Activation of Dll4/Notch Signaling and Hypoxia-Inducible Factor-1 Alpha Facilitates Lymphangiogenesis in Lacrimal Glands in Dry Eye

Purpose By using hypoxia-inducible factor-1 alpha conditional knockout (HIF-1α CKO) mice and a dry eye (DE) mouse model, we aimed to determine the role played by delta-like ligand 4 (Dll4)/Notch signaling and HIF-1α in the lymphangiogenesis of lacrimal glands (LGs). Methods C57BL/6 mice were housed in a controlled-environment chamber for DE induction. During DE induction, the expression level of Dll4/Notch signaling and lymphangiogenesis in LGs was measured by quantitative RT-PCR, immunoblot, and immunofluorescence staining. Next, lymphangiogenesis was measured after Dll4/Notch signal inhibition by anti-Dll4 antibody or γ-secretase inhibitor. Using HIF-1α CKO mice, the expression of Dll4/Notch signaling and lymphangiogenesis in LGs of DE-induced HIF-1α CKO mice were assessed. Additionally, the infiltration of CD45+ cells in LGs was assessed by immunohistochemical (IHC) staining and flow cytometry for each condition. Results DE significantly upregulated Dll4/Notch and lymphangiogenesis in LGs. Inhibition of Dll4/Notch significantly suppressed lymphangiogenesis in LGs. Compared to wild-type (WT) mice, DE induced HIF-1α CKO mice showed markedly low levels of Dll4/Notch and lymphangiogenesis. Inhibition of lymphangiogenesis by Dll4/Notch suppression resulted in increased CD45+ cell infiltration in LGs. Likewise, CD45+ cells infiltrated more in the LGs of HIF-1α CKO DE mice than in non-DE HIF-1α CKO mice. Conclusions Dll4/Notch signaling and HIF-1α are closely related to lymphangiogenesis in DE-induced LGs. Lymphangiogenesis stimulated by Dll4/Notch and HIF-1α may play a role in protecting LGs from DE-induced inflammation by aiding the clearance of immune cells from LGs.


Introduction
Dry eye (DE) is a highly prevalent ocular inflammatory disorder affecting millions of people worldwide. However, disparities in the definition, diagnostic criteria, and treatment guidelines of the condition suggest that DE is a complicated heterogeneous disease involving many different pathophysiologic mechanisms. [1,2] Although most DE patients complain of discomfort on the ocular surface area, the lacrimal gland (LG) is a major target organ of DE pathogenesis for both non-Sjögren DE and Sjögren syndrome. [3,4] Inflammatory cytokines, inflammatory cells, and matrix proteases were upregulated after DE stress in human and mouse LGs. [5][6][7] Despite the importance of LGs and inflammation in DE pathophysiology, the exact mechanisms underlying increased inflammation in LGs affected by DE remain unknown.
According to previous studies, inflammatory conditions gave rise to new lymphatics extending into the cornea despite its immune privilege. [8,9] Function-wise, lymphatics in cornea may facilitate the exit of antigen-presenting cells and antigenic material from the cornea to regional lymph nodes, thus promoting the induction of an adaptive immune response. [10] Similar to DE induced cornea, we found an increase of lymphatic vessels (LVs) in the LGs of a DE-induced mouse model. [11] By using immunofluorescence staining as well as immunoblot, upregulation of a well-known marker related to LV formation, lymphatic vessel endothelial hyaluronan receptor 1 (LYVE-1), was observed in LGs after DE stress. [11] Nonetheless, the underlying molecular mechanism for lymphatics growth in LGs and their functional role in the development of DE pathology have not yet been investigated.
Notch has been identified as an important factor for lymphangiogenesis interacting with hypoxia-inducible factor-1 alpha (HIF-1α). [12][13][14] Notch signaling performs diverse functions mediated by Notch receptors (Notch 1 -Notch 4), Delta-like ligands (Dll1, Dll3, Dll4), and Jagged ligands (Jagged 1 and Jagged 2). [15,16] A previous research proved that the blocking Notch signaling pathway reduced LV sprouting during early postnatal development of wound healing in mouse dermis. [17] Moreover, conditional inhibition of Notch gene produced a disruption of normal ocular surface homeostasis, implying an important role played by Notch in the development of ocular surface disorders. [18] The purpose of this study is to investigate DE-induced lymphangiogenesis in LGs, focusing on Dll4/Notch signaling and its relationship to HIF-1α activation by using mouse DE model.

Animal treatment and DE induction
Six-to eight-week-old male C57BL/6 mice (Charles River Laboratory, Wilmington, MA) were used in accordance with the standards set forth in the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. The research protocol was approved by the Institutional Animal Care and Use Committee of the Yonsei University College of Medicine. DE was induced in the mice by placing them in a controlled-environment chamber (CEC). To achieve maximum ocular surface dryness, mice in the CEC (with a relative humidity below 13%) were given subcutaneous injections of 0.1 mL scopolamine hydrobromide (5 mg/mL; Sigma-Aldrich Chemical Co., St. Louis, MO) three times a day.
To generate HIF-1α conditional knockout (CKO), mouse mammary tumor virus (MMTV)-Cre mice (The Jackson Laboratory, Bar Harbor, ME) and HIF-1α floxed mice (The Jackson Laboratory) were used. The viral MMTV promoter directed the expression of Cre recombinase in the secretory epithelium of mammary glands, salivary glands, and LGs. The down regulation of HIF-1α in LGs was confirmed by immunoblot and IHC staining. The detailed methods for generating HIF-1α CKO have been described in our earlier work. [11] Tissue preparation After 10 days of DE induction, the mice were sacrificed and LGs were collected. Each tissue was halved and was either fixed in 3.7% paraformaldehyde for immunofluorescence staining or stored at -70°C for quantitative Real-Time-PCR (qRT-PCR) and immunoblotting.
Tissue RNA extraction and qRT-PCR RNA was isolated using an RNeasy Micro Kit (QIAGEN, Valencia, CA) from mouse LGs, and reverse transcription was performed using a Superscript III Kit (Invitrogen, Carlsbad, CA). Real-time qPCR was performed using SYBR1 Premix Ex Taq (Takara Bio Inc., Otsu, Japan) with a StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA). Preformulated primers were used in order to evaluate mRNA expression. Detailed primer information is described in Fig A in

Statistical analysis
The statistical analysis of more than three groups was performed by One-way ANOVA. As post hoc analyses, the subgroups were analyzed by Newman-Keuls method, a stepwise multiple comparisons procedure used to identify sample means that are significantly different from each other. Additionally, Student t-test was performed to compare the two group samples. Dunnett's test was also used to compare each treated group to the control group. A p-value of < 0.05 was considered to indicate a significant difference.

DE induction upregulates Dll4/Notch signaling in LGs
The expression level of Notch signaling was evaluated during DE induction. The qPCR results showed a significant increase of NOTCH1 and NOTCH2 expression at Day 10 (NOTCH1: mean 6.5 fold change, p = 0.034; NOTCH2: mean 5.1 fold change, p = 0.045). The mRNA level of Dll4, Jagged (JAG) 1 and JAG2 increased with significance at Day 10 (Dll4: mean 2.2 fold change, p = 0.032; JAG1: mean 1.5 fold change, p = 0.043; JAG2: mean 1.69 fold change, p = 0.041). The mRNA level of NOTCH1 and Dll4 at Days 2, 4, 6, 8, and 10 was measured. NOTCH1 and Dll4 expression started to rise at Day 2 (NOTCH1: p = 0.022; Dll4: p = 0.034). NOTCH1 reached the peak at Day 4 and declined significantly at Day 8. For Dll4, the mRNA level peaked at Day 6, with significant downregulation starting on Day 8 ( Fig 1B). Immunoblot and densitometry of NOTCH1 showed peaks at Day 6 and Day 10, while Dll4 showed the highest level at Day 8 ( Fig 1C).

DE induction upregulates lymphangiogenesis in LGs
To detect the level of LV and BV formation in LG in DE, the mRNA level of LYVE-1, PECAM, VEGF-C, VEGF-D, and VEGFR3 was measured. LYVE-1 is a powerful indicator for LVs, and PECAM is a representative marker for angiogenesis. These markers were used to assess the effects of DE stress on LV development and angiogenesis. Interestingly, the mRNA level of LYVE-1 showed biphasic humps, peaking twice at Day 4 and Day 10. A significant drop from Day 6 to Day 8 was observed (p = 0.028). From Day 8 to Day 10, LYVE-1 increased significantly with a mean 11.2 fold change (p = 0.045). After the mRNA level of PECAM showed an increase at Day 2 of DE induction (p = 0.245), there was significant downregulation of PECAM from Day 4 to Day 6 (p = 0.008), reaching the lowest level at Day 8 (Fig 2A).
The change of expression of VEGF-C, VEGF-D, and VEGFR3 was evaluated at Day 10 by qPCR. All three were upregulated with a mean 2.1 fold increase for VEGF-C (p = 0.036), a 7.3 fold increase for VEGF-D (p = 0.029), and a mean 1.8 fold increase for VEGFR3 (p = 0.047) ( Fig 2B). Conclusively, immunofluorescence staining demonstrated an increased expression of LYVE-1 + cells at Day 10 ( Fig 2C).   LYVE-1 by GSI and anti-Dll4 antibody was confirmed by immunoblot and immunofluorescence staining (Fig 3B and 3C).

Downregulation of Dll4/Notch signaling and lymphangiogenesis in DEinduced LG of HIF-1α CKO mice
HIF-1α has been regarded as an important regulator for lymphangiogenesis in mouse models. [14] Additionally, we have previously discovered that HIF-1α CKO mice exhibit reduced expression of lymphatics. [11] Therefore, a study for demonstrating the interaction between HIF-1α, Dll4/Notch 1 signaling, and lymphangiogenesis in DE LGs was performed.

CD45 + cell infiltration increases after inhibition of lymphangiogenesis in LGs in DE
To study the role played by lymphangiogenesis in immune cell infiltration in LGs, the change of CD45 + cells in LGs during DE induction was measured using IHC staining and flow cytometry. Among inflammatory cell markers, CD45 antigen (leukocyte common antigen) was chosen because it is universally expressed in almost all hematolymphoid cells, including T lymphocytes, B lymphocytes, granulocytes, monocytes, and macrophages. [21] Therefore, the analysis of CD45 + cell infiltration in LGs may provide an understanding of the overall change of inflammatory status of DE induced LGs.
During DE induction, CD45 + cells increased, reaching the highest peak at Day 6. At Day 10, CD45 + cells decreased drastically (p = 0.00085), reaching almost the same level as the control group (Fig 5A & 5B). According to the flow cytometry data, the actual percentage of CD45 + cells changed from 3.5% (Day 0) to 3.8% (Day 10) when DE was induced in WT mice (data not shown).

Effects of Dll4/Notch signaling on lymphangiogenesis in LGs in DE
The results of this study proved Dll4/Notch-induced lymphangiogenesis in DE LGs by qPCR, immunoblot and immunofluorescence staining (Figs 1-3). Similar to our results, Notch induced lymphangiogenesis has been reported by several previous works. Fatima et al. demonstrated the downregulation of VEGF-C and VEGFR3 in Notch-1 mutant lymphatic endothelial cells. [22] In addition, Niessen et al. manifested the downregulation of VEGF-C/VEGFR3 signaling by blocking Dll4/Notch 1 pathway in mouse dermis model. [17] Therefore, these previous results support the lymphangiogenic role of Dll4/Notch signaling as it has been demonstrated in this study.
Meanwhile, we could not determine which Notch receptors and ligands are exactly responsible for lymphangiogenesis from the present work. However, among the members of the Notch family, the mRNA levels of NOTCH1, NOTCH2, Dll4, Jagged1, and Jagged 2 were increased by DE stress. Yet, the mRNA fold increase of NOTCH1 and Dll4 was significantly higher than other Notch receptors and ligands (Fig 1A). Immunoblot assay also revealed high expression of NOTCH1 and Dll4 after DE stress (Fig 1C). Therefore, Dll4/Notch 1 may be the main Notch subtype in inducing lymphangiogenic pathways. Moreover, earlier studies have pointed to Dll4/Notch 1 signaling in particular as the main subtype of Notch signaling related to lymphangiogenesis. [17,22,23] As opposed to the findings of previous studies where Dll4/Notch 1 signaling mainly regulated VEGF-C/VEGFR3 expression, the results of this study show that VEGF-C is subtly affected by the inhibition of Notch signaling by GSI or anti-Dll4 antibody. Per contra, VEGF-D was inhibited with good significance (Fig 3A). These results imply that Dll4/Notch 1 regulates VEGF-D/VEGFR3 expression rather than VEGF-C/VEGFR3 expression in the process of DE-induced lymphangiogenesis.
Another interesting finding is that the mRNA levels of LYVE-1 showed biphasic peak, at Day 4 and Day 10 of DE induction (Fig 2A). However, by the treatment of GSI or anti-Dll4 antibody, LYVE-1 expression was inhibited only at Day 10 (Fig 3). Although we could not unveil which cells were responsible for the early LYVE-1 expression, it might be caused by the LYVE-1 + bone marrow derived cells rather than by the lymphatic endothelium. Previous studies showed increased LYVE-1 expression by activated macrophages of various tissues in response to inflammation. [20,24] Our data also shows early CD45 + cell infiltration in DE induced LGs (Fig 5A & 5B), which may demonstrate potentials for LYVE-1 + cell infiltration at Day 4. Additionally, despite the fact that mRNA level exhibited a high peak at Day 4, the actual LVs were found at Day 10 with immunostaining, but not at Day 4 ( Fig 2C). This suggests that four days induction may be insufficient for forming mature LVs, implying that the mRNA Dll4/Notch and HIF-1α in LG Lymphangiogenesis peak at Day 4 does not represent LV structure. Future studies for identifying the exact identity of the early mRNA expression of LYVE-1 will be performed for clarification of this finding.

HIF-1α regulates Dll4/Notch signaling and lymphangiogenesis in DEinduced LGs
In DE induced HIF-1α CKO mice, the expressions of DLL4, NOTCH1, LYVE-1, and Podoplanin were inhibited (Fig 4). This indicates that Dll4/Notch signaling and LV formation is regulated by HIF-1α. Bridges et al. supported HIF-1α regulated Notch signaling in lung epithelium, where Notch signaling and lymphangiogenesis was increased by up-regulation of HIF-1α. [13]  In addition, VEGFs, transcribed by the activation of HIF-1α, have been shown to interact with Notch activated VEGFR3 to form new LVs. [25,26] Indeed, our data shows that HIF-1α knockout suppresses Notch signaling and lymphangiogenesis. However, since we did not investigate intracellular signaling for Dll4/Notch induction by HIF-1α, the temporal relationship between HIF-1α and Notch signaling in LGs requires further evaluation.

LVs help reduce inflammatory cell infiltration in DE-induced LGs
The results of this study show diminished CD45 + cell infiltration after the completion of LV formation, while blocking lymphangiogenesis enhanced CD45 + cell infiltration. These findings indicate that newly formed LVs help reduce inflammatory cells in DE induced LGs.
In previous studies, lymphangiogenesis has been helpful in resolving inflammation induced tissue damage in inflammatory diseases. [27,28] In skin inflammatory disease, lymphangiogenesis regulated fluid drainage, immune cell migration, and the removal of inflammatory cells, thereby accelerating the resolution of inflammation. [29,30] Additionally, VEGF-C/VEGFR3 induced lymphangiogenesis accelerated clearance of inflammatory cells and bacterial antigens from inflamed colon to draining lymph nodes in inflammatory bowel disease. [31] Likewise, the results of this study show that lymphangiogenesis during DE-induction aid the resolution of DE-induced inflammation by clearing CD45 + cells from LGs.
There are some limitations of this study. The mouse model for investigating DE may be quite different from the human DE. Human pathologic studies will be needed to confirm the role of Notch signaling induced lymphangiogenesis of LGs. Also, scopolamine, which was used as a routine protocol for DE induction, may have altered the expression of Dll4/Notch pathway and may have affected our results. In addition, while we have only focused on HIF-1α and Notch signaling, other known pathways related to lymphangiogenesis, such as nuclear factor-kappaB (NF-κB) and Janus kinase/signal transducers and activators of transcription (JAK/ STAT), may have been activated by DE stress and may have influenced our results. [26,32] Future studies with these major pathways are needed to fully understand LV formation in DE LGs.
In conclusion, this study demonstrates the relationship between Dll4/Notch signaling, HIF-1α activation, and lymphangiogenesis in LGs in DE. By aiding the reduction of inflammatory cells from LGs, LVs are important in lessening inflammatory damage. Further study for investigating the neural network between the cornea and the LG may explain more fully how the Notch system is activated in LGs. Moreover, cross-talk between VEGFs, Notch signaling, and HIF-1α should be investigated in order to clarify lymphangiogenesis in DE-induced LG.