© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
LETTERS
Lymphatic endothelial progenitor cells contribute to de novo lymphangiogenesis in human renal transplants Dontscho Kerjaschki1, Nicole Huttary1, Ingrid Raab1, Heinz Regele1, Katalin Bojarski-Nagy1, Gregor Bartel1, Stefan M Kro¨ber2, Hildegard Greinix3, Agathe Rosenmaier3, Franz Karlhofer4, Nikolaus Wick1 & Peter R Mazal1 De novo lymphangiogenesis influences the course of different human diseases as diverse as chronic renal transplant rejection1 and tumor metastasis2,3. The cellular mechanisms of lymphangiogenesis in human diseases are currently unknown, and could involve division of local preexisting endothelial cells or incorporation of circulating progenitors. We analyzed renal tissues of individuals with gender-mismatched transplants who had transplant rejection and high rates of overall lymphatic endothelial proliferation as well as massive chronic inflammation. Donor-derived cells were detected by in situ hybridization of the Y chromosome. We compared these tissues with biopsies of essentially normal skin and intestine, and two rare carcinomas with low rates of lymphatic endothelial proliferation that were derived from individuals with gendermismatched bone marrow transplants. Here, we provide evidence for the participation of recipient-derived lymphatic progenitor cells in renal transplants. In contrast, lymphatic vessels of normal tissues and those around post-transplant carcinomas did not incorporate donor-derived progenitors. This indicates a stepwise mechanism of inflammation-associated de novo lymphangiogenesis, implying that potential lymphatic progenitor cells derive from the circulation, transmigrate through the connective tissue stroma, presumably in the form of macrophages, and finally incorporate into the growing lymphatic vessel. De novo development of blood vessels (‘adult vasculogenesis’)4 is important for cancer growth5 and transplant survival1. For example, circulating endothelial progenitor cells are integrated into tumor blood vessels both in humans and in animal models, and thus provide a potential therapeutic target, as well as a surrogate marker for antiangiogenic tumor therapy6. In contrast, nothing is known about the mechanisms of de novo lymphangiogenesis in human diseases. Therefore, we examined whether circulating lymphatic endothelial progenitors have any role in this process, similar to the case with blood vessels, or alternatively, lymphatic networks grow by division of
local endothelial cells. For this purpose, we used human tissues from male renal transplant recipients with a female donor kidney, or tissue from female bone marrow recipients who received a graft from a male donor. We detected the nuclei of progenitor-derived lymphatic endothelial cells through colocalization of the transcription factor Prox-1 (ref. 7) by immunohistochemistry, and the Y chromosome by in situ hybridization8. We chose nephrectomy specimens of rejected kidney grafts (n ¼ 6) that showed inflammation-associated, extensive de novo lymphangiogenesis1 (Fig. 1a,b), whereas in normal human kidney only few lymphatic vessels were localized in the adventitia of large to middle-sized arteries. We examined normal or minimally inflamed skin and intestinal tract biopsies from recipients after bone marrow transplantation (n ¼ 32) as controls that presumably reflect lymphatic endothelial cell turnover in normal tissues. We also identified in a worldwide search a post–bone marrow transplant mammary carcinoma and a colorectal carcinoma in appropriately gender-mismatched recipients, each of which showed an elaborate peritumoral lymphatic vasculature and desmoplasia. Collectively, the results indicate that in renal explants, 47 out of 1,005 (4.5%; range, 2.7–7%) Prox-1+ lymphatic endothelial nuclei contained a single Y chromosome, and were therefore derived from circulating progenitors of the host’s genotype (Fig. 1d–h). These Y chromosome+ endothelial cells accounted for 12.9% of the 281 lymphatic vessels encountered, suggesting that these vessels serve as focal sites of de novo angiogenesis (Table 1). We did not observe Y chromosome+ pericytes. Fusion of endothelial progenitors with preexisting endothelial cells was discounted as we were unable to detect more than two sex chromosomes by double localization of X and Y chromosomes as assessed by fluorescent in situ hybridization in a large number (n ¼ 7,135) of nuclei (Supplementary Table 1 online). In contrast, 746 Prox-1+ nuclei of lymphatic endothelial cells in skin and gastrointestinal biopsies of individuals who had undergone bone marrow transplant did not contain a Y chromosome. Also, all 97 peritumoral lymphatic vessels (Fig. 2a–d) and their 338 Prox-1+ nuclei were devoid of Y chromosome+ lymphatic endothelial cells (Table 1). As we used paraffin sections that were 4 mm thick, it is
1Department of Pathology, Medical University of Vienna - Allgemeines Krankenhaus Wa ¨ hringer Gu¨rtel 18 - 20, A 1090 Vienna, Austria. 2Department of Pathology, University of Tu¨bingen, Liebermeisterstrae 8, D 72076 Tu¨bingen Germany. 3Austrian Bone Marrow Donor Registry, Florianigasse 38/12, A 1080 Vienna, Austria. 4Department of Dermatology, Medical University of Vienna - Allgemeines Krankenhaus Wa ¨ hringer Gu¨rtel 18 - 20, A 1090 Vienna, Austria. Correspondence should be addressed to D.K. (
[email protected]).
Received 25 May 2005; accepted 8 November 2005; published online 15 January 2006; doi:10.1038/nm1340
230
VOLUME 12
[
NUMBER 2
[
FEBRUARY 2006
NATURE MEDICINE
LETTERS G
G
Art
Art G
G G
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
a
b
c Ly Ly
Ly
d
G
Ly
f
G
e
Ly
Ly
g
h
probable that all figures obtained are underestimates, because some Y chromosomes could have been located outside the plane of the section. These results depend crucially on the techniques used to precisely identify Y chromosome+ cells as lymphatic endothelial cells, and to avoid confusion with occasional Y chromosome+ inflammatory cells in transit through the vascular wall. In addition to using antibodies to Prox-1 to identify lymphatic endothelial cells, the identity of these cells was confirmed by consecutive immunolabeling of the same sections with antibodies specific for the lymphatic endothelial membrane proteins podoplanin9 and LYVE-1 (ref. 10), followed by antibodies to CD45, and staining with Giemsa dye (Fig. 1c–h). We also noted that expression of LYVE-1 was variable in lymphatic vessels within inflamed areas, whereas expression of podoplanin remained constant (data not shown). Lymphatic endothelial cells were consistently negative for Cd11b (data not shown) and CD68, as found previously1,3. CD45 was expressed exclusively by perivascular mononuclear cells (Fig. 1c), but not by lymphatic endothelial cells. The contribution of cell division to renal lymphangiogenesis was confirmed by the endothelial expression of Ki67 (also known as MIB-1) in 2.3% of all lymphatic vessels; expression was found primarily in large lymphatic vessels (Fig. 2e–g and Supplementary Table 2 online). In contrast, no Ki67+ endothelial cells were encountered in normal tissues of bone marrow transplant recipients, nor in the Figure 2 Distribution of lymphatic vessels in post-transplant carcinomas arising in recipients of gender-mismatched bone marrow transplants. A colorectal carcinoma (a,b) and a mammary carcinoma (c,d,h). The peritumoral stroma adjacent to the carcinomas (CA) is fibrotic and contains moderate mononuclear inflammatory infiltrates in H&E stain (a,c). Lymphatic vessels are immunohistochemically visualized by podoplanin (b,d). (e–h) Double localization of Ki67 (green) and Prox-1 (red) in renal transplant NTX 11316/04 (e–g) and mammary carcinoma (h). Proliferating lymphatic endothelial cells occur relatively frequently in the transplants (white arrowheads), and were not found in the carcinomas. Ki67+ cells are recognized as nonendothelial by lack of reactivity with Prox-1 (open arrowheads in e–g). Ly, lymphatic vessel. Original magnification in a–d, 100; in e–h, 620.
NATURE MEDICINE
VOLUME 12
[
NUMBER 2
[
FEBRUARY 2006
Figure 1 Localization of lymphatic vessels by immunohistochemistry using antibodies to podoplanin in normal human renal cortex from a tumor nephrectomy (a), and in an explanted kidney (subject NTX 11316/04) of a male recipient and a female donor (b). In the normal kidney cortex, lymphatic vessels accompany the interstitial arteries (Art). In contrast, in b there is massive amplification of the lymphatic network around the arteries, and also branching out into the tubulo-interstitial space. Original magnification in a and b, 100. Inset in b shows localization of Id-1 in nuclei (arrowheads) of a small podoplanin+ (red) vessel (original magnification, 400). G, glomerulus. (c) Localization of CD45 (green) and Prox-1 (red) in the renal transplant NTX 11316/04, indicating that CD45 is restricted to the perivascular inflammatory infiltrate, and is not expressed in lymphatic endothelium. (d–h) A single lymphatic vessel in the explanted kidney stained consecutively with Giemsa (d), hybridized with a probe specific for the Y chromosome and localized by immunoperoxidase (reaction product false-colored in blue to enhance visibility, e). (f,g) Immunolabeling with antibodies specific for podoplanin (f) and Prox-1 (g). Original magnification in d–g, 620. (h) Superimposition of the signals for podoplanin (green), Prox-1 (red) and the Y chromosome (blue). Arrowheads indicate three nuclei of the small lymphatic vessel (L) that contain a recipient-specific Y chromosome, and thus originate from recipient-derived progenitors (original magnification, 780). Insert in h shows a small lymphatic vessel in renal explant NTX 8547/98, with green Y chromosome, and Prox-1 indicated by red immunofluorescence (original magnification, 600). Of four lymphatic endothelial nuclei, two contain a Y chromosome (arrowheads).
peritumoral stroma of the colorectal and mammary carcinomas (Fig. 2h and Supplementary Table 2 online). Our results show that recipient-derived cells function as progenitors of lymphatic endothelial cells, and contribute to the overall lymphangiogenesis in renal transplants; however, identification of their origin is complicated by their elaborate itinerary in reaching their final destination. The presumed first step in the life of lymphatic progenitor cells, analogous to that of hemangioblasts11, is their creation and release from the bone marrow into the circulation. Current data indicate that circulating precursors of blood vessel endothelial cells are mobilized from the bone marrow by different
a
b
CA
c
CA
d
CA
CA
e
g
f Ly
Ly
CA
h Ly Ly
231
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
LETTERS factors, and are recognizable in the circulation Table 1 Host-derived lymphatic endothelial progenitor cells in lymphatic vessels as vascular endothelial growth factor receptor Number of Number of (VEGFR)-2+ angiopoietic precursor cells12 Y+ Prox-1+ vessels/ Y+ Prox-1+ nuclei/ Recipient/donor that require the expression of the inhitotal nuclei total vessels gender bitor of differentiation gene (Id-1)13. Here Sample we report that Id-1 is also expressed in the NTX 8,547/98 M/F 10/225 (4.4%) 6/45 (13.3%) nuclei of some podoplanin+ lymphatic NTX 3,631/05 M/F 6/223 (2.7%) 5/28 (17.9%) endothelial cells in renal grafts (Fig. 1b), NTX 11,316/04 M/F 7/101 (7.0%) 5/27 (18.5%) although in low numbers (Supplementary NTX 4,890/04 M/F 8/144 (5.5%) 6/52 (11.5%) Table 3 online), suggesting that lymphatic NTX 9,090/03 M/F 8/214 (3.7%) 6/74 (8.1%) endothelial precursors could also originate NTX 23,265/04 M/F 8/128 (6.3%) 8/54 (14.8%) from the bone marrow. What, then, is the Average 7.83/172.5 (4.5%) 6/46.6 (12.9%) F/M 0/746 0/842 identity of lymphatic progenitors in the cir- BMTX, normal tissue, n ¼ 32 F/M 0/157 0/48 culation? In humans, two populations of BMTX, colorectal cancer 12,498/04 F/M 0/181 0/49 VEGFR-3+ candidate lymphatic progenitor BMTX, mammary cancer 16/96 cells have been identified so far, on the basis BMTX, bone marrow transplant; NTX, renal transplant; M, male; F, female. of their expression of canonical lymphatic endothelial markers in vitro. First, a small subpopulation of CD133+VEGFR-3+CD34+ cells has been identified14. that isolated human naive CD14+ monocytes form aggregates and Second, we have established the occurrence of a major subpopulation extensions that express podoplanin and variably express LYVE-1 after of CD14+VEGFR-3+CD31+VEGFR-2– monocytes1,3 in human blood extended tissue culture (Fig. 3c). Also, the mRNA encoding podothat can be stimulated in vitro to express VEGF-C3 as well as the planin was upregulated under culturing conditions, whereas mRNA lymphatic endothelial marker podoplanin, whereas the expression of encoding LVYE-1 was variably expressed (Fig. 3b). Recently, incorLYVE-1 on these cells remains variable (Fig. 3a,b). It is thus probable poration of transdifferentiated monocytes-macrophages into growing that one or both types of circulating potential progenitors participate lymphatic vessels was directly shown in a mouse corneal transplantation model16. But, once integrated into the lymphatic endothelial in lymphangiogenesis. Before circulating lymphatic progenitors integrate into expanding layer, the donor-derived endothelial cells do not retain any macrolymphatic vessels, they must migrate across the connective tissue phage markers, such as CD45 (Fig. 1c), CD11b (data not shown) or stroma. This obviously distinguishes lymphatic progenitors from CD68 (refs. 1,3). Together, these findings indicate that macrophages circulating hemangioblasts that reach their site of integration directly may serve a dual role in lymphangiogenesis17, either acting directly as from the blood stream in a single step. A major obstacle in following lymphatic endothelial progenitors, or indirectly as a major source of this migratory pathway in tissues is the potential progenitor’s lack of VEGF-C and other lymphangiogenic factors that stimulate division of lineage-specific indicators, and thus expression of VEGFR-3 is cur- local preexistent lymphatic endothelial cells. Appositional growth of rently accepted as an indicative surrogate marker3,14. Potential candi- lymphatic vessels cannot account for the entire de novo expansion of dates for lymphatic progenitors are VEGFR-3+ tissue macrophages of the lymphatic vasculature in transplants: whereas lymphatic vasculahigh developmental plasticity that transdifferentiate into lymphatic ture expands by a factor of o50 times over normal renal cortex1, the endothelial cells. A similar mechanism has also been proposed for fraction of donor-derived lymphatic endothelial cells amounts to only blood vessel endothelia15. In favor of this hypothesis is the finding 4.5%. Thus, it follows that this type of neoangiogenesis must be driven by a combination of division of preexistent endothelial cells and focal incorporation of progenitor cells. Indeed, lymphangiogenesis a c in transplants is associated with a relatively large number of Ki67+ endothelial cells (Supplementary Table 2 online) that occur in large, periarterial lymphatics, whereas appositional growth with donor-derived precursors is restricted to a small number of lymphatic vessels with small diameter. This GAPDH Podoplanin LYVE-1 VEGFR-3 CD31 VEGFR-2 b suggests a specific functional regionalization 1,500 1,000 of this complex process, which is also asso500 ciated with chronic inflammatory infiltrates, suggesting a regulatory role for inflammatory 0 6 36 0 6 36 0 6 36 0 6 36 0 6 36 0 6 36 Hours mediators. The lack of Ki67+ lymphatic + endothelial cells in normal tissues, and also Figure 3 Isolation and in vitro differentiation of VEGFR-3 monocytes. (a) Localization of VEGFR-3 on the surface and the endoplasmic reticulum of FACS-purified CD14+ human monocytes. (b) PCR of in the peritumoral stroma, indicates a slow mRNA encoding GAPDH, podoplanin, LYVE-1, VEGFR-3, CD31 and VEGFR-2 from freshly isolated growth rate that correlates with the absence CD14+VEGFR-3+ monocytes, or after culturing for 6 and 36 h, indicating that expression of mRNA of lymphatic progenitors (Supplementary encoding podoplanin is induced; expression of mRNA encoding VEGFR-3, CD31 and GAPDH remains Table 2 online). This situation is similar to unchanged, expression of mRNA encoding LYVE-1 is variable and mRNA encoding VEGFR-2 is not the case of bone marrow–depleted irradiated + + expressed. (c) Purified CD14 VEGFR-3 monocytes that were cultured for 13 d form aggregates that mice that were repleted with green fluorescontain several LYVE-1+ cells (green) and form podoplanin+ extensions (red). Blue indicates DAPI in nuclei. Original magnification, 1,200. Arrowheads: in a, surface of monocyte; in c, podoplanin+ cells. cent bone marrow18,19. Although the results
232
VOLUME 12
[
NUMBER 2
[
FEBRUARY 2006
NATURE MEDICINE
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
LETTERS from the two rare carcinomas in this study remain to be confirmed, they show that peritumoral lymphangiogenesis in these two specific cases occurs without incorporation of circulating lymphatic progenitors, similar to normal tissues. In conclusion, we propose that in renal transplants de novo lymphangiogenesis involves incorporation of recipient-derived lymphatic progenitors, in addition to division of preexistent endothelial cells. In contrast, we found no evidence for lymphatic precursors in normal tissues, indicating that they grow and are maintained by local endothelial cell division. Our results, together with recently published data on a mouse corneal transplantation model16, highlight a dual role in lymphangiogenesis for macrophages that have the potential to transdifferentiate into lymphatic endothelia or, alternatively, to differentiate into VEGF-C–producing cells that drive division of endothelial cells17,19. METHODS Tissue samples. All information pertaining to subjects and all human samples were used in compliance with Austrian legislation. Archival paraffin blocks were collected by the Departments of Pathology of the Universities of Vienna and Tu¨bingen, and the skin and intestinal biopsies of bone marrow– transplanted individuals were from the Department of Dermatology of the University of Vienna. We used nephrectomy specimens from six male recipients who had received female grafts ranging from 26 to 62 months before nephrectomy, indicated by chronic transplant rejection and renal failure. We selected biopsies with minimal graft-versus-host reaction or normal tissue of skin, stomach and colon of 32 female recipients who were previously (interval between transplantation and biopsy ranging from 2 to 43 months) grafted with male bone marrow as antileukemic therapy. Immunocytochemistry. We dewaxed, hydrated and processed 4 mm thick paraffin sections for immunolabeling, as previously described1,3. We performed antigen retrieval by microwaving the samples in 10 mM citrate buffer, pH 6.0, for 15 min at 620 W. We performed localization of lymphatic vessels in paraffin sections with rabbit podoplanin-specific Ig (5 mg/ml) or with monoclonal human podoplanin–specific mouse IgG (Bender Med Systems BMS 1105; 1 mg/ml), or with monoclonal human MIB-1–specific (Ki67 antigen) mouse IgG (DAKO), human LYVE-1–specific rabbit IgG or human Prox-1–specific rabbit IgG (AngioBio) using a biotin-streptavidin–horseradish peroxidase or alkaline phosphatase method, as previously described2. For immunofluorescence, we used appropriate secondary antibodies labeled with fluorochromes Alexa 488, Alexa 594 or Alexa 633 (Molecular Probes). Double-labeling experiments were controlled by omitting the primary antibodies. In situ hybridization. For fluorescence in situ hybridization (FISH), we dewaxed 4 mm thick paraffin sections with xylene for 10 min (three times), air-dried and rinsed them four times for 2 min in 0.3 M glycine buffer and 2 saline sodium citrate buffer (SSC, Sigma). We then incubated slides in 1% NP40 (Pierce) in PBS for 30 min at 37 1C. We added 10 ml CEP X/Y DNA probes (Vysis-Abbott), and immediately covered the slides with glass cover slips, sealed them with rubber cement, and denaturated the DNA at 81 1C using a metal box for exactly 10 min. For hybridization, we incubated the slides overnight at 37 1C, and removed the unbound probe by three washes with a combination of 50% formamide and 2 SSC for 10 min, followed by 2 SSC and 2 SSC/0.5% NP40 for 5 min at 45 1C. When X/Y FISH was combined with Prox-1 immunofluorescence, we incubated with Prox-1–specific antibody for 12 h at 37 1C after the FISH procedure. We then washed slides three times for 10 min with PBS, and incubated them with biotinylated rabbit-specific antibody for 45 min at 20 1C, followed by streptavidin labeled with Alexa 594 for 30 min. When X/Y FISH was combined with localization of podoplanin, we incubated with monoclonal podoplanin-specific antibody for 45 min at 20 1C, followed by detection with mouse-specific Alexa 488–conjugated goat IgG. We observed the slides in a Zeiss Axiovert fluorescent microscope equipped with a video-capturing system. Photographs were taken in single or combined color channels. As an alternative to FISH, we also detected the Y chromosome by immunoperoxidase, using SpoT-Light Chromosome Y Probe (Zymed,
NATURE MEDICINE
VOLUME 12
[
NUMBER 2
[
FEBRUARY 2006
Invitrogen), and a similar hybridization and washing protocol as that in FISH. We detected Y chromosomes by incubating with horseradish peroxidase– streptavidin conjugate (Sigma), washing in a mixture of PBS and Tween 20 (Sigma), developing with diaminobenzidine (Sigma) and signal enhancement with 2% CuSO4 for 10 min. We viewed Y chromosomes using light microscopy, and in the digital photographs the chromosomes were false-colored with blue using Adobe Photoshop to facilitate visualization. This pattern was subsequently superimposed electronically onto photographs of the same tissue area. We counted Prox-1–reactive nuclei of lymphatic vessels and determined the percentage of nuclei containing a Y chromosome. After these procedures, we removed cover slips from some slides, and then performed indirect immunofluorescence using CD45-specific antibody, and Giemsa or H&E stain. Preparation, culturing and analysis of monocytes. We separated and quantified CD14+VEGFR-3+ monocytes from whole human citrate blood by FACS, immunolabeled them for VEGFR-3 and examined them by confocal microscopy, as previously described3. For in vitro experiments, we prepared monocytesmacrophages by separation of lymphocytes by sheep erythrocyte rosetting20, followed by adhesion of macrophages on plastic dishes for 2 h. We collected adherent cells and seeded them onto fibronectin-coated chamber slides (Costar), and cultured them in EGM-2MV BulletKit media (Cambrex Bio Science) for 3–13 d. For indirect immunofluorescence, we washed cells in Hanks balanced salt solution, fixed them in methanol-acetone for 5 min at 4 1C, air-dried them, blocked Fc receptors by incubation with Endobulin S/D (human immunoglobulin, 10 mg/ml, Baxter Bioscience), and then incubated them with monoclonal human podoplanin–specific mouse IgG (0.1 mg/ml) and human LYVE-1– specific rabbit IgG (1 mg/ml, AngioBio). For immunofluorescence, we labeled appropriate secondary antibodies with fluorochromes Alexa 488, Alexa 594 or Alexa 633(Molecular Probes). We stained nuclei with DAPI 18860 (Serva). We performed PCR as previously described3. We prepared RNA with TRIreagent (Molecular Research Center) from FACS-sorted monocytes before, 6 and 36 h after culturing in EGM-2MV medium. We used the following primers: GAPDH (which encodes GAPDH) forward 5¢-TGAAGGTCGGAGTCAACG GATTTGGT-3¢, reverse 5¢-CATGTGGGCCATGAGGTCCACCAC-3¢; PDPN (which encodes podoplanin) forward 5¢-TGGAAGGTGTCAGCTCTGCTC-3¢, reverse 5¢-ACGTTGGCAGCAGGGCGTAAC-3¢; XLKD1 (which encodes LYVE1) forward 5¢-GCCAGGTGCTTCAGCCTGGTG-3¢, reverse 5¢-CTTCAGCTTC CAGGCATCGCACGG-3¢; FLT4 (which encodes VEGFR-3) forward 5¢-CAG ACGGGCAGGAGGTGGTGTG-3¢, reverse 5¢-CGGCTGTGACGCGAGTAGAT GC-3¢; PECAM1 (which encodes CD31) forward 5¢-CAGAGGAAAAGGCCC CAATACACT-3¢, reverse 5¢-AACTGGGCATCATAAGAAATCCTG-3¢; KDR (which encodes VEGFR-2) forward 5¢-CGGTCAACAAAGTCGGGAGAGG-3¢, reverse 5¢-TGGCATCATAAGGCAGTCGTTCAC-3¢. We separated the amplified PCR products by electrophoresis in 1% agarose gels and stained them with ethidium bromide. Note: Supplementary information is available on the Nature Medicine website. ACKNOWLEDGMENTS This work was supported in part by the European Union 5th Framework Project ‘‘Chronic Kidney Disease’’ (QLG1-CZ-2000-00619), the 6th Framework Integrated Project ‘‘Lymphangiogenomics’’ (LSGH-2004-503573) and from the Center of Excellence for Clinical and Experimental Oncology (CLEXO) to D.K. We are indebted to K.H. Mu¨ller-Hermelink and H. Einsele, University Wu¨rzburg, and B. Bu¨ltmann, University Tu¨bingen, for their help in allocating the archival tumor tissue blocks. COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. Published online at http://www.nature.com/naturemedicine/ Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions/ 1. Kerjaschki, D. et al. Lymphatic neoangiogenesis in human kidney transplants is associated with immunologically active lymphocytic infiltrates. J. Am. Soc. Nephrol. 15, 603–612 (2004). 2. Saharinen, P., Tammela, T., Karkkainen, M.J. & Alitalo, K. Lymphatic vasculature: development, molecular regulation and role in tumor metastasis and inflammation. Trends Immunol. 25, 387–395 (2004).
233
© 2006 Nature Publishing Group http://www.nature.com/naturemedicine
LETTERS 3. Schoppmann, S. et al. Tumor-associated macrophages express lymphatic endothelial growth factors and are related to peritumoral lymphangiogenesis. Am. J. Pathol. 161, 947–956 (2002). 4. Tepper, O.M. et al. Adult vasculogenesis occurs through in situ recruitment, proliferation, and tubulization of circulating bone marrow-derived cells. Blood 105, 1068– 1077 (2005). 5. Folkman, J. Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat. Med. 1, 27–31 (1995). 6. Shaked, Y. et al. Genetic heterogeneity of the vasculogenic phenotype parallels angiogenesis: Implications for cellular surrogate marker analysis of antiangiogenesis. Cancer Cell 7, 101–111 (2005). 7. Hong, Y.K. et al. Prox1 is a master control gene in the program specifying lymphatic endothelial cell fate. Dev. Dyn. 225, 351–357 (2002). 8. Mazal, P.R. et al. Derivation of nephrogenic adenomas from renal tubular cells in kidney-transplant recipients. N. Engl. J. Med. 347, 653–659 (2002). 9. Breiteneder-Geleff, S. et al. Angiosarcomas express mixed endothelial phenotypes of blood and lymphatic capillaries: podoplanin as a specific marker for lymphatic endothelium. Am. J. Pathol. 154, 385–394 (1999). 10. Jackson, D.G. Biology of the lymphatic marker LYVE-1 and applications in research into lymphatic trafficking and lymphangiogenesis. APMIS 112, 526–538 (2004). 11. Rafii, S. & Lyden, D. Therapeutic stem and progenitor cell transplantation for organ vascularization and regeneration. Nat. Med. 9, 702–712 (2003).
234
12. Asahara, T. et al. VEGF contributes to postnatal neovascularization by mobilizing bone marrow-derived endothelial progenitor cells. EMBO J. 18, 3964–3972 (1999). 13. Benezra, R., Rafii, S. & Lyden, D. The Id proteins and angiogenesis. Oncogene 20, 8334–8341 (2001). 14. Salven, P. et al. VEGFR-3 and CD133 identify a population of CD34+ lymphatic/ vascular endothelial precursor cells. Blood 101, 168–172 (2003). 15. Ito, W.D. & Khmelevski, E. Tissue macrophages: ‘‘satellite cells’’ for growing collateral vessels? A hypothesis. Endothelium 10, 233–235 (2003). 16. Maruyama, K. et al. Inflammation-induced lymphangiogenesis in the cornea arises from CD11b-positive macrophages. J. Clin. Invest. 115, 2363–2372 (2005). 17. Kerjaschki, D. The crucial role of macrophages in lymphangiogenesis. J. Clin. Invest. 115, 2316–2319 (2005). 18. Rajantie, I. et al. Adult bone marrow-derived cells recruited during angiogenesis comprise precursors for periendothelial vascular mural cells. Blood 104, 2084– 2086 (2004). 19. He, Y. et al. Preexisting lymphatic endothelium but not endothelial progenitor cells are essential for tumor lymphangiogenesis and lymphatic metastasis. Cancer Res. 64, 3737–3740 (2004). 20. Plunkett, M. et al. Rosetting of activated human T lymphocytes with autologous erythrocytes. Definition of the receptor and ligand molecules as CD2 and lymphocyte function-associated antigen (LFA-3). J. Exp. Med. 165, 664–676 (1987).
VOLUME 12
[
NUMBER 2
[
FEBRUARY 2006
NATURE MEDICINE
