Modeling lymphangiogenesis: Pairing in vitro and in vivo metrics

Lymphangiogenesis is the mechanism by which the lymphatic system develops and expands new vessels facilitating fluid drainage and immune cell trafficking. Models to study lymphangiogenesis are necessary for a better understanding of the underlying mechanisms and to identify or test new therapeutic agents that target lymphangiogenesis. Across the lymphatic literature, multiple models have been developed to study lymphangiogenesis in vitro and in vivo. In vitro, lymphangiogenesis can be modeled with varying complexity, from monolayers to hydrogels to explants, with common metrics for characterizing proliferation, migration, and sprouting of lymphatic endothelial cells (LECs) and vessels. In comparison, in vivo models of lymphangiogenesis often use genetically modified zebrafish and mice, with in situ mouse models in the ear, cornea, hind leg, and tail. In vivo metrics, such as activation of LECs, number of new lymphatic vessels, and sprouting, mirror those most used in vitro, with the addition of lymphatic vessel hyperplasia and drainage. The impacts of lymphangiogenesis vary by context of tissue and pathology. Therapeutic targeting of lymphangiogenesis can have paradoxical effects depending on the pathology including lymphedema, cancer, organ transplant, and inflammation. In this review, we describe and compare lymphangiogenic outcomes and metrics between in vitro and in vivo studies, specifically reviewing only those publications in which both testing formats are used. We find that in vitro studies correlate well with in vivo in wound healing and development, but not in the reproductive tract or the complex tumor microenvironment. Considerations for improving in vitro models are to increase complexity with perfusable microfluidic devices, co‐cultures with tissue‐specific support cells, the inclusion of fluid flow, and pairing in vitro models of differing complexities. We believe that these changes would strengthen the correlation between in vitro and in vivo outcomes, giving more insight into lymphangiogenesis in healthy and pathological states.

new lymphatic vessels, and sprouting, mirror those most used in vitro, with the addition of lymphatic vessel hyperplasia and drainage. The impacts of lymphangiogenesis vary by context of tissue and pathology. Therapeutic targeting of lymphangiogenesis can have paradoxical effects depending on the pathology including lymphedema, cancer, organ transplant, and inflammation. In this review, we describe and compare lymphangiogenic outcomes and metrics between in vitro and in vivo studies, specifically reviewing only those publications in which both testing formats are used. We find that in vitro studies correlate well with in vivo in wound healing and development, but not in the reproductive tract or the complex tumor microenvironment. Considerations for improving in vitro models are to increase complexity with perfusable microfluidic devices, co-cultures with tissue-specific support cells, the inclusion of fluid flow, and pairing in vitro models of differing complexities. We believe that these changes would

| INTRODUC TI ON
The lymphatic system is responsible for maintaining fluid homeostasis in the body and aiding immune cell trafficking from lymph nodes to injury sites. A healthy lymphatic system is integral to overall health, and its disruption is apparent in many disorders. The lymphatic system mirrors the blood capillaries, reaching every part of the body except the epithelium, sclera, cartilage, brain parenchyma, and intima of blood vessels. There are specialized lymphatic vessels in the meninges surrounding the central nervous system (meningeal lymphatics), and the small intestines (lacteals). The lymphatics within tissues are comprised of initial and collecting lymphatics. The initial lymphatics are thin-walled cylinders made up of lymphatic endothelial cells (LECs) that lack a basement membrane to allow the filtering of the interstitial fluid from tissues. These lymphatic capillaries empty into larger mature collecting lymphatic vessels surrounded by pericytes and smooth muscle cells that transport lymph (the interstitial fluid that entered the lymphatic system) using one-way valves.
These mature vessels drain and filter lymph through the various lymph nodes before entering the thoracic duct to empty into the venous circulation.
Lymphangiogenesis is the formation of new lymphatic vessels.
The process may start from LEC progenitor cells during embryonic development (considered neo-lymphangiogenesis henceforth) or extend from an existing lymphatic vessel (sprouting lymphangiogenesis). LECs initiating from an existing vessel proliferate and migrate into the extracellular matrix (ECM) after activation. Proteins in the vascular endothelial growth factor (VEGF) family that promote vascular genesis also direct LEC behavior. VEGF-C is the central protein that modulates LEC maintenance, migration, and proliferation with VEGF-D as a close paralog that aids in generating lymphatic vessel growth. 1,2 VEGF-C primarily binds with the tyrosine kinase receptor VEGF receptor-3 (VEGFR-3) that is expressed most abundantly on LECs, and activates downstream molecular pathways that regulate LEC migration and proliferation. 3,4 VEGF-C can also activate VEGFR-2 to form a heterodimer with VEGFR-3 to promote lymphatic vessel development. 5,6 The proliferating cells begin to sprout in the surrounding ECM to form naïve lymphatic capillaries, which undergo remodeling for maturation. 7 Enhanced proliferation of LECs in the presence of excess VEGF-C is suggested to cause lymphatic hyperplasia, which can be part of normal physiological lymphangiogenesis but also contributes to a number of pathological states. 8,9 Though the human lymphatic system was anatomically defined and received its name in the 16th and 17th centuries, 21stcentury research aims to characterize its function and development in healthy and pathological tissues. 10 For example, in a diseased or disordered state, the lymphatic system can be blocked, leading to fluid accumulation and fibrosis seen as secondary lymphedema. A report has suggested that patients with head and neck cancers have a 75% prevalence rate for lymphedema after resection surgeries and radiation therapies. 11 Likewise, the lymphatic system can facilitate tumor metastasis to lymph nodes and other organs. Lymph node metastasis is correlated with poor prognosis in a variety of cancers. [12][13][14] Therefore, modulating lymphangiogenesis as a therapeutic approach is often investigated as enhancing it can relieve fluid buildup in the case of cancer-associated lymphedema, while inhibiting it preclinically can slow the metastasis of malignancies.
In order to study and understand lymphangiogenesis and the lymphatic system, researchers use a variety of in vitro and in vivo models to simultaneously examine multiple elements of lymphatic physiology. Here, we aim to describe how lymphangiogenesis can be studied across different test environments, specifically in vitro and in vivo, and how outcomes in these contexts correlate. In this review, the 'hallmarks' of lymphangiogenesis refer to the phenomena (e.g., sprouting), while 'metrics' are how the hallmarks are quantified (e.g., number of new vessels). Only articles that have both in vitro and in vivo experiments are discussed to better pair results across diseases and tissues to inform how lymphangiogenesis is studied and how to interpret results seen in different contexts by examining a breadth of normal and disease states.

| Cell sourcing
Though lymphatics are found throughout the body, a few key primary sources are typically utilized for cell culture. Dermal LECs, isolated from human neonatal foreskin or adult skin, and lymph node LECs are commercially available. LECs can also be isolated from primary tissues using specific cell surface markers such as lymphatic endothelial hyaluronan receptor-1 (LYVE-1), a transmembrane receptor for hyaluronan and a CD44 homolog, 15 podoplanin/D2-40, a transmembrane glycoprotein that can also appear on thymic epithelium and peripheral lymphoid stromal cells, 16 and the aforementioned VEGFR-3/Flt-4. 17,18 Nuclear prospero-related homeobox 1 (PROX-1) protein can also identify LECs. 19,20 Other tissue sources for commercial LECs include human lung, 21,22 rat mesentery, 23 and mouse mesentery. 24 Culturing primary LECs in vitro comes with a set strengthen the correlation between in vitro and in vivo outcomes, giving more insight into lymphangiogenesis in healthy and pathological states.

K E Y W O R D S
lymphatics, in vitro models, regenerative medicine of limitations. Like many primary cell types, primary LECs are limited in passage number, require specific cell seeding densities, may grow slowly, and require plate coatings of fibronectin or collagen which could alter cellular behaviors. 25 Additionally, the shear stresses that may occur during standard passaging procedures can activate LECs. 26,27 There is further evidence that LECs may dedifferentiate or have changes in transcriptional profiles in vitro due to a lack of important microenvironmental cues. 28 LECs can be immortalized for more stable and long-lasting cultures. 29,30 Lymphatic endothelial cells can also be differentiated from a variety of cell sources including adipose-derived stem cells (ADSCs), 31 human pluripotent stem cells, 32 embryonic stem cells, 33,34 and blood endothelial cells (BECs). 35 Differentiation is typically confirmed with the markers listed above for primary LECs. In the endothelial cell lineage, PROX-1 is key to differentiating cells toward a lymphatic phenotype. 35

| In vitro models and metrics of lymphangiogenesis
There are many important elements to creating an in vitro model that adequately recapitulates a tissue that include physical, chemical, and biological cues. 40 Though in vitro models may lack all of the elements of the in vivo environment, they allow for more precise control over the microenvironment and can reduce potential confounding factors. In vitro models of lymphangiogenesis can vary greatly in complexity from monolayers to explants. Each model has an associated set of outcomes that researchers have identified as pro-lymphangiogenic behaviors ( Figure 1).
In 2D culture, simple metrics such as migration can easily be quantified using scratch assays and tissue culture insert migration assays. A scratch assay is performed by growing a confluent monolayer of cells and then physically creating a "scratch" with a pipet tip and tracking cell migration into the empty space over time, as described by Liang et al. 41 This technique is simple, inexpensive, and can be performed with any adherent cell type. Simple 2D cultures can also be used to quantify proliferation via bromodeoxyuridine (BrdU) and MTT (chemical name 3-(4,5-dimethylthiazol-2-yl)-2,5-d iphenyltetrazolium bromide) assays. 42 A tissue culture insert assay allows the quantification of cellular transmigration. Originally called the Boyden chamber assay, tissue culture insert assays are now a common methodology whereby cells are seeded as a monolayer or atop a layer of basement membrane on a porous membrane, with a chemoattractant placed in the media beneath. 43 After a set amount of time, the number of cells that have migrated to the underside of the tissue culture insert can be quantified with a simple nuclear stain. 44 The transmigration assay can also be adapted to examine 3D behaviors with hydrogel systems, interstitial fluid flow, and tumor microenvironment interactions. [45][46][47][48] These models provide straightforward methodology for examining migration and proliferation, two outcomes involved in lymphangiogenesis.
A step up in complexity for in vitro models incorporates hydrogels or spheroids to observe lymphangiogenesis as the formation of vessel-like structures. First performed in 1998 by Kubota et al.,49 F I G U R E 1 An overview of components of lymphangiogenesis and relevant in vitro models. Lymphangiogenesis is characterized by proliferation, migration, and sprouting. In vitro models of varying complexity can be employed to investigate these processes. Quantitative metrics often extracted from such assays include percent proliferation, migration and invasion, the distance of migration, and length and number of new vessels or sprouts. the tube formation assay involves plating endothelial cells on top of basement membrane extract (i.e., Matrigel) and has become ubiquitous for observing angiogenesis in vitro. 50 The cells can readily adhere to the matrix and form tubular structures that resemble networks within 24 h. Additionally, LEC spheroids can be embedded within a hydrogel to observe how the cells invade the surrounding microenvironment, called the spheroid sprouting assay. This assay recapitulates lymphangiogenesis by allowing new sprouts to form from pre-existing structures. 51 Nascent work using self-assembled lymphatic networks in hydrogels is also underway, 52,53 with a focus on microfluidic devices. 30,[54][55][56][57][58][59][60][61] Recent microfluidic models use hydrogel (often collagen)-laden polydimethylsiloxane (PDMS) with hollow channels for seeding lymphatic endothelial cells, called gravitational lumen patterning, and allow for flow through the lumen, which can be pressure-head induced or provided by pumps. 30,56,59,62 In these devices, there is precise control of the microenvironment An alternative ex vivo assay uses rat mesentery tissue with live imaging of lymphatic vascular development for 5 days. 66 However, the inability to perform these assays with human lymphatics is a limitation of the current ex vivo systems. Mouse and human endothelial cells respond differently to inflammation 67 and anti-angiogenic drugs are more effective in mouse models than in human patients. 68 Additionally, though many subsets of LECs have similar gene expression between mice and humans, there are some differences, such as the high expression of IL-6, MMP-2, and lysyl oxidase in humans but not mice, pointing to differences in basal inflammatory states and matrix remodeling mechanisms. 69 Beyond migration and proliferation, the crux of many lymphangiogenesis-specific studies revolves around VEGF-C and -D and their receptors, VEGFR-2 and VEGFR-3.
VEGF-C and -D can be quantified at the protein level with ELISA or Western blot. The receptors can be visualized via immunocytochemistry or flow cytometry. These techniques are compatible with most in vitro assays. Further, LECs can be visualized using antibodies for LYVE-1, PROX-1, podoplanin, and CD31/PECAM-1, allowing the quantification of LEC morphology and junctions. It is important to note that CD31 is a better marker for LECs in vitro, as in vivo, CD31 expression is higher in blood vasculature than in lymphatic vasculature. 70,71 Outside of lymphangiogenesis, functional LEC assays can include characterization of permeability, immune cell trafficking, and fluid transport, as discussed in recent reviews. [72][73][74] These metrics can be combined with lymphangiogenesis but often focus on the function of pre-existing lymphatic networks, such as in perfusable microfluidic devices.

| IN VIVO MODEL S AND ME TRI C S OF LYMPH ANG IOG ENE S IS
Measuring lymphangiogenesis in vivo is an essential step in research because the native microenvironment with its 3D structure, support cells, fluid flow, and circulating proteins are challenging to recapitulate in vitro as described above. Lymphatic system research commonly uses lower-order vertebrate animal models (i.e., zebrafish and mice), though higher-order vertebrate animal models, such as sheep, are useful in better mimicking human lymphatic function due to the similarity in size. Though zebrafish do not recapitulate all elements of the human lymphatic system (i.e. lymph nodes) they share cellular and molecular characteristics of the early development of lymphatic vessels and the thoracic duct in vertebrates that closely parallels lymphangiogenesis in mammals. 75,76 The ability to externally fertilize zebrafish embryos and ease of genetic modification makes them a higher throughput model organism. 76 Their transparency, aided by expression of the green fluorescent protein (GFP) transgene that produces GFP in endothelial cells, allows for in vivo optical imaging of the vascular systems. Studies focus on the development and measurement of the thoracic duct as the duct length can translate to broader lymphatic system defects. 76,77 Likewise, mouse models are used for their small size, cost efficiency, and similarities to the human genome. 76 They also can be genetically modified to visualize the lymphatic system. Researchers have added fluorescent tags to lymphatic vessels via the Prox1 gene in Prox1-EGFP or Prox1-tdTomato mice. [78][79][80] Another tactic has been to modify the mouse's genetic code so that EGFP-luciferase is co-expressed with the Vegfr3 gene to monitor lymphangiogenesis in vivo. 81 Within mice, scientists commonly utilize the ear, cornea, hind leg, and tail as locations to visualize the lymphatic system and its development in situ ( Figure 2). Mouse ears have many lymphatic vessels close to the surface and can be used as dermal sheets or in sponge assays. 76,77 The mouse ear sponge assay simultaneously examines the proliferation, migration, and tubulogenesis of LECs in a gelatin sponge implanted into the upper mouse ear. 82 The mouse cornea is a unique location to observe lymphangiogenesis since it is avascular, providing an environment to observe lymphangiogenesis without interference from the native network. 76,77 This model was first described by Cao et al. 83 but has been modified to look at inflammation-induced lymphangiogenesis by placing a suture in the cornea. 84 Another form of visualizing lymphangiogenesis, specifically invasion, utilizes an implanted collagen window in a mouse tail, as created by Boardman and Swartz in 2003. 18 A collagen dermal equivalent is used as a scaffold in an excised section midway up the tail to visualize cell migration and proliferation in an acellular and regenerative format. 85 A similar technique is performed outside of the tail, as seen with the Matrigel plug assay, where Matrigel is subcutaneously injected and then excised later to visualize the vascularization of the overlaying skin. 86 Researchers have also implanted chronic windows to use intravital microscopy to visualize native lymphangiogenesis in a number of tissues and contexts. 87,88 In addition to the models and techniques used to visualize the growth of new lymphatic vessels, in vivo studies also examine lymphatic drainage to probe the functionality of new and existing lymphatic vessels. This can be monitored by injecting Evans blue into the mouse's tail or rear footpad to observe the drainage of the hind leg. 89,90 Micro-lymphangiography is another in situ technique that uses fluorescent isothiocyanate (FITC) or a macromolecule homolog injected intradermally in the tail to reveal the hexagonal fluid channels and network of functional lymphatic capillaries. 18,85,90 Alternative methods to image the lymphatic system and its dynamics include microbubbles with ultrasound, 89 microbead injection and tracking, 91 lymphoscintigraphy, 92,93 and other imaging technologies recently reviewed. 94 With any in vivo model, histology and immunofluorescence can be used to demonstrate how LEC-specific positive staining denotes the luminal structures of vessels while maintaining the native tissue organization. Due to the multiple cell types in tissues, it is crucial to make a distinction between lymphatic and non-lymphatic endothelial structures using colocalization of the standard LEC markers (LYVE-1, PROX-1, and podoplanin) with CD31/PECAM-1 to differentiate BECs, 85 and F4/80 to mark macrophages. 85,95 The colocalization of LEC markers can also distinguish the maturity of vessels with CC chemokine ligand 21 (CCL21) for initial lymphatic capillaries 96 and alpha-smooth muscle actin to denote collecting vessels. 91,97 Researchers have also used transmission electron microscopy (TEM) to visualize matrix remodeling and LEC migration using ultrathin sections stained with uranyl acetate and lead citrate. 77 Multiple metrics have been used to determine if lymphangiogenesis is indeed occurring as described in Figure 2. Researchers can monitor the expression of VEGF-C and VEGF-D in vivo using the same protein analysis methods mentioned earlier. 80,95 For instance, one can measure the activation of VEGFR-3 to denote the initiation of lymphangiogenesis. 18,95 Quantifying the cell density of positively stained LECs (i.e. proliferating LECs using co-stains with BrdU) or the percentage of F I G U R E 2 Hallmarks of lymphangiogenesis in vivo include new vessels, hyperplasia, sprouting, or activation. In vivo models of lymphangiogenesis typically use genetically modified zebrafish and mice. Some examples of in situ mouse models are the cornea, ear, hind leg, and tail. Histology and lymphatic drainage are used to quantify lymphangiogenesis. the fluorescent area using the LEC-specific markers can display LEC activation and number of lymphatic vessels. 85 Counting the number of LEC+ lumens, the lymphatic vessel density (total count over a specific area), and the area of lymphatic vessel coverage identifies the creation of new lymphatic vessels in tissue cross-sections. 77,95,98 In models where the 3D lymphatic network can be visualized, studies denote the increase in lymphatic vessels or tubulogenesis by quantifying the nodes (points where multiple vessels meet) and bifurcations/branches (number of additional vessels connected to a selected vessel). 76,77,84 Lymphatic hyperplasia can be quantified by measuring the diameters of lymphatic vessels as well as the cross-sectional area of each vessel to identify the start of lymphangiogenesis, initiated by increased levels of VEGF-C from proliferating LECs. 8,9,77,84,91 Any model used to capture human physiology and disease has drawbacks and weaknesses. While in vivo animal models allow for experiments in native tissue environments, 99

| Lymphatic development and dysregulation
Focusing on lymphatic development is important to understanding how lymphangiogenesis occurs in the embryo and how lymphatics could regenerate to remedy pathologies like lymphedema.
Detry et al. 103 explored the step-by-step mechanism of lymphangiogenesis by using injury models in vivo and bridged the gap across experimental platforms using an ex vivo LRA. In vitro and in vivo, LEC sprouting can be characterized by the elongation of cells into thin structures that probe the tissue microenvironment, the formation of vacuoles, and the remodeling of a chemotactic matrix that promotes tubulogenesis and migration. In this case, the selection of the LRA allows for sprouting lymphangiogenesis in vitro, showing vessel formation from an existing lymphatic vessel. Therefore, the in vitro and in vivo models demonstrate comparable results. 103 In more basic models, the lack of 3D structures or pre-existing vessels could limit whether the results correlate to in vivo work.
Lymphedema is a disorder where fluid accumulates in a tissue due to a damaged or blocked lymphatic system that can be congenital or secondary to injury. Studies aim to stimulate lymph- Lee et al. 105 found that hypoxic media from MSCs aided lymphangiogenesis, shown in vitro by proliferation, migration, and tube formation, and in vivo by LYVE-1 staining and lymphangiography.
In these cases, positive impacts in vitro correlated with functional outcomes and edema relief in vivo; these studies focus on LECs in a regenerative capacity and not on recapitulating the disordered state.
In contrast to lymphedema treatments, where lymphangiogenesis is stimulated, treatments for cornea-related inflammation aim to inhibit lymphangiogenesis as it creates a hostile environment for transplants. Here, studies often focus on pharmacological inhibitors, such as Nintedanib 106 and Toluquinol. 76 Yuen et al. 84 used 3D cultures for tube formation assays alongside corneal neovascularization to explore Angiopoietin-2's role in inflammatory lymphangiogenesis and how its suppression could inhibit lymphangiogenesis.
The use of the cornea model makes these findings applicable to the broader field of lymphatics research as it is a good model of neolymphangiogenesis while matching the simplicity of in vitro models due to limited crosstalk with the tissue microenvironment. However, overall, research on lymphatics is beginning to move toward more complex in vitro modeling via microfluidic devices. 56 These models could be leveraged to develop in vitro studies that more accurately replicate the characteristics of lymphedema and other lymphaticrelated disorders.

| Lymphatics in the skin
Many models of lymphangiogenesis focus on dermal lymphatics, as they are well-studied, easy to access, and derive from human tissues. Thus, dermal LECs have been the standard for in vitro cell culture. Gordon et al. 107 investigated the impact of macrophages on dermal lymphatic vessel development. Macrophages, an immune cell responsible for phagocytosis of pathogens, can cause and alleviate inflammation at a site of injury and regulate angiogenesis and lymphangiogenesis. 108 During development, a subset of LYVE-1+ macrophages colocalizes to embryonic lymphatic vasculature, sometimes even incorporating into vessels. However, transdifferentiation of these macrophages was disproven as they do not express Prox-1.
In PU.1 knockout mice that lack macrophages, lymphatic vessels were hyperplastic, with large vessel diameters and enhanced proliferation, but no change in branching ( Figure 3B,C). These results suggest that macrophages can negatively regulate LEC proliferation. However, in vitro, macrophages directly co-cultured with LECs enhance proliferation ( Figure 3D), while macrophage conditioned media did not, suggesting that direct crosstalk between macrophages and LECs is necessary for positive impacts on proliferation in vitro. 107 Macrophages require a broad spectrum of cues to regulate their pro-or anti-inflammatory behavior, and in vitro, these cues are not easily recapitulated. Gordon et al. also noted that macrophage transdifferentiation is potentially easier to observe in vitro.
Here, it is necessary to comment that simplified in vitro systems may not always recapitulate findings seen in vivo, especially when complex multicellular interactions are involved.
Lymphangiogenesis during wound healing is a longtime focus of study in the skin. For direct comparison of wound healing assays in vivo and in vitro, studies have paired the in vivo ear biopsy wound margin closing assay with in vitro scratch and tube (or chord) formation assays. 109 In patients with diabetes, therapeutic strategies for improving wound healing are particularly important. Wu et al. 110 examined lymphangiogenesis in type 2 diabetes in vitro and in vivo.
Epsins 1 and 2 were identified as potential therapeutic targets due to their roles in regulating the spatial availability of VEGFR-2 and VEGFR-3. In diabetic mice, epsin deficiency with VEGF-C supplementation led to enhanced corneal lymphangiogenesis in vivo, shown by increased vessel density and branching ( Figure 3E,F), and in vitro, shown by increased proliferation, migration, and tube formation ( Figure 3G). Further, epsin deficiency relieved tail edema in diabetic mice. 110

| Lymphatics in adipose tissue
Adipose tissue is the master regulator of metabolism in the body.
Adipocytes intrinsically support vascularization in vitro 116,117 and thus could be harnessed for lymphatic regeneration. Many lymphatic-related diseases are interconnected with adipose tissue dysfunction. Lipedema, for example, is a disorder characterized by dysfunctional drainage of adipose tissue lymphatics; this is an understudied and underdiagnosed disease, but work is underway to address it. In obesity, there is conflicting evidence on how obesity and high-fat diets may interact to impair lymphatics. 93 121 However, it's important to note that xenograft models do not have a fully functioning immune system, and thus differ from both syngeneic models and humans in ways that likely affect and are affected by lymphatics. 123 Tumors with more fluid drainage (seen by contrast-enhanced MRI) would also have more and larger lymphatic vessels, which correlated with more metastasis to lymph nodes. 121 In vivo tumor studies estab-  Figure 4C). Across these studies, it seems that modeling the tumor microenvironment with 3D co-cultures is important to translation in vivo and should become standard practice to address the complexities of breast cancer pathophysiology. 47,132

| Lymphatics in the cardiothoracic system
In the cardiovascular system, lymphangiogenesis for regeneration post-myocardial infarction (MI) is the primary pathological focus.
After MI, impaired lymphatics lead to edema, inflammation, and fibrosis. 133 Interestingly, an increased number of lymphatic vessels is observed post-MI, but the remodeling of the collecting vessels causes impaired fluid transport. In short, even though there are more vessels after MI, they are dysfunctional. 134

| Lymphatics in the abdomen
Though the abdomen is home to the specialized gut lymphatic vessels (lacteals), studies here primarily focus on lymphatic function and drug transport rather than lymphangiogenesis. Thus, we refer the interested reader to a recent review. 144 As often noted in this review, lymphangiogenesis can be beneficial or detrimental to organ function, and thus there is a careful balance needed. In the kidney, lymphangiogenesis can be associated with fibrosis and renal disease. 145 Kinashi et al. 146 131 In vivo, TNFα related lymphangiogenesis is abolished by VEGFR3 blocking and Tnfr1 knockout. Mouse corneas are stained for LYVE-1 (lymphatics) and CD31 (blood vessels) after treatment with VEGF-C, TNFα, TNFα and VEGFR3 blocking antibody, or TNFα and Tnfr-1 knockout (D). In vitro, TNFα causes alignment of LECs. This is prevented by blocking TNFR1, but not VEGFR-3 (E). Scale bars are 200 μm (D) and 100 μm (E). Reproduced with publisher permission from Ji et al. 143 effective in modeling the lymphangiogenic migratory activity seen in mice with unilateral ureteral obstructions. 95 In contrast, immunosuppressive drugs are used during kidney transplants to prevent lymphangiogenesis, negatively impacting organ function as perivascular lymphatic density is correlated with glomerular filtration rate. 147,148 Huber et al. 148 showed that immunosuppressive drugs, even at low concentrations, negatively impact lymphangiogenesis in vitro by decreasing LEC proliferation and mi-

| Lymphatics in the reproductive tract
The lymphatic system can be found throughout the reproductive tract, and lymphangiogenesis is an integral part of normal and pathologic physiology. Svingen et al. 79  Lymphangiogenesis associated with pregnancy can be characterized as lymphatic vessel hyperplasia. 96,150 Placental cytotrophoblasts stimulate LEC migration in vitro as seen by chemotaxis toward cytotrophoblast-conditioned media in a tissue culture insert assay. 96 Thus, there are unique intercellular interactions in the reproductive tract, though more complex models would offer better mechanistic insight.
Lymphangiogenesis is also associated with the gynecological malignancy endometriosis, an estrogen-driven disease where endometrial tissue is found outside the uterus. Endometriosis affects 10%-15% of females of reproductive age, and there is no cure. 98 The pathophysiology is still undetermined, though some hypothesize that it originates and disseminates through the lymphatic system.
Researchers have investigated how VEGF-C regulation can be leveraged as a potential therapeutic. 98,151 Inflammation caused by neighboring immune and support cells is a significant focus of the study of this disease. Li et al. 151 showed lymphatic invasion into endometriotic lesions by the amount of LYVE-1+ cell expression in vivo compared to using tube formation and migration assays to denote lymphatic vessel development in vitro. They found that VEGF-C is a potential diagnostic biomarker, and treatment of Lenvatinib can possibly alleviate endometriosis by reducing the size of endometriotic lesions and decreasing lymphangiogenesis. 151 One drawback to some of the current studies is the reliance on immunohistochemistry, since many lymphatic vessels require further distinguishment from immune cells with additional LEC markers. Regardless, the structure and function of the lymphatic system in the reproductive tract and the development of in vitro models of this tissue will greatly advance this field of study and our understanding of gynecological pathologies.

| PER S PEC TIVE ON MODEL S OF LYMPH ANG I OG ENE S IS: IN VITRO AND IN VIVO
Overall, the highlighted articles show that in vitro and in vivo results often correspond to one another. However, the assays and analyses done in either context vary from lab to lab, organ to organ, and pathology to pathology. In studies of lymphatic development, in vitro assays often mimic well with neo-lymphangiogenesis, as they  55,56,63 Another critical gap in the literature is a lack of in vitro models of the lymph node that allow for the study of lymphangiogenesis. The lymph node is responsible for immunosurveillance and adaptive immunity. In cancer research, tumor crosstalk and chemotherapies can cause lymphangiogenesis in the lymph node. 47,152,153 As many tumors metastasize to the lymph node and evade the immune system, it is essential to study. Lymph node lymphangiogenesis is also impacted by immunization, inflammation, and immune cell crosstalk. 154,155 In vitro models for the study of lymph node lymphangiogenesis can provide insight into autoimmunity and cancer metastasis, as well as the potential to screen therapeutics and vaccines. The development of microphysiological lymph node models is underway, 156 and applicable ex vivo models are also available. [157][158][159] Further, a recent in vivo imaging technique using microbubbles and ultrasound can aid in visualizing lymph node vasculature and flow dynamics so that this vital data can be translated to microfluidics. 89 Specifically, within the context of this review, the correspondence of in vitro to in vivo outcomes depends on the metrics evaluated and the tissue or pathology of interest. Some metrics of lymphangiogenesis typically require more physiological context to interpret. One example is proliferation, which can translate to remodeled hyperplastic vessels or new vessel generation. 8,9,107,136,141,148 Therefore, interpreting in vitro LEC proliferation results can be complex -relating to pathological or beneficial outcomes. One recommendation is that proliferation assays be paired with another metric, specifically an assay that allows for lymphangiogenesis from an existing lymphatic vessel as seen in vivo, to provide more informative results.  107 where macrophages promoted LEC proliferation in vitro but suppressed it in vivo. VEGF-C is known to indirectly recruit proteolytically active macrophages in vivo, aiding LEC migration, highlighting the spectrum of roles macrophages play in lymphangiogenesis. 161 Interestingly, this phenomenon of macrophage or monocyte transdifferentiation into LECs could offer an interesting in vitro system for both cell sourcing and lymphatic study.
For example, in the melanoma tumor microenvironment, LYVE-1+ macrophages mimic sprouting lymphatic vessels. 162 and in renal fibrosis, M1 macrophages can transdifferentiate into LECs. 163 These vascular-associated macrophages are thought to be derived from endothelial progenitor cells. 108 Inflammation is key to causing this F I G U R E 5 The complexity of in vitro models should increase to better compare in vitro and in vivo outcomes. Neighboring cells and interstitial fluid in the microenvironment are critical factors to recapitulate. Some considerations are: (1) using 3-dimensional models that allow for sprouting or using a hydrogel that replicates the ECM and includes support cells (i.e. fibroblasts) and immune cells (i.e. macrophages). (2) including different types of fluid flow to better mimic in vivo conditions since lymphatics play key roles in fluid transport and experience both shear stress and interstitial fluid flow in the body. (3) moving toward microfluidic devices that are perfusable to incorporate the benefits of a controlled 3D microenvironment, fluid flow, and allow for functional assessment (i.e. immune cell trafficking, barrier function) of new lymphatic vessels. transdifferentiation, 164,165 and harnessing this phenomenon in vitro, described by Zhigeng et al., 166 could be a promising source of LECs for studying pathologies with lymphatic involvement.
Moreover, in Ji et al.'s 143 study on TNFα in ovarian cancer, blocking VEGFR-3 did not suppress TNF-alpha's effects on LECs in vitro, but it did suppress TNFα-related lymphangiogenesis in vivo.
In breast cancer, platinum agents induce lymphangiogenic effects in vitro and in vivo alone. However, taxanes, another class of breast cancer chemotherapy, require the presence of tumor cells to induce the same effects. 47,129 These studies directly demonstrate that the complex in vivo environment provides signals that impact lymphangiogenesis and supporting cells, and that these cues may not be recapitulated in vitro without careful model design and cell sourcing.
Lymphatic endothelial cell migration is impacted by ECM composition, alignment, and dimensionality. 77,167 Detry et al. 77 showed that sprouting fails to occur when an LRA is performed in Matrigel instead of collagen and that excessive or insufficient collagen impairs LEC sprouting. Further, a key ECM element that is often missing from in vitro lymphangiogenesis assays is hyaluronan. 168 Cell migration in 2D differs significantly from 3D. In 2D assays, such as scratch assays, there is potential for single or collective cell migration. 44 In 3D, proteolytic degradation of the matrix must occur. As mentioned above, macrophages are recruited in vivo to aid matrix remodeling during lymphangiogenesis. 161 Support cells often aid in proteolytic degradation during lymphangiogenesis. 168 Thus, it is essential to design in vitro models that approach the complexity of the tissue microenvironment so that in vitro results can translate to in vivo outcomes.
In contrast, in regenerative applications, where the goal is induction or regrowth of lymphatics, in vitro results typically translate to functional relief of edema. Without the added complexity of the tumor, LEC behavior seems more predictable. This is also seen in studies that simply aim to examine how anti-cancer drugs can suppress LEC migration and proliferation. When selecting an in vitro assay, it is necessary to consider the underlying hypotheses, desired therapeutic outcomes, and the complexity of the native environment that needs to be replicated when interpreting results between in vitro and in vivo.
Furthermore, there is a distinct gap between studies that neglect fluid flow and those that account for it. Even low interstitial fluid flow velocity still affect lymphangiogenic characteristics including, but not limited to, junction integrity, protein expression, cell-cell interactions, and sprouting. 169 We recommend that experiments measuring sprouting and maturation use fluid flow to mimic the forces in vivo. Previous research has shown that interstitial fluid flow can guide lymphangiogenesis by enhancing morphogenic effects of ECM-bound growth factors (like VEGF-C) and inducing the expression of essential genes required for lymphatic vessel development. 18,56,85,91 The direction of lymph flow has been shown to guide fluid channeling, LEC migration, and functional vessel formation. 18 Sweet et al. 91 reveal that though fluid flow is not required for the primary formation of lymphatic vessels, it is necessary for its remodeling into mature collecting vessels by using knockdown experiments in vitro and in vivo. Fluid accumulation and stretch on the cells cause shear stress, making LECs proliferate, elongate, and sprout. 170,171 Adding interstitial flow can be as easy as calculating the necessary pressure head for a tissue culture insert [45][46][47][48] or using a radial flow chamber, 172 and shear stress experiments can be performed on a parallel plate flow chamber. 91 These modifications can enhance the physiological relevance of lymphatic study in vitro.
As lymphangiogenesis has been overlooked in the past compared to other physiological players, many current studies rely on the fact that drugs already approved for their anti-angiogenic effects also tend to hinder lymphangiogenesis. 173 Preclinical models have played a vital role in uncovering anti-lymphangiogenic therapies, such as antibodies targeting the VEGF-C/VEGFR-2/3 axes, leading to recent clinical trials. 174 As preclinical models of lymphangiogenesis carefully increase in complexity and relevance, their ability to uncover and explore potential therapeutics will lead to immense progress toward treating lymphatic-involved pathologies. Further, the development of models and knowledge surrounding lymphatics in other tissues (such as gynecological tissues) and diseases, will offer even more opportunity for novel models and discoveries surrounding these important vessels.

| CON CLUS ION
Within this review, we have discussed models and metrics of lymphangiogenesis both in vitro and in vivo. In vitro models of lymphangiogenesis correlate best with in vivo data in applications such as wound healing, where lymphatics play a beneficial role. LEC behavior is altered by an extensive range of supporting cells in the tumor microenvironment and other pathological states. Thus, these cues can be more challenging to recapitulate in vitro. We recommend developing in vitro models with physiologically relevant microenvironments and fluid flow, while carefully considering underlying hypotheses or outcomes to be tested, to facilitate the correlation of lymphangiogenesis in vitro and in vivo.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare no conflict of interests.

C I TAT I O N D I V E R S I T Y S TAT E M E N T 175
Recent work in several fields of science has identified a bias in citation practices such that papers from women and other minority scholars are under-cited relative to the number of such papers in the field. [176][177][178][179] Here we sought to proactively consider choosing references that reflect the diversity of the field in thought, form of contribution, gender, race, ethnicity, and other factors. First, we obtained the predicted gender of the first and last author of each reference by using databases that store the probability of a first name being carried by a woman. 177 By this measure (and excluding self-citations to the first and last authors of our current paper), our references contain 17.97% woman(first)/woman(last), 17.4% man/woman, 25.9% woman/man, and 38.72% man/man. This method is limited in that (a) names, pronouns, and social media profiles used to construct the databases may not, in every case, be indicative of gender identity and (b) it cannot account for intersex, non-binary, or transgender people.
Second, we obtained predicted racial/ethnic category of the first and last author of each reference by databases that store the probability of a first and last name being carried by an author of color. 180 By this measure (and excluding self-citations), our references contain 35.26% author of color (first)/author of color(last), 14.07% white author/author of color, 20.0% author of color/white author, and 30.67% white author/white author. This method is limited in that (a) names and Florida Voter Data to make the predictions may not be indicative of racial/ethnic identity, and (b) it cannot account for Indigenous and mixed-race authors, or those who may face differential biases due to the ambiguous racialization or ethnicization of their names. We look forward to future work that could help us to better understand how to support equitable practices in science.