes1 double staining, Notch1 is detected by ? and Hes1 is shown as red nuclear staining using the Bond Polymer Refine Red Detection (Leica).

Notch1 staining was scored based on a 2-dimension system, including the extent of the cells stained and the intensity. The extent of cells stained was scored at a 0–3 scale: 0, (0–4% cells staining); 1, (5–33% cells staining); 2, (34–66% cells staining); 3, (67–100% of cells staining). The intensity was scored at a 1–3 scale: 1, weak; 2, moderate; 3, strong staining. The final score was reached by the product of the extent and intensity. Hes1 was scored based on the percentage of nuclei staining: 0, (0–4%); 1, (5–25%); 2, (26–50%); 3 (51–75%) and 4 (76–100% cell staining). The acinar, ductal, and endocrine compartments are separately evaluated. The scores were independently recorded by 2 pathologists (H.H. and W.X), and the average was used in the final score. The cases are classified as positive if the average score of Notch1 ≥ 2, while the Hes1 ≥1, respectively.

Statistics

The comparison of Notch1 and Hes1 staining among different ages of fetal pancreata, and the comparison among chronic pancreatitis, PanINs and PDAC were evaluated by Mann Whitney U Test (http://elegans.som.vcu.edu/~leon/stats/utest.cgi). Fisher’s exact test was used to compare the Notch1/Hes1 expression in PDACs with different pathological variables due to the small sample size.

The study is approved by the Institutional Review Board (IRB) at University Hospital Case Medical Center.

Results

Notch 1 and Hes1 expression during human fetal pancreatic development

Histologically, at 18 weeks of gestation (n=15), the fetal pancreas starts to form the lobular structure surrounded by highly cellular mesenchyme (Figure 1A). At 21 weeks of gestation (n=15), the fetal pancreas have well developed lobular structure with both distinctive endocrine and exocrine components (Figure 1D). The fetal pancreas is essentially similar to the adult pancreas histologically at 3 months after birth (n=13) (Figure 1G). At 18 weeks of gestation, the expression of Notch1 is mainly identified in acinar cells and centroacinar cells, with a focal membranous and cytoplasmic pattern (Figure 1B). The level of Notch1 expression in the pancreatic acini gradually increases during pancreatic maturation (Figure 1E) and reaches plateau right after birth (Figure 1H). In comparison, pancreatic ducts and endocrine cells do not express Notch1 (Figure 1B, E, and H). Notch1 signaling activation is evidently active in acinar cells and centroacinar cells during the pancreatic embryonic development, which is supported by the nuclear co-expression of Hes1 in these cells (Figure 1C, 1F). However, Notch1 signaling appears to be inactive in the acinar cells derived from pancreas older than 3 months after birth, which show negative nuclear expression of Hes1 (Figure 1I) despite that there is strong membrane staining of Notch1 protein.

Figure 1.

Expression of Notch1 and HES1 during human fetal pancreas development. At 18 weeks of gestation (G18W), the pancreas has primitive lobular structure with cellular mesenchyme (A); At 21 weeks of gestation (G21W), the fetal pancreas has well developed lobular structure with endocrine and exocrine components (D); At 3 month old after full term birth, the histology of pancreas is similar to adult pancreas (G). The Notch1 expression is identified in acinar cells and centroacinar cells. The level of Notch 1 expression increases during pancreatic maturation (B, E, and H). The Hes1 is co-expressed with Notch1 in pancreatic acinar cells at G18W(C) and G21W (F), however, the Hes1 expression is undetectable in the mature pancreas (I).

As summarized in table 1, in acinar cells, the scores of Notch1 immunostaining were much higher in 0–3 months old pancreata than those at 18 weeks and 21 weeks gestational ages (P<0.01 by Mann Whitney U Test); while the scores of Hes1 immunostaining were much lower in 0–3 months old pancreata compared to those at 18 weeks and 21 weeks gestational ages. In developed ductal cells, no significant immunostaining of Notch1 and Hes1 were detected among pancreata at different gestational ages. Overall, the expression pattern of Notch1 and Hes1 in the human fetal pancreas is comparable to that of mouse pancreas. However, it is impossible to evaluate Notch signaling in the earlier stage of human pancreas development due to the unavailability of tissue samples.

Table 1.

Scores of Notch 1 and Hes1 expression in pancreata at different gestational stages

Gestational age	18 weeks (n=15)	21 weeks (n=15)	3 months (n=13)
Acinar
Notch1 (mean)	2.1	3.5	10.6
	P value	p<0.01	p<0.01
Hes1 (mean)	3.6	3.2	0.4
	P value*	p=0.10	p<0.01; p<0.01
Ducts
Notch1(mean)	0.33	0.20	0
	P value*	p=0.26	p=0.06; p=0.17
Hes1(mean)	0.12	0.06	0
	P value*	p=0.37	p=0.26; p=0.10

^*Mann Whitney U Test

Notch1 expression changes during human chronic pancreatitis and pancreatic carcinogenesis

As described above, Notch1 is expressed on the pancreatic acinar and centroacinar cells, but not on mature ductal cells of the developed pancreas. In chronic pancreatitis, the level of Notch1 expression is inversely related to the degree of the damage to the pancreas. The loss of Notch1 expression in the acinar cells starts at the basal side and gradually involves the luminal side (Figure 2A). The Notch1 protein is undetectable in PanIN1, PanIN 2 and PanIN3 (Figure 2B–D). However, Notch1 expression is detectable on 50% (33/66) of PDAC in our study (Figure 2E–F). The expression features of Notch1 in pancreatic cancer cells are of variably luminal pattern. The summary of Notch1 expression from all 66 surgical cases of pancreatic ductal adenocarcinoma is tabulated in the Table 2.

Figure 2.

Notch signaling pathway is re-activated in pancreatic ductal adenocarcinoma. A-F. Immunostain of Notch 1 in chronic pancreatitis, PanINs and PDACs. A. Chronic pancreatitis, with variable reduced Notch1 expression in acinar cells, X100; B-D, PanIN1,2,3, showing mucinous ductal epithelial with no Notch1 expression, X100, X200, X200 respectively; E&F, PDAC with luminal Notch 1 expression, X100 and X200 respectively.

G-I. Double immunostains of Notch1 (brown) and Hes1 (red). G, normal pancreas, showing strong diffuse Notch1 membranous staining with no Hes1 expression; H, PanIN1, with no Notch1 and Hes1 expression; I, PDAC, with luminal (brown) Notch1 expression and focal Hes1 (red) nuclear staining.

Table 2.

Notch1 expression in 66 cases of pancreatic ductal adenocarcinoma

	Positivity (%)	Expression pattern
Normal Pancreas	100	Diffuse membrane
Chronic pancreatitis	100	Variable membrane
PanIN 1	0	NA
PanIN 2	0	NA
PanIN 3	0	NA
Ductal adenocarcinoma	50	Variable, luminal

By double immunostaining of Notch1 and Hes1, normal pancreas have strong membranous Notch1 protein expression in acinar cells (not in ductal cells) but no detectable Hes1 expression (Figure 2G), suggesting that Notch1 is not active in normal pancreas tissues. PanINs have neither Notch1 nor Hes1 expression (Figure 2H). However, PDACs with luminal Notch1 expression all have positive Hes1 expression (Figure 2I).

Our data suggests that 50% (33/66) of PDACs have re-activated Notch1/Hes1 signaling, while Notch1 is dormant in normal pancreas and pancreatic precursor lesions. Demographic variables such as age and gender, as well as pathological variables including tumor size and the presence of lymph node metastasis have no correlation with Notch1 activation in our sampled PDACs, as summarized in Table 3.

Table 3.

Demographic and pathological variables of pancreatic ductal adenocarcinomas with Notch1 pathway activation

	Overall (N=66)	Notch1 activated (N=33)	Notch1 non-activated (N=33)	P value
F/M	29/37	13/20	16/17
Age (means, years)	63.6	62.4	64.8
Tumor size
≤ 2.0 cm	6	4	2
> 2.0cm	60	29	31	*P=0.23
Lymph node
negative	10	4	6
Positive	56	29	27	*P=0.21

Fisher’s exact test

Discussion

Notch signaling is an important pathway involved in embryonic development to cell proliferation and survival in many organ systems, including pancreas. Among 4 members of Notch family protein, only Notch 1 and Notch 2 are involved in pancreatic embryonic development and pancreatic carcinogenesis. In the pancreas, Notch1, Notch2 and Notch4 are mainly present in vascular endothelium, while Notch3 is only found in vascular smooth muscle cells^{6, 16}.

In mice, Notch1/Hes1 pathway is activated early during the pancreatic development, followed by a late Notch 2 activation mainly in branch ductal epithelium, which is considered the source of endocrine and exocrine stem cells¹⁶. In mouse models with Hes1 deficiency, severe pancreatic exocrine hypoplasia due to depletion of pancreatic epithelial precursors is observed, suggesting a critical role of Notch signaling activation through Hes1 in pancreas development⁹. In addition, in mouse pancreas, ectopic Notch activation results in an expansion of undifferentiated ductal precursor cells¹⁰. Therefore, Notch1/Hes1 signaling pathway is important in regulating pancreatic organ development, and hence may be re-activated in carcinogenesis.

In humans, pancreatic embryonic development begins with the forming of dorsal and ventral buds. During the 6^th week of gestation, the ventral bud migrates posteriorly and merges with dorsal bud, resulting in the formation of definitive pancreas¹⁷. The differentiation process goes through two different pathways, corresponding to the exocrine and the endocrine component of pancreas. In our study, we showed that the fetal pancreas starts to form the lobular structure surrounded by highly cellular interstitial tissue at 18 weeks of gestation, and had well developed lobular structure with both distinctive endocrine and exocrine components at 21–24 weeks of gestation. The histology the fetal pancreas was histologically similar to the adult pancreas around birth (after 36 weeks). At 18 weeks of gestation, Notch1 expression is detectable and is widely distributed in acinar cells and centroacinar cells, with a focal membranous and cytoplasmic pattern. Importantly, this pattern of Notch1 expression is accompanied with strong nuclear Hes1 expression, indicating that Notch1 is highly activated at this stage. Although membrane Notch1 staining keeps increasing in density with the advance of gestational age, its downstream molecule Hes1 staining steadily decreases with the maturation of pancreatic organ. Because the monoclonal antibody we used can detect both the full length and the activated NICD, we believe that the paradox increase in the membrane staining of Notch1 in the context of near-absent expression of Hes1 represents inactivation of Notch1 after pancreas is fully developed.

We found that normal human pancreas, regardless of their age, always show diffuse membranous Notch1 expression, but have neither cytoplasmic Notch1 nor nuclear Hes1 expression. This indicates that in normal pancreas, the cell surface Notch1 protein expression is increased but no or little activated form of Notch1 is released and hence Notch1 signaling is not activated. During chronic pancreatitis, the Notch1 pathway is activated with decreased intensity of Notch1 membrane staining but increased cytoplasmic staining, and increased Hes1 expression for regenerative function. In comparison, half of the PDACs show increased expression of both Notch1 and Hes1. The immunostaining of Hes1 in the double staining is universally weaker in intensity when compared to that in the single Hes1 staining. This is likely due to the weakening of the Hes1 staining relative to the Notch1 staining in the presence of counterstaining. Despite of the technical issue, the nuclear signal of Hes1 is intense enough to identify the positive cases from the negative cases. Therefore, our data suggest that Noch1 signaling pathway is active during pancreatic organogenesis, and is inactivated after the organ is developed. Notch1 pathway apparently is re-activated in responding to regenerative demand, and is significantly up-regulated in PDACs, and hence Notch1 may function as a feto-oncoprotein as it is activated during both organogenesis and carcinogenesis. However, it is neither a sensitive nor specific marker for PDAC as only subset of PDAC cases show activated Notch1.

Further, our data suggest that Notch1 is not active in PanIN cases, and the activation of Nocth1 is neither associated with the tumor size nor the lymph node status. Although Notch1 has been reported as a key member involved in the pancreatic carcinogenesis^{10, 18–20}, the role of Notch1 signaling in pancreatic carcinogenesis may be more complex depending on the cell-context and its interacting downstream signaling molecules or pathways. To support this notion, recent studies by Hanlon et al showed that contrary to what is expected, loss of Notch1 in a mouse model of PDAC induced by one oncogenic allele of K-ras (Kras-G12D) in fact promoted the tumor incidence and progression. However, Hes1 and Hey1 expression was not decreased in tumors that lack Notch1 when compared to those with intact Notch1, indicating a Notch1-independent activation of Hes1 or Hey1 may be responsible for the oncogenesis in this mouse model of PDAC²¹. Other studies also suggest that Notch2 but not Notch1 expression was found increased in PDAC²². Furthermore, it was shown by Mzur et al that Notch2, but not Notch1, is the key regulator of PanIN progression and malignant transformation by boosting Myc signaling in Kras-driven pancreatic carcinogenesis¹⁴. However, contradictory to the report from Mazur et al, Hes1 expression was not observed by us in PanIN. The discrepancy may be explained by different antibodies used in different studies. Alternatively, Hes1 is likely up-regulated by Kras G12D independent of either Notch1 or Notch2 activation in mouse models of pancreatic cancer, as suggested by the fact that Hes1 expression is maintained even after ablation of either Notch receptor. Nevertheless, we did not find any case that displays Hes1 expression without Nocth1 expression, supporting that Hes1 is a direct and specific target of Notch1 in human PDACs which show activated Notch1 signaling. Unfortunately the expression of Notch2 as assessed by immunohistochemistry did not yield interpretable data after we tested several commercially available Notch2 antibodies (data not shown). Further characterization of Notch2 in human PDAC tissue samples is warranted in the future in order to elucidate distinctive roles and perhaps cooperative function as well of individual Notch receptor in pancreatic cancer development.

Our data further suggests that Notch signaling may not be universally activated in human PDAC, and therefore Notch inhibitors like GSI may be only effective in a subgroup of PDAC patients with activated Notch signaling. Further study of the relationship between cancer tissue Notch activation levels and their response to Notch inhibitors in patient with advanced pancreatic cancer will be a necessary step for optimal personalized patient care.