Anatomy and physiology

Unlike blood capillaries, lymphatic capillaries consist of overlapping oak leaf-shaped endothelial cells (ECs), with incomplete or no basement membrane. Lymph capillaries lack smooth muscle cells or pericyte coverage, but are equipped with thin fibrillar structures called anchoring filaments, which permit expansion and prevent collapse. By connecting the abluminal surface of LECs to the surrounding interstitial matrix, anchoring filaments enable these cells to respond to changes in interstitial fluid pressure, as in oedema, when high interstitial pressure tends to collapse LVs (Ref. Reference Wiig and Swartz2). The precollecting segment does exhibit some coverage (Ref. Reference Florey3); however, collecting vessels are fully equipped with perivascular SMCs (Ref. Reference Reddy and Staub4). To ensure their unidirectional fluid movement, LVs possess two kinds of valves to prevent lymph backflow: overlapping endothelial junctions in the initial lymphatics serve as a primary valve (also called a microvalve) which allows fluid to flow from the interstitial space into the lumen of the LVs while considerably reducing the leakage back into the interstitium, and the traditional intraluminal valve serves as a secondary valve (Ref. Reference Trzewik5).

With their special structural and functional properties, LVs are central to a wide variety of actions in the body. They play a central role in the maintenance of tissue fluid homoeostasis, immune surveillance and trafficking, and fatty acid absorption. LVs drain excess fluid entering the interstitial space from the blood capillaries, and take up extravasated cells and macromolecules and transport them unidirectionally to the draining lymph node, thoracic duct and into the inferior vena cava. LVs thus complete one of the most important physiological functions, by returning interstitial tissue fluid back to the blood circulation. LVs are not simple ducts, since LECs actively participate in a variety of vital biological processes such as reorganisation and remodelling of the extracellular matrix (ECM), modulation of capillary morphogenesis, cellular migration and more importantly, immune response and peripheral tolerance (Ref. Reference Wiig and Swartz2).

Visualisation and phenotyping of LECs

Until recently, our understanding of the lymphatic vascular system was much more limited than of its counterpart blood vasculature. The main reason was a lack of reliable tools to differentiate between LECs and blood endothelial cells (BECs). With the discovery of several markers predominantly expressed on LECs, knowledge in this field has substantially improved in recent years. In this respect, although several markers have been identified to date, the extent of their expression and reliability in all pathological conditions has long been a matter of debate. The consensus, however, is that there still is no perfect and ideal marker available that works reliably in all disease conditions in different organs. Several markers have been proposed in the literature (Table 1); however, the most extensively tested and studied markers yielding satisfactory results are LV endothelial hyaluronan receptor 1 (LYVE-1), vascular endothelial growth factor receptor 3 (VEGFR-3), prosperorelated homeobox transcription factor 1 (Prox1) and mucin-type transmembrane glycoprotein podoplanin. Care should be taken when using these markers, as many other cells and structures can show positivity. Therefore, multiple markers need to be used for confirmation.

Table 1

Summary of markers for lymphatic vessels

LECs at developmental stages

The heterogeneity of LECs at different developmental life stages is a complicated matter and the origins of LVs have been controversial. To date, two major hypotheses have been put forward to explain the origins of LVs: one proposes the budding of primary lymph sacs ‘centrifugally’ by endothelial budding and lymphangiogenesis; the other involves a role for mesenchymal cells. The first centrifugal sprouting hypothesis was proposed by Florence Sabin, based on her observation of LVs sprouting from the jugular veins of pig foetus (Ref. Reference Sabin6). This hypothesis is widely supported, principally by mouse molecular genetics (Ref. Reference Srinivasan7). Several years later, a second alternative model was proposed by Huntington and McClure. This hypothesis proposes that the initial lymph sacs develop in the mesenchyme while, in contrast to the previous model, their origins are independent of the veins (Ref. Reference Huntington and McClure8). However, it has been shown that LVs start growing from distinct ECs which express Prox-1 in the anterior cardinal vein (Ref. Reference Wigle and Oliver9), and that the establishment of blood circulation, which happens in humans at around the sixth to seventh week of embryonic development, is a prerequisite for LV development. Later, this Prox-1-expressing EC population migrates and establishes the primary lymph sac and then, by remodelling, forms the LV system (Ref. Reference van der Putte10). The reader is referred to recent excellent reviews (Refs Reference Schulte-Merker, Sabine and Petrova11, Reference Yang and Oliver12, Reference Koltowska13) for further details on lymphatic development.

LVs in kidney

Before the discovery of specific markers, observations by light or electron microscopy and dye injection studies had revealed no major differences in the distribution of the renal lymphatic system in mammals (Refs Reference Albertine and O'Morchoe14, Reference Cuttino15, Reference Peirce16, Reference Eliska17, Reference Kriz18). Later studies taking advantage of more specific markers such as podoplanin, LYVE-1, Prox-1 and VEGFR-3 confirmed the previous findings on the distribution of renal lymphatics (Refs Reference Lee19, Reference Tanabe20, Reference Matsui21, Reference Voss22). D2–40, a monoclonal antibody which recognises human podoplanin, validated the pattern of human renal LVs that had been uncovered by previous investigations (Refs Reference Sakamoto23, Reference Ishikawa24, Reference Bonsib25). This antibody is a promising marker for the study of human LVs and is in use in daily clinical practice.

Using these novel markers, recent studies have provided in-depth information on the development of renal LVs in the embryonic and adult rodent kidney. Although the extra-renal lymphatic network in mice is well developed at embryonic day (E) 12, the first intrarenal LV is observed at E13 (Ref. Reference Lee19). In contrast, no LVs can be found in developing rat kidney until E20, at which point the first LVs are observed in the renal hilus and interstitium around the pelvis (Ref. Reference Tanabe20). There are no LVs at that point around the arcuate arteries and in the cortical area. However, one day after birth, several small LVs can be observed around the branches of arcuate veins and arteries in the corticomedullary region; since no LVs can be found at this time on the cortex, these LVs appear to be unconnected to the cortical area. The first small LVs in the cortex of rat kidneys are observed around the interlobular arteries at days 10 to 20 after birth (Ref. Reference Tanabe20). LVs accompany intrarenal arteries and veins up to the interlobular branch in the cortex in adult rat kidneys, and occasionally some LVs can be observed near glomeruli. Although there can be several LVs in the renal pelvis, virtually no LVs have been reported in the parenchymal tissue in the cortex and medulla (Refs Reference Tanabe20, Reference Matsui21).

Two lymphatic systems have been observed in human kidneys. One is the intrarenal LVs, which show the same pattern of distribution as in rat kidneys. They follow the intrarenal distribution of blood vessels up to the level of interlobular arteries and veins in the cortex. LVs in the cortical area are occasionally observed in the tubulointerstitial space and around glomeruli, whereas LVs are absent or very rarely found in the outer or inner medulla. LVs are also found in the adipose tissue of the hilum, around the blood vessels beneath the interstitium of the pelvic transitional epithelium, and also in the vicinity of the hilar and sinus arteries and veins (Refs Reference Ishikawa24, Reference Bonsib25). The second capsular lymphatic system is found in addition to these blood vasculature-accompanied intrarenal LVs (Refs Reference Holmes, O'Morchoe and O'Morchoe26, Reference Bell27, Reference Clark and Cuttino28). As the number of LVs decrease in the outer cortex, these capsular LVs mostly drain superficial cortical lymph and run parallel to the renal capsule, eventually connecting to LVs in the renal sinus in the hilum. In healthy kidneys, these capsular LVs drain little fluid, but if there is a urinary tract obstruction (UUO), these LVs become the main route of lymph drainage (Ref. Reference Holmes, O'Morchoe and O'Morchoe26). The specific roles of this dual lymphatic network in renal health and disease remain unclear and warrant future investigation. A schematic representation of the anatomy of renal lymphatics is presented in Figure 1.

Figure 1

Anatomy of renal lymphatics. Renal lymph vessel after entering the kidney follows the same hierarchical distribution as of renal veins and arteries, up to the level of interlobular arteries. In normal kidney, there is almost no lymphatic vessel (LV) in cortical interstitium and few may observe in close proximity to glomeruli. In the medullary part, LVs are absent or very rarely observe. Kidney capsule has also some LVs. (Modified with permission from ref. 75)

Inflammation and inflammatory cells

Lymphangiogenesis has been reported to occur at sites of tissue inflammation, following immunisation and bacterial infection. It is also involved in many inflammatory conditions such as psoriasis, chronic airway inflammation, Crohn's disease, inflammatory bowel disease, rheumatoid arthritis, transplant rejection and tumour metastasis (Ref. Reference Alitalo38). An increasing body of data supports those hypotheses proposing the existence of mutual interaction between inflammation and lymphangiogenesis. LVs form the exit route for inflammatory cells, cytokines, chemokines, inflammatory exudate fluid and macromolecules from tissues to lymph nodes, where the immune response arises. LECs efficiently react to microenvironmental alterations. In this sense, the inflammatory process can affect LVs, while LVs also play a crucial role in promoting inflammatory responses. A wealth of evidence from many investigations has shown that LVs are not just passive conduits but active players in many biological responses such as immunoregulation, which is necessary for immune surveillance and inflammation (Ref. Reference Card, Yu and Swartz39).

Fibrosis

Fibrosis has a close connection to inflammatory cells and mediators, as clearly stated in a recent comprehensive review ‘no inflammation, no fibrosis’ (Ref. Reference Wick76). Therefore, it is highly likely that lymphangiogenesis in the kidney or other organs are associated with fibrosis. Several studies of distinct organs and disease conditions, and renal diseases, have reported that new LV formation occurs in diseases accompanied by fibrotic conditions (Refs Reference Matsui21, Reference Sakamoto23, Reference Lee67, Reference Suzuki69, Reference Zampell77). However, the causal relationship between lymphangiogenesis and fibrosis, evidenced by collagen deposition and scar formation, has not been completely elucidated. Lymphangiogenesis is proposed to be involved in the pathogenesis of idiopathic pulmonary fibrosis (Ref. Reference El-Chemaly78). Recently, ectopic secretion of PDGF-B by LECs, which signals through its PDGFR-ß receptor on mural cells, was shown to recruit aberrant mural cells on LVs. This attachment of mural cells to LVs hampers their drainage capacity, and this defect in the lung's LVs has been reported to be a critical step in progressing to pulmonary fibrosis. This process occurs in the early stage of pulmonary fibrosis development, and by blocking the PDGF-B/PDGFR-ß signalling, the authors were able to prevent recruitment of mural cells to LVs, to restore lymphatic drainage and to ameliorate pulmonary fibrosis (Ref. Reference Meinecke79). This elegant study revealed a crucial role for LVs in the pathogenesis of fibrotic diseases, although it is not yet clear whether the same process holds true for other fibrotic conditions, such as renal fibrosis. Sakamoto et al. (Ref. Reference Sakamoto23) evaluated biopsies from several different chronic renal diseases for LVs by immunohistochemistry. They observed significant increases in LV density (LVD) in tubulointerstitial lesions, which correlate with the degree of tissue interstitial fibrosis. CCR7-positive fibrocytes migrate to the kidney in response to CCL21 chemokine and contribute to the pathogenesis of renal fibrosis. Blocking this CCL21/CCR7 pathway has shown to reduce kidney fibrosis (Ref. Reference Sakai80). As LECs have been shown to actively express and secrete CCL21 chemokine in the kidney (Ref. Reference Kerjaschki45), it is likely that blocking LEC expression of CCL21 could prevent renal fibrosis. We found that prominent lymphangiogenesis in proteinuric kidneys occurs before the influx of macrophages, deposition of collagens and interstitial fibrosis (Ref. Reference Huang72). It appears that lymphangiogenesis would occurs before clear signs of fibrosis, at least in our rat proteinuric model, and this is more in line with a profibrotic response.

TGF-β showed both inhibitory and stimulatory effect on lymphangiogenesis (Refs. Reference Oka81, Reference Kinashi82, Reference James, Nalbandian and Mukouyama83), and LECs are also able to secrete TGF-β (Refs. Reference Podgrabinska84, Reference Malhotra85). Whether LECs in renal disease conditions can secrete TGF-β and play a direct role in fibrogenesis needs further investigations. Zhao et al. (Ref. Reference Zhao86) recently proposed that VEGF-C directly induces fibrogenesis. It is therefore very intriguing to examine VEGFR-3 decoy receptor strategies to catch away and block VEGF-C in some renal diseases with, for example, VEGF-C secretion by activated tubular epithelial cells. By this intervention both fibrosis and lymphangiogenesis should be prevented at the same time.

Figure 2 shows how activated renal lymphatic endothelium contributes to inflammation and fibrosis, and therefore could substantially contribute to tubulointerstitial remodelling in kidney diseases.

Figure 2

The connection between renal lymphangiogenesis, inflammation and fibrosis in the progression of tubulointerstitial remodelling in kidney diseases. In many renal diseases, the initial insults via different mechanisms are able to activate tubular epithelial cells. Upon activation, injured tubular cells start making and secreting a variety of chemokines and meditators which have huge impacts on several key processes in kidney diseases, such as renal inflammation and fibrosis. The reciprocal connection between inflammation and fibrosis are very well studied and known. Both infiltrated inflammatory cells, (myo)fibroblasts and their mediators are shown to promote indirectly renal lymphangiogenesis and remodelling. However, these effects are mutual. In addition to indirect effects, activated tubular epithelial cells (mostly of proximal tubules) secrete specific lymphangiogenic growth factors such as VEGF-C and VEGF-D; therefore directly stimulate the formation of new lymphatic vessel in tubulointerstitial area. Lymphatic endothelial cells under this inflammatory microenvironment also secrete several chemokines as among which CCL21, which can attract CCR7-expressing inflammatory cells and fibrocytes into the perilymphovascular space, and in this way exacerbate further inflammatory and fibrotic reactions.

Transplantation

The exact role of LVs and lymphangiogenesis in organ transplantation remains under debate. Lymphangiogenesis is suggested to be beneficial, for example, by draining inflammatory cells and cytokines, but could also be detrimental, by transporting antigen-presenting cells to lymph nodes and initiating immune responses. Inhibiting lymphangiogenesis in rodent corneal transplantation models was reported to be beneficial by promoting allotransplant survival and appears to be a promising therapeutic tool in enhancing corneal transplant outcomes (Refs Reference Ling87, Reference Tang88, Reference Dietrich89, Reference Uehara90, Reference Zheng, Lin and Ling91). Corneal lymphangiogenesis appears to have negative effects on human cornea graft survival, and has been suggested as a sign of poor prognosis in corneal transplant rejection (Ref. Reference Zheng, Lin and Ling91). Lymphangiogenesis has also been shown to be a common phenomenon in solid organ transplantation, at least in the lungs, kidneys and heart (Refs Reference Geissler92, Reference Dashkevich93, Reference Stuht94). Specific blocking of lymphangiogenesis in experimental heart and islet transplantation models enhances graft survival (Refs Reference Yin55, Reference Nykanen95). Nykänen et al. (Ref. Reference Nykanen95) convincingly demonstrated in experimental mouse cardiac allografts that targeting LVs with anti-VEGFR-3 antibody decreases the number of LECs that express CCL21, infiltrating CD4 + T cells and arteriosclerosis, suggesting a novel intervention to regulate alloimmune activation in solid organ transplantation. Adair et al. (Ref. Reference Adair66) showed that almost all rejected kidney transplantation samples had increased LV numbers as a common phenomenon. Biopsies from human renal allografts showed active lymphangiogenesis, with an approximately 50-fold increase in LV numbers being associated with nodular mononuclear infiltrates (Ref. Reference Kerjaschki45). LECs have been reported to express CCL21 chemokine and attract CCR7+ cells, such as lymphocyte and DCs, providing evidence of an active role for these LECs in organising nodular infiltrates around LVs in the kidney (Ref. Reference Kerjaschki45). This pioneer study highlighted the newly formed LVs and perilymphovascular infiltrates as important origins for alloimmune responses, which have a central role in renal transplant rejection. We, among others, also observed renal lymphangiogenesis in experimental renal transplantation rat models (Refs Reference Vass96, Reference Rienstra97, Reference Palin, Savikko and Koskinen98). Inhibition of the mammalian target of rapamycin (mTOR) blocked lymphangiogenesis, and this anti-lymphangiogenic effect can most probably explain prolonged lymphoedaema in renal transplanted patients treated with rapamycin (Ref. Reference Huber99). In line with these findings, the beneficial effect of sirolimus, another mTOR inhibitor, in ameliorating experimental chronic kidney allograft injury has been related largely to its anti-lymphangiogenic properties (Ref. Reference Palin, Savikko and Koskinen98). In addition to sirolimus-related lymphedema, Ersoy. and Koca: (Ref. Reference Ersoy and Koca100) recently reported everolimus-related lymphedaema in renal transplantation, which indicates the anti-lymphangiogenic activity of everolimus. The therapeutic value of these mTOR inhibitors in blocking lymphangiogenesis remains to be elucidated in some conditions where new lymph vessel formation is detrimental. Blocking lymphangiogenesis seems to be a double-edged sword intervention after organ transplantation. It might be beneficial by interrupting the route of antigen-presenting cells (APCs) to reach lymph node and preventing raising of alloimmune response or detrimental by hampering the clearance of inflammatory cells, interstitial fluids and macromolecules. This area needs further experimental and clinical studies in order to prove the pros and cons of therapeutic values of targeting renal lymphatic and lymphangiogenesis in the course of kidney transplantation.

Renal microenviromental homoeostasis and oedema

LVs have a central role in maintaining fluid transport in the body, and disruption of this pivotal function leads to accumulation of tissue fluid and oedema formation. Interstitial fluid pressure increases in conditions where lymphatic drainage is insufficient, such as in lymphoedaema, causing tissue swelling, increased hydraulic conductivity and remodelling of interstitial matrix (Ref. Reference Mendez101). Renal interstitial oedema and swelling are common as a consequence of the tubulointerstitial injuries in chronic kidney diseases associated with proteinuria (Ref. Reference Rodriguez-Iturbe, Herrera-Acosta and Johnson102). The establishment of renal lymphatic circulation appears to be crucial to renal homoeostasis and function.

Experimental disruption of lymphatic circulation in rat kidneys causes protein-rich fluid retention in renal interstitium, fibrosis, renal cell apoptosis and severe proteinuria, and collectively leads to marked chronic kidney failure (Refs Reference Zhang103, Reference Zhang104). On the other hand, lymphangiogenic therapies have shown promising results in restoring lymphatic function and drainage, thereby ameliorating tissue oedema (Ref. Reference Choi105). Provoking lymphangiogenesis by viral vector overexpression of VEGF-C or exogenous administration of recombinant VEGF-C profoundly decreased interstitial fluid accumulation in experimental lymphoedaema models (Refs Reference Szuba106, Reference Cheung107). Of remarkable interest is a recent study by Planas-Paz et al. (Ref. Reference Planas-Paz108) which convincingly showed that increased interstitial fluid pressure can mechanically prompt lymphangiogenesis in embryos. They demonstrated that mechanical stretching of LECs induces their proliferation in embryonic life, leading to increased LV's size to improve their ability to drain fluid. This mechanism could be important for oedematous disorders in adult life.

Fascinating discoveries from Titze's lab (Refs Reference Machnik109, Reference Machnik110, Reference Wiig111) on the subject of nonosmotical storage of sodium in the subcutaneous interstitium involving immune cells and LVs, usher in a new era in the sodium homoeostasis and hypertension field. This unique sodium storage subsequently induces lymphangiogenesis in the skin via the macrophage/VEGF-C signalling pathway, thereby creating buffering phenomena to halt hypertension in the salt-overload state. This lymphangiogenic response to high-salt diet appears to be specific to the skin and is not observed in the kidney. Interestingly, we found that plasma VEGF-C is modulated by salt intake, at least in proteinuric patients with chronic kidney disease, with higher VEGF-C levels being observed during high salt intake. We also reported increased VEGF-C level in proteinuric patients compared with healthy control individuals for both sodium intakes (Ref. Reference Slagman112). Circulating VEGF-C is also higher in preeclamptic women compared with normal pregnant women (Ref. Reference Lely113). Although the response of VEGF-C to altered sodium intake in preeclamptic women remains unclear, this does suggest that increase in the lymphangiogenic state also occurs in preeclampsia, as a parallel to our observation of CKD proteinuric patients (Ref. Reference Toering, Navis and Lely114). These data emphasise that blood pressure regulation is not only a kidney affair and actually is more complex, highlighting the role of LVs and immune cells in hypertensive disorders. How these advances can be transferred to daily clinical practice for preventive or curative therapy is the issue of current research area in this field.

Proteinuria

Many renal diseases are associated with proteinuria. Since proteinuria is independently associated with the decline in renal function, anti-proteinuric treatment comprises a major cornerstone in clinical nephrology. Nevertheless, annihilation of proteinuria is very difficult, and most patients slowly progress towards renal failure. This indicates the need for additional treatment modalities: not only in trying to reduce proteinuria even further, but also to reduce the harmful effects of proteinuria downstream. Overexposure to filtered plasma proteins injures and activates tubular epithelia cells, and upon activation they start secreting several kinds of chemokines and mediators (Ref. Reference Mono115). We recently showed in a time-course study of experimental rat proteinuric nephropathy that proteinuria can trigger renal lymph vessel formation (Ref. Reference Yazdani70). Secretion of VEGF-C by activated proximal tubular epithelial cells secondary to proteinuria was suggested to be the main source for promoting proteinuria-driven renal lymphangiogenesis. We observed in this model that early proteinuria cannot evoke renal lymphangiogenesis up to six weeks. However, prolonged and chronic proteinuria by activation of the tubular epithelial cells can trigger massive renal lymphangiogenesis at week 12. In addition, anti-proteinuric intervention prevented this renal lymphangiogenesis. We found that renal lymphangiogenesis occurs in line with tubular activation (evidenced by osteopontin expression), profibrotic reactions (interstitial α-SMA); however, before the influx of ED1-positive macrophages, and before deposition of collagen I and III. This suggests that renal lymphangiogenesis is an early response, at least in our model, and starts prior to the development of fibrosis and inflammation. As already mentioned, LECs also express CCL21, which can attract CCR7+ fibrocytes and immune cells and are therefore able to promote renal fibrosis. Although we did not investigate whether LECs express this chemokine at the early phase (6 weeks), blocking lymphangiogenesis appears to be a promising intervention to prevent proteinuria-mediated tubulointerstitial inflammation and fibrosis, and is currently under investigation. Another observation was concomitant tubular expression of osteopontin which, after VEGF-C, is a potent lymphangiogenic factor, at least in the melanoma cell line (Ref. Reference Liersch116). To date, the role of osteopontin in renal lymphangiogenesis remains unknown.

Diabetic nephropathy

Hyperglycaemia has been linked to LV remodelling and lymphangiogenesis in various organs (Refs Reference Moriguchi117, Reference Haemmerle118, Reference Kivela119, Reference Yin120). Sakamoto et al. (Ref. Reference Sakamoto23) reported that lymphangiogenesis and VEGF-C expression was significantly higher in human biopsies with type 2 diabetic nephropathy than in nondiabetic renal diseases. Notably, there was no correlation between renal lymphangiogenesis, the extent of fibrosis and inflammation in these biopsies, which could suggest that diabetic conditions have a direct impact on the growth of renal lymphatics. Similarly, oedematous renal tissues with nephropathy and renal lymphangiogenesis have been observed in diabetic mouse kidneys (Ref. Reference Uchiyama121). However, the effect of glucose and advanced glycation end-products on VEGF-C production and other lymphangiogenic mediators which orchestrate renal lymphangiogenesis in diabetic nephropathy and its impact on renal pathophysiology has not yet been fully explored.

Renal cancer

The potent impact of lymphatic invasion and lymphangiogenesis in cancer development and metastasis has been extensively investigated, and has been suggested to be associated with poor prognosis in several human malignant tumours (Ref. Reference Karaman and Detmar122). However, the significance of lymphangiogenesis, compared with angiogenesis, is unclear in renal cancers, even in the most common kidney tumour, renal cell carcinoma. Despite the fact that the VEGF-C/D/VEGFR-3-axis has been reported to be important in lymphatic tumour spread (Ref. Reference Lin123), others (Ref. Reference Bierer124) found no correlation for this axis with clinicopathological parameters. Using the D2–40 antibody and applying an immunohistochemistry approach, one study (Ref. Reference Horiguchi125) reported that intratumoural LVs and lymphatic invasion are associated with renal cell carcinoma tumour aggressiveness, and that patients with intratumoural LVs have poor prognosis. Two other studies reported that intratumoural LVs rather than peritumoural LVs correlate with prognostic values such as distant metastasis, and can be used as indicators of high risk of renal cell carcinoma metastasis (Refs Reference Ozardili126, Reference Debinski127). However, other studies aiming to evaluate lymphangiogenesis in renal cell carcinoma were largely unable to find intratumoural LVs and no correlation between these findings (intratumoural and peritumoural LVs) and prognostic clinical parameters was found (Refs Reference Bonsib25, Reference Baldewijns128, Reference Iwata129). LV involvement within the sinus and renal pelvis have also been proposed to be sources for tumour cell spread via LVs and lymph node metastasis in clear cell renal cell carcinoma (Ref. Reference Ishikawa130). Recently, in a large single institution analysis of renal cell carcinoma patients, capillary-lymphatic invasion was suggested to be independently associated with renal cell carcinoma death and metastasis (Ref. Reference Eisenberg131). Although this needs to be validated by further studies, Kroeger et al. (Ref. Reference Kroeger132) reported for a small cohort that clear cell renal cell carcinoma with lymphatic spread can be characterised by unique molecular, clinicopathological and genetic changes, such as high epithelial VEGFR-2 protein expression and low chromosome 3p loss at the genetic level. Other studies have revealed no induction of VEGF-C in renal cell carcinoma (Ref. Reference Voss22) and a very limited ongoing lymphangiogenesis (Ref. Reference Iwata129). Collectively, the conflicting data on lymphangiogenesis in renal cell carcinoma does not yet allow any clear conclusion on its potential roles in tumour growth and metastasis and warrant future investigations.