Advertisement

Expression of Lymphangiogenic Factors and Evidence of Intratumoral Lymphangiogenesis in Pancreatic Endocrine Tumors

      Lymphangiogenesis is thought to promote the progression of malignant tumors. Because the lymphangiogenic factors vascular endothelial factor (VEGF)-C and -D are expressed in endocrine cells, we investigated their expression in pancreatic endocrine tumors (PETs) and correlated these data and intratumoral lymph vessel density (iLVD) with clinicopathological features. Lymph vessels were identified with anti-podoplanin antiserum and with podoplanin/proliferating cell nuclear antigen double labeling. PETs (n = 104) were investigated by immunohistochemical staining for VEGF, basic fibroblast growth factor, and VEGF-C expression. VEGF-C and VEGF-D mRNA were quantified by real-time reverse transcriptase-polymerase chain reaction. PETs showed higher iLVD than normal pancreata, but iLVD did not discriminate between benign and malignant PETs. In PETs proliferating lymph vessels were identified. High iLVD was associated with lymph vessel invasion and it was more frequent in angioinvasive/metastatic tumors than in grossly invasive tumors. VEGF-C expression correlated with iLVD as well as with glucagon and pancreatic polypeptide expression. PETs show intratumoral lymphangiogenesis, which is associated with VEGF-C expression in tumor cells. The association between iLVD and lymph vessel invasion and angioinvasive/metastatic features in PETs suggests that lymphangiogenesis may promote malignant progression of PETs. PET is the first human tumor entity in which VEGF-C-related intratumoral lymphangiogenesis has been demonstrated.
      The biological behavior of pancreatic endocrine tumors (PET) is difficult to predict on the basis of histological criteria. In the absence of clear signs of malignancy, such as invasion of adjacent organs, angioinvasion, or metastasis, the prognosis remains uncertain. Because most human carcinomas metastasize via lymphatic invasion, lymph node metastasis is a key prognostic factor for the clinical outcome. By which mechanism tumor cells spread through the lymph vessels is unknown. One proposed mechanism is the induction of new lymph vessels by tumor or inflammatory cells, facilitating lymphangioinvasion. Intratumoral lymph vessels have been detected in head and neck squamous cell carcinomas
      • Beasley NJ
      • Prevo R
      • Banerji S
      • Leek RD
      • Moore J
      • van Trappen P
      • Cox G
      • Harris AL
      • Jackson DG
      Intratumoral lymphangiogenesis and lymph node metastasis in head and neck cancer.
      • Maula SM
      • Luukkaa M
      • Grenman R
      • Jackson D
      • Jalkanen S
      • Ristamaki R
      Intratumoral lymphatics are essential for the metastatic spread and prognosis in squamous cell carcinomas of the head and neck region.
      and in cutaneous melanomas.
      • Straume O
      • Jackson DG
      • Akslen LA
      Independent prognostic impact of lymphatic vessel density and presence of low-grade lymphangiogenesis in cutaneous melanoma.
      • Dadras SS
      • Paul T
      • Bertoncini J
      • Brown LF
      • Muzikansky A
      • Jackson DG
      • Ellwanger U
      • Garbe C
      • Mihm MC
      • Detmar M
      Tumor lymphangiogenesis: a novel prognostic indicator for cutaneous melanoma metastasis and survival.
      In squamous cell carcinomas intratumoral lymph vessel density (iLVD) has been shown to correlate with lymph node metastasis, whereas the results for melanomas are inconsistent. Beasley and colleagues
      • Beasley NJ
      • Prevo R
      • Banerji S
      • Leek RD
      • Moore J
      • van Trappen P
      • Cox G
      • Harris AL
      • Jackson DG
      Intratumoral lymphangiogenesis and lymph node metastasis in head and neck cancer.
      and Straume and colleagues
      • Straume O
      • Jackson DG
      • Akslen LA
      Independent prognostic impact of lymphatic vessel density and presence of low-grade lymphangiogenesis in cutaneous melanoma.
      showed that intratumoral lymphatic endothelial cells were capable of proliferation, suggesting de novo lymphangiogenesis. No evidence of intratumoral lymph vessels was found in ovarian,
      • Birner P
      • Schindl M
      • Obermair A
      • Plank C
      • Breitenecker G
      • Kowalski H
      • Oberhuber G
      Lymphatic microvessel density in epithelial ovarian cancer: its impact on prognosis.
      liver,
      • Mouta Carreira C
      • Nasser SM
      • di Tomaso E
      • Padera TP
      • Boucher Y
      • Tomarev SI
      • Jain RK
      lYVE-1 is not restricted to the lymph vessels: expression in normal liver blood sinusoids and down-regulation in human liver cancer and cirrhosis.
      breast,
      • Schoppmann SF
      • Birner P
      • Studer P
      • Breiteneder-Geleff S
      Lymphatic microvessel density and lymphovascular invasion assessed by anti-podoplanin immunostaining in human breast cancer.
      • Williams CS
      • Leek RD
      • Robson AM
      • Banerji S
      • Prevo R
      • Harris AL
      • Jackson DG
      Absence of lymphangiogenesis and intratumoural lymph vessels in human metastatic breast cancer.
      or cervical carcinomas.
      • Schoppmann SF
      • Birner P
      • Stockl J
      • Kalt R
      • Ullrich R
      • Caucig C
      • Kriehuber E
      • Nagy K
      • Alitalo K
      • Kerjaschki D
      Tumor-associated macrophages express lymphatic endothelial growth factors and are related to peritumoral lymphangiogenesis.
      Vascular endothelial growth factor (VEGF)-C and VEGF-D are cytokines possessing structural homology with VEGF and platelet-derived growth factor.
      • Joukov V
      • Pajusola K
      • Kaipainen A
      • Chilov D
      • Lahtinen I
      • Kukk E
      • Saksela O
      • Kalkkinen N
      • Alitalo K
      A novel vascular endothelial growth factor, VEGF-C, is a ligand for the Flt4 (VEGFR-3) and KDR (VEGFR-2) receptor tyrosine kinases.
      • Joukov V
      • Sorsa T
      • Kumar V
      • Jeltsch M
      • Claesson-Welsh L
      • Cao Y
      • Saksela O
      • Kalkkinen N
      • Alitalo K
      Proteolytic processing regulates receptor specificity and activity of VEGF-C.
      • Achen MG
      • Jeltsch M
      • Kukk E
      • Makinen T
      • Vitali A
      • Wilks AF
      • Alitalo K
      • Stacker SA
      Vascular endothelial growth factor D (VEGF-D) is a ligand for the tyrosine kinases VEGF receptor 2 (Flk1) and VEGF receptor 3 (Flt4).
      They are secreted glycoproteins, both of which can act through the tyrosine-kinase receptors VEGF-R2 and VEGF-R3.
      • Joukov V
      • Sorsa T
      • Kumar V
      • Jeltsch M
      • Claesson-Welsh L
      • Cao Y
      • Saksela O
      • Kalkkinen N
      • Alitalo K
      Proteolytic processing regulates receptor specificity and activity of VEGF-C.
      • Stacker SA
      • Stenvers K
      • Caesar C
      • Vitali A
      • Domagala T
      • Nice E
      • Roufail S
      • Simpson RJ
      • Moritz R
      • Karpanen T
      • Alitalo K
      • Achen MG
      Biosynthesis of vascular endothelial growth factor-D involves proteolytic processing which generates non-covalent homodimers.
      Because VEGF-R3 is almost exclusively expressed in lymphatic endothelial cells, VEGF-C and VEGF-D stimulate the growth of these cells and thus induce lymphangiogenesis.
      • Joukov V
      • Sorsa T
      • Kumar V
      • Jeltsch M
      • Claesson-Welsh L
      • Cao Y
      • Saksela O
      • Kalkkinen N
      • Alitalo K
      Proteolytic processing regulates receptor specificity and activity of VEGF-C.
      • Stacker SA
      • Stenvers K
      • Caesar C
      • Vitali A
      • Domagala T
      • Nice E
      • Roufail S
      • Simpson RJ
      • Moritz R
      • Karpanen T
      • Alitalo K
      • Achen MG
      Biosynthesis of vascular endothelial growth factor-D involves proteolytic processing which generates non-covalent homodimers.
      • Kukk E
      • Lymboussaki A
      • Taira S
      • Kaipainen A
      • Jeltsch M
      • Joukov V
      • Alitalo K
      VEGF-C receptor binding and pattern of expression with VEGFR-3 suggests a role in lymphatic vascular development.
      • Jeltsch M
      • Kaipainen A
      • Joukov V
      • Meng X
      • Lakso M
      • Rauvala H
      • Swartz M
      • Fukumura D
      • Jain RK
      • Alitalo K
      Hyperplasia of lymphatic vessels in VEGF-C transgenic mice.
      VEGF-C expression has been reported in several types of cancer, such as carcinomas of the esophagus, stomach, colorectum,
      • Duff SE
      • Li C
      • Jeziorska M
      • Kumar S
      • Saunders MP
      • Sherlock D
      • O'Dwyer ST
      • Jayson GC
      Vascular endothelial growth factors C and D and lymphangiogenesis in gastrointestinal tract malignancy.
      head and neck,
      • Beasley NJ
      • Prevo R
      • Banerji S
      • Leek RD
      • Moore J
      • van Trappen P
      • Cox G
      • Harris AL
      • Jackson DG
      Intratumoral lymphangiogenesis and lymph node metastasis in head and neck cancer.
      • Maula SM
      • Luukkaa M
      • Grenman R
      • Jackson D
      • Jalkanen S
      • Ristamaki R
      Intratumoral lymphatics are essential for the metastatic spread and prognosis in squamous cell carcinomas of the head and neck region.
      breast,
      • Valtola R
      • Salven P
      • Heikkila P
      • Taipale J
      • Joensuu H
      • Rehn M
      • Pihlajaniemi T
      • Weich H
      • deWaal R
      • Alitalo K
      VEGFR-3 and its ligand VEGF-C are associated with angiogenesis in breast cancer.
      • Yang W
      • Klos K
      • Yang Y
      • Smith TL
      • Shi D
      • Yu D
      ErbB2 overexpression correlates with increased expression of vascular endothelial growth factors A, C, and D in human breast carcinoma.
      • Nakamura Y
      • Yasuoka H
      • Tsujimoto M
      • Yang Q
      • Tsukiyama A
      • Imabun S
      • Nakahara M
      • Nakao K
      • Nakamura M
      • Mori I
      • Kakudo K
      Clinicopathological significance of vascular endothelial growth factor-C in breast carcinoma with long-term follow-up.
      and cervix
      • Schoppmann SF
      • Birner P
      • Stockl J
      • Kalt R
      • Ullrich R
      • Caucig C
      • Kriehuber E
      • Nagy K
      • Alitalo K
      • Kerjaschki D
      Tumor-associated macrophages express lymphatic endothelial growth factors and are related to peritumoral lymphangiogenesis.
      and in melanomas.
      • Straume O
      • Jackson DG
      • Akslen LA
      Independent prognostic impact of lymphatic vessel density and presence of low-grade lymphangiogenesis in cutaneous melanoma.
      These studies demonstrated that tumor cells or tumor-associated macrophages served as a cellular source of VEGF-C. VEGF-D expression was detected in colorectal,
      • White JD
      • Hewett PW
      • Kosuge D
      • McCulloch T
      • Enholm BC
      • Carmichael J
      • Murray JC
      Vascular endothelial growth factor-D expression is an independent prognostic marker for survival in colorectal carcinoma.
      • Hanrahan V
      • Currie MJ
      • Gunningham SP
      • Morrin HR
      • Scott PA
      • Robinson BA
      • Fox SB
      The angiogenic switch for vascular endothelial growth factor (VEGF)-A, VEGF-B, VEGF-C, and VEGF-D in the adenoma-carcinoma sequence during colorectal cancer progression.
      • Funaki H
      • Nishimura G
      • Harada S
      • Ninomiya I
      • Terada I
      • Fushida S
      • Tani T
      • Fujimura T
      • Kayahara M
      • Shimizu K
      • Ohta T
      • Miwa K
      Expression of vascular endothelial growth factor D is associated with lymph node metastasis in human colorectal carcinoma.
      ovarian,
      • Yokoyama Y
      • Charnock-Jones DS
      • Licence D
      • Yanaihara A
      • Hastings JM
      • Holland CM
      • Emoto M
      • Umemoto M
      • Sakamoto T
      • Sato S
      • Mizunuma H
      • Smith SK
      Vascular endothelial growth factor-D is an independent prognostic factor in epithelial ovarian carcinoma.
      endometrial,
      • Yokoyama Y
      • Charnock-Jones DS
      • Licence D
      • Yanaihara A
      • Hastings JM
      • Holland CM
      • Emoto M
      • Sakamoto A
      • Sakamoto T
      • Maruyama H
      • Sato S
      • Mizunuma H
      • Smith SK
      Expression of vascular endothelial growth factor (VEGF)-D and its receptor, VEGF receptor 3, as a prognostic factor in endometrial carcinoma.
      and breast carcinomas,
      • Yang W
      • Klos K
      • Yang Y
      • Smith TL
      • Shi D
      • Yu D
      ErbB2 overexpression correlates with increased expression of vascular endothelial growth factors A, C, and D in human breast carcinoma.
      • Nakamura Y
      • Yasuoka H
      • Tsujimoto M
      • Yang Q
      • Imabun S
      • Nakahara M
      • Nakao K
      • Nakamura M
      • Mori I
      • Kakudo K
      Prognostic significance of vascular endothelial growth factor D in breast carcinoma with long-term follow-up.
      and in glioblastomas
      • Debinski W
      • Slagle-Webb B
      • Achen MG
      • Stacker SA
      • Tulchinsky E
      • Gillespie GY
      • Gibo DM
      VEGF-D is an X-linked/AP-1 regulated putative onco-angiogen in human glioblastoma multiforme.
      and melanomas.
      • Achen MG
      • Williams RA
      • Minekus MP
      • Thornton GE
      • Stenvers K
      • Rogers PA
      • Lederman F
      • Roufail S
      • Stacker SA
      Localization of vascular endothelial growth factor-D in malignant melanoma suggests a role in tumour angiogenesis.
      In nonneoplastic cells, VEGF-C and VEGF-D were found to be expressed mainly in the neuroendocrine cell system, often in close proximity to fenestrated vessels positive for VEGF-R2 and VEGF-R3. VEGF-C was identified in pituitary cells, pancreatic α-cells, adrenal medullary cells, and serotonin-producing cells of the gastrointestinal tract, whereas VEGF-D was detected in the cortex of the adrenal gland and in gastrin-producing cells.
      • Partanen TA
      • Arola J
      • Saaristo A
      • Jussila L
      • Ora A
      • Miettinen M
      • Stacker SA
      • Achen MG
      • Alitalo K
      VEGF-C and VEGF-D expression in neuroendocrine cells and their receptor, VEGFR-3, in fenestrated blood vessels in human tissues.
      Because malignant PETs frequently exhibit lymphatic invasion and lymph node metastases, we investigated whether these tumors are capable of inducing lymphangiogenesis that may promote tumor progression. On the basis of the distribution pattern of lymphangiogenic factors in normal endocrine cells, we hypothesized that PETs may express VEGF-C or VEGF-D and that this expression is related to lymphangiogenesis and biological features.

      Materials and Methods

      Tissues

      Formalin- or Bouin-fixed paraffin-embedded tissue blocks from 111 PETs from the period 1972 to 2002 were investigated. They were retrieved from the consultation files of the Department of Pathology of the University of Kiel. Only cases with complete patient records, including gender, age, clinical manifestation, size and localization of tumor, and presence of metastasis were considered. Table 1 summarizes the most important clinicopathological characteristics of the investigated PETs. All benign functionally active PETs were insulinomas. Two of them arose in patients with multiple endocrine neoplasia type I. The malignant functionally active PETs consisted of 11 insulinomas, 7 gastrinomas, 4 glucagonomas, 2 VIPomas, and 1 ACTH-producing tumor. Clinical symptoms and elevated serum hormone levels defined functionally active PETs. PETs that were ≤2 cm in diameter and showed no signs of invasive growth or metastasis were considered benign. Malignant PETs showed blood or lymph vessel invasion (detected with endothelial cell-specific immunostains), tumor infiltration of adjacent organs, or histologically verified lymph node or blood-borne metastases. To increase the number of glucagon-expressing tumors, seven glucagon-expressing PETs (one glucagonoma, six functionally inactive) that lacked signs of malignancy but were larger than 2 cm were also included in the study. Immunohistochemically, all tumors expressed synaptophysin and/or chromogranin and were well differentiated according to the World Health Organization classification.
      • Solcia E
      • Kloppel G
      • Sobin LH
      Table 1Clinicopathological Features of 111 PETs
      GenderAge (years)Tumor size
      Maximum diameter in cm.
      No. of tumorsFMMedianSDRangeMedianSDRange
      Functionally active
          Benign3219134819.617–881.50.420.5–2
          Malignant2514115115.224–7942.462–11
      Functionally inactive
          Benign7526114.835–7521.670.7–6
      This group contains one large cystic tumor (6 cm) that had already been noted 10 years before the surgical resection and showed no signs of malignant transformation.
          Malignant4023175714.23–7553.032–12
      Probably malignant
      This group was included only to test the correlation between glucagon expression, VEGF-C expression, and LVD.
      7525014.717–5953.62.2–12
      Total11166455416.263–8833.040.5–12
      * Maximum diameter in cm.
      This group contains one large cystic tumor (6 cm) that had already been noted 10 years before the surgical resection and showed no signs of malignant transformation.
      This group was included only to test the correlation between glucagon expression, VEGF-C expression, and LVD.
      Freshly cryoconserved tumor tissue from six pancreatic PETs served as source of reference RNA for the quantitative polymerase chain reaction (PCR) investigations. These tumors were collected according to the protocol approved by the ethics committee of the University of Kiel Hospitals (permission number, 110/99). This protocol prescribed signed informed consent by all patients.
      Normal pancreatic tissues removed from 12 individuals who died of suicide were obtained from the Department of Forensic Medicine of Semmelweis University (Budapest, Hungary). The pancreata were removed within 1 hour after death with the permission of the local Ethical Commission for Scientific Research (permission number, 140-1/1996) and immediately fixed in 4% buffered formalin. Histological examination of the pancreas tissues revealed no signs of pancreatic disease. During the forensic autopsy no signs of organic disease affecting the pancreas were seen.

      Immunohistochemistry

      Tumor tissues were fixed in formalin or in Bouin's fixative and routinely processed for paraffin sectioning. Three-μm-thin paraffin sections were deparaffinized, rehydrated, and immunohistochemical stains were performed according to routine methods. Before application of the primary antibody, blocking with nonimmune serum was performed for 20 minutes. Table 2 lists the positive controls, demasking methods, antibodies, and detection systems. For negative controls, the primary antibodies were omitted or the primary anti-basic fibroblast growth factor antibody was incubated with specific blocking peptide in 10-fold molar excess before the staining. For the VEGF-D staining, the primary antibody was replaced with control mouse IgG2a (DAKO Cytomation, Glostrup, Denmark).
      Table 2Antibodies, Conditions, and Controls for Immunohistochemical Reactions
      AntigenPrimary antibody clone/codeSourceConcentration/dilutionAntigen demaskingSecondary antibodyDetection systemPositive control
      SynaptophysinRabbit/A0010Dakocytomation, Glostrup, Denmark6 μg/mlPressure cooker, TEC buffer 3 minutesRabbit Vectastain peroxidase kit; Vector Laboratories, Burlingame, CAPancreas
      ChromograninLK2H10/E001LINARIS, Wertheim-Bettingen, Germany1:2Mouse Vectastain peroxidase kit, Vector LaboratoriesPancreas
      InsulinHB125/02911Biogenex, San Ramon, CA1:40Mouse Vectastain peroxidase kit, Vector LaboratoriesPancreas
      GlucagonRabbit/039PBiogenex1:60Rabbit Vectastain peroxidase kit, Vector LaboratoriesPancreas
      Pancreatic polypeptideRabbitR.E. Chance, Indianapolis, IN1:5000Rabbit Vectastain peroxidase kit, Vector LaboratoriesPancreas
      SomatostatinRabbit/A0566DAKO1:200Rabbit Vectastain peroxidase kit, Vector LaboratoriesPancreas
      GastrinRabbitPaesel, Frankfurt, Germany1:3000Rabbit Vectastain peroxidase kit, Vector LaboratoriesDuodenum
      Vasoactive intestinal peptideRabbit/18-0080ZYMED, San Francisco, CA1:10Rabbit Vectastain peroxidase kit, Vector LaboratoriesFunctionally active VIPoma
      CD34QBEND10/0786Immunotech, Marseille, France1 μg/mlMouse Vectastain peroxidase kit, Vector LaboratoriesEndothelial cells
      PodoplaninRabbit321:2000Pressure cooker, TEC buffer, 3 minutesRabbit Envision kit, DAKOLymphangiom
      VEGFGoat/AF293 NAR&D Systems, Wiesbaden, Germany10 μg/mlPressure cooker, TEC buffer, 3 minutesBiotin conjugated anti-goat IgG, 705-065-147; Jackson Laboratories, West Grove, PAABC Elite Kit, Vector LaboratoriesFetal kidney
      VEGF-CGoat/AF752R&D Systems0.5 μg/mlPressure cooker, Glycol containing solution,
      Biologo, Kiel, Germany.
      5 minutes
      Biotin conjugated anti-goat IgG, 705-065-147; Jackson LaboratoriesABC Elite Kit, Vector Laboratoriesα-Cells in Langerhans-islets
      VEGF-DMouse/MAB 286R&D Systems15 μg/ml, overnightPressure cooker, Glycol containing solution,
      Biologo, Kiel, Germany.
      5 minutes
      Mouse Envision kit, DakoCytomationGastrin-producing cells of stomach
      bFGFRabbit/sc-79Santa Cruz Biotechnology, Santa Cruz, CA, USA1 μg/mlPressure cooker, TEC buffer, 3 minutesRabbit Vectastain peroxidase kit, Vector LaboratoriesChronic pancreatitis
      * Biologo, Kiel, Germany.
      Expression of hormones, VEGF, VEGF-C, VEGF-D, and bFGF was semiquantitatively evaluated as mild (<10% of tumor cells stained), moderate (10 to 50%), or strong (>50%) by one investigator (B.S.) without knowledge of the other tumor features. Hormone expression was carefully analyzed in consecutive sections to avoid misinterpretation of entrapped pancreatic islets. iLVD and peritumoral lymph vessel density (pLVD) were determined by the hot spot method.
      • Weidner N
      • Semple JP
      • Welch WR
      • Folkman J
      Tumor angiogenesis and metastasis—correlation in invasive breast carcinoma.
      All areas (0.75 mm2) with high lymph vessel density stained with anti-podoplanin antiserum were photographed with a Nikon Coolpix 990 digital camera mounted on an Axioskop 50 microscope (Zeiss, Oberkochen, Germany). Intratumoral hot spots in all PETs and peritumoral hot spots (areas in the fibrous capsule of the tumor or in peritumoral tissues without close relationship to tumor cells) in 80 PETs were evaluated. The localization of the 349 digitalized intratumoral hot spots was recorded as peripheral (within 4 mm of the tumor border) or central. Podoplanin-positive vessels with or without lumen were counted with Scion Image Software (Scion Corp., Frederick, MD). Vessels were considered to be invaded if vascular spaces lined by podoplanin-positive (lymph vessel) or CD34-positive (blood vessel) endothelial cells contained definite tumor cell complexes. LVD was determined with the same method in 12 normal pancreata (one section from the head, body, and tail from each pancreas) and peripancreatic fat tissues (one section from each pancreas). Overall 244 hot spots were analyzed in normal pancreatic and peripancreatic tissue.

      Double Immunolabeling

      Tissue sections were incubated in 1% sodium borohydride in phosphate-buffered saline for 30 minutes to reduce autofluorescence and subsequently stained with 1% Sudan Black B (Sigma, Taufkirchen, Germany) for 10 minutes. Nonspecific binding sites were blocked in blocking buffer containing 0.1% bovine serum albumin and 0.2% glycine in Tris-buffered saline for 1 hour at room temperature. Incubation of antibodies was performed overnight in Tris-buffered saline. Primary antibody dilutions were as follows: rabbit anti-podoplanin antiserum, 1:400; monoclonal anti-proliferating-cell nuclear antigen (PCNA) (Santa Cruz Biotechnology, Santa Cruz, CA), 1:300; mouse monoclonal anti-glucagon antibody, 1:30; and goat anti-VEGF-C antibody, 1:50. After extensive washing in Tris-buffered saline, secondary antibodies were incubated for 1 hour at 37°C (goat anti-rabbit Alexa Fluor 488, 1:1000; goat anti-mouse Alexa Fluor 546, 1:1000; donkey anti-goat Alexa Fluor 488, 1:1000; all from Molecular Probes, Eugene, OR). Sections were counterstained for nuclei using Hoechst A33528 dye at a concentration of 0.025 mg/ml. After extensive washing in Tris-buffered saline, sections were mounted using ProLong anti-fade mounting medium (Molecular Probes). Laser-scanning analysis was performed using a Zeiss LSM510 confocal laser-scanning microscope (Carl Zeiss Jena, Jena, Germany). All recordings were done using multitracking with ×400 original magnification and pinhole diameter set at 1.0 Airy unit.
      For colorometric double labeling, PCNA immunostaining was developed with alkaline phosphatase-conjugated secondary antibody and podoplanin staining was developed with peroxidase-conjugated secondary antibody using an Envision double-labeling kit (DAKO Cytomation) according to the manufacturer's instructions. The rate of the double-labeled vessels was determined by counting the nuclei of intratumoral podoplanin-positive microvessels (100 to 350 nuclei in each tumor).

      RNA Extraction and Quantitative PCR

      Paraffin blocks from 29 formalin-fixed PETs, which consisted of >90% tumor tissue were selected for quantitative PCR. Immunohistochemically, 10 tumors showed no or low VEGF-C expression; 19 tumors revealed moderate or strong VEGF-C positivity. Ten 5-μm-thin sections were cut into tubes, deparaffinized, washed in ethanol, air-dried, and resuspended in 80 μl of 60 mg/ml Proteinase K (Sigma) plus 720 μl of digestion buffer as described by Godfrey and colleagues
      • Godfrey TE
      • Kim SH
      • Chavira M
      • Ruff DW
      • Warren RS
      • Gray JW
      • Jensen RH
      Quantitative mRNA expression analysis from formalin-fixed, paraffin-embedded tissues using 5′ nuclease quantitative reverse transcription-polymerase chain reaction.
      overnight at 55°C in a shaker. RNA extraction was performed with an RNAeasy kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. To remove remnants of genomic DNA, extracted RNA was treated with RNase-free DNase Set (Qiagen) for 20 minutes. The RNA concentration was determined by spectrophotometer (Spectramax 250; Molecular Devices, Sunnyvale, CA) at 260 nm. The purity of the RNA extracts was monitored by OD 260/280 measurement. RNA extracts were aliquoted and stored at −20°C. The cryoconserved tissues were cut into pieces, transferred to tubes, homogenized by pellet pestles, pushed through a 21-gauge needle, and subjected to RNA extraction with the same method.
      Primers were designed to span introns in the genomic DNA. Primers and probes for quantitative PCR are listed in Table 3. The levels of VEGF-C and VEGF-D mRNA were quantified by real-time reverse transcriptase (RT)-PCR using an ABI Prism 7000 sequence detection system (Applied Biosystem, Foster City, CA). One-step RT-PCR (QuantiTect Probe RT-PCR, Qiagen) was performed using 100 ng of RNA, 10 pmol of forward and reverse primer, respectively, and 3 pmol of probe in each reaction. The mRNA for VEGF-C and VEGF-D and the control GAPDH were each amplified in separate tubes. VEGF-C and VEGF-D quantification was performed in triplicate and in duplicate, respectively. DNA for the PCR standard was obtained from RNA from fresh PET tissue by RT-PCR using the same primers as for the quantitative RT-PCR. Standard curves were constructed with 10-fold serial dilutions of gel-purified DNA. After performing a linear regression analysis for the standard dilutions the values for the experimental samples were extrapolated and expressed as corresponding attomol of the standard DNA. The values in attomol for VEGF-C and VEGF-D were divided by the value for GAPDH.
      Table 3Oligonucleotide Primers and Probes Used for Quantitative PCR
      Target5′ Primer3′ PrimerProbeSize of amplicon (bp)Annealing temperature
      VEGF-C15′-TCAAGGACAGAAGAGACTATAAAATTTGC5′-ACTCCAAACTCCTTCCCCACAT6FAM-ATACACACCTCCCGTGGCATGCATTGT-TAMRA13760
      VEGF-D5′-TACTCTTCCCCAGCTCACTG5′-TACTCTTCCCCAGCTCACTG6FAM-CAAAGAACTCAGTGCAGCCCTAGAGAAACGTTAMRA11660
      GAPDH315′-ATTCCACCCATGGCAAATTC5′-TGGGATTTCCATTGATGACAAGVIC-CAAGCTTCCCGTTCTCAGCC-TAMRA7260

      Statistics

      The association between LVDs (continuous variable) and clinicopathological features or cytokine expression was analyzed with the Mann-Whitney U-test and the Kruskal-Wallis H-test. Correlations between VEGF-C expression and other tumor parameters were determined with Fisher's exact test. iLVD and mRNA expression levels determined by quantitative PCR were divided into quartiles before their association with VEGF-C protein expression was tested using a one-sided Kendall-Tau b test. Comparisons between continuous variables were made with Spearman's rank test. Linear regression analysis was applied for testing LVD influencing factors. Logistic regression analysis was performed to test cause variables for VEGF-C expression. P values <0.05 were considered to be significant. All statistical tests were performed using SPSS 10.1 software (SPSS, Chicago, IL).

      Results

      Lymph Vessel Density in PETs and Normal Pancreatic and Peripancreatic Tissues

      Podoplanin marked thin-walled vessels
      • Breiteneder-Geleff S
      • Soleiman A
      • Kowalski H
      • Horvat R
      • Amann G
      • Kriehuber E
      • Diem K
      • Weninger W
      • Tschachler E
      • Alitalo K
      • Kerjaschki D
      Angiosarcomas express mixed endothelial phenotypes of blood and lymphatic capillaries: podoplanin as a specific marker for lymphatic endothelium.
      without recognizable muscle layer or pericytes corresponded morphologically to lymph vessels whereas other cell types did not show podoplanin expression. The vast majority of the lymph vessels in normal pancreata had wide open lumina and were in close proximity to ducts, blood vessels, or nerves (Figure 1a). The majority of the intratumoral lymph vessels were collapsed. Lymph vessels were unevenly distributed throughout the tumor. Eighty-five percent of the lymph vessel hot spots were located in a rim of 4 mm at the periphery of the tumors (Figure 1b).
      Figure thumbnail gr1
      Figure 1a and b: Lymph vessels marked by podoplanin immunolabeling in normal pancreas (a) and in a PET (b). c: Lymph vessel invasion by tumor cells in a malignant PET (podoplanin immunostain). d and e: Double immunolabeling of proliferating lymph vessels (arrows) marked by chromogenic substrates for PCNA (red) and podoplanin (brown) and by fluorochromes for PCNA (red) and podoplanin (green). f: VEGF-C immunolabeling in a peritumoral Langerhans islet (L), note the glucagon cell pattern (arrows). g: VEGF-C expression in a PET. h: Fluorescent double labeling of VEGF-C (green) and glucagon (red) in a glucagon-producing PET; note the cellular co-localization (yellow). i: VEGF-D expression in a PET; inset: positive control for VEGF-D in endocrine cells of the gastric mucosa. j: Negative control for VEGF-D immunolabeling replacing the primary antibody with nonspecific mouse IgG2; inset: the same negative control in the gastric mucosa stained with control mouse immunoglobulin (see Materials and Methods). k: VEGF-A expression in a PET. l: bFGF expression in a PET, significant immunolabeling of endothelial (arrows) and tumor cells (T). Original magnifications: ×200 (a, b); ×400 (i, j); ×500 (g); ×800 (c, f, k, l); ×1000 (d, e, h, and insets in i, j).
      Median LVDs for normal pancreas were 11.5/0.75 mm2 (mean, 11.9; SD, 3.3; range, 6 to 19), for peripancreatic adipose tissue 8.5/0.75 mm2 (mean, 9.4; SD, 3; range, 6 to 14), iLVD 16/0.75 mm2 (mean, 20.4; SD, 16.9; range 0 to 87), and pLVD 14/0.75 mm2 (mean, 15.5; SD, 7.6; range 4 to 45) for all PETs. Comparison of LVDs in normal tissues versus tumors revealed a significant increase of both iLVD (Mann-Whitney U-test, P = 0.004) and pLVD (Mann-Whitney U-test, P = 0.001) in PETs. Tumor size was not related to iLVD (Kendall-Tau b test; correlation coefficient, −0.31; P = 0.675), but showed a positive correlation with pLVD when tumor size was divided into quartiles and tested with Kendall-Tau b test (correlation coefficient, 0.295; P = 0.001). There was no correlation between iLVD and pLVD (Spearman's rank test, P = 0.582). Eleven PETs did not show intratumoral podoplanin-expressing lymph vessels. In these cases lymphatics in the surrounding tissue served as positive control to rule out failure of the immunohistochemical detection method.
      Table 4 summarizes the correlations between LVDs and clinicopathological features and hormone and cytokine expression. iLVD showed no correlation with biological behavior or functional activity of PETs (Kruskal-Wallis analysis, P = 0.96). High to moderate glucagon expression correlated with iLVD (Mann-Whitney U-test, P < 0.001), but no other hormone expression in the tumor cells did. If the glucagon-expressing tumors were divided into expressing and nonexpressing subgroups, the significance of the glucagon-iLVD association remained high (Mann-Whitney U-test, P < 0.001). Similar results were obtained if the seven additional, probably malignant, glucagon-expressing PETs were included in the analysis. There was a clear association between iLVD and the presence of lymph vessel invasion (Mann-Whitney U-test, P = 0.004) (Figure 1c), but not of lymph node metastasis (Mann-Whitney U-test, P = 0.39). Interestingly, iLVD exhibited a significant correlation with VEGF-C expression (Mann-Whitney U-test, P < 0.001), without being associated with VEGF or bFGF expression. In a multivariate linear regression analysis, moderate to strong glucagon expression (r2 = 0.21; P < 0.001; B value, 28.7; 95% confidence intervals, 17.7 to 39.7) and VEGF-C expression (r2 = 0.064; P = 0.003; B value, 11.5; 95% confidence intervals, 3.9 to 19.2) remained independent factors of high iLVD.
      Table 4Lymph Vessel Densities Related to Clinicopathological Features and Cytokine Expression in 104 PETs
      Intratumoral lymphatic vessel densityPeritumoral lymphatic vessel density
      No. of tumorsLVD (mean)SDP value
      Mann-Whitney U-test, if not otherwise specified.
      No. of tumorsLVD (mean)SDP value
      Mann-Whitney U-test, if not otherwise specified.
      Biological behavior
          Benign all3919.5143012.85.9
          Malignant all6520.918.40.925117.280.004
      Functionally active
          Benign322014.82511.23.5
          Malignant2520.314.81914.44.8
      Functionally inactive
          Benign717.110.6520.89
          Malignant4021.220.60.96
      Kruskal-Wallis test.
      3218.99.10.001
      Kruskal-Wallis test.
      Lymph vessel invasion
      Malignant tumors only.
          Absent4516.6163717.68.3
          Present2030.420.30.0041416.17.10.69
      Lymph node metastasis
      Malignant tumors only.
          Absent4122.920.43415.76.7
          Present2417.414.10.391720.29.70.14
      Hormone expression in tumor cells
          Insulin
              Absent/weak6921.118.35417.67.9
              Moderate/strong3518.913.60.932711.34.3<0.001
          Glucagon
              Absent/weak9618.214.47415.67.9
              Moderate/strong846.922.1<0.001715.32.90.65
              Absent6415.913.64814.78.6
              Present4027.619.10.0013316.85.60.108
          Glucagon
      Seven glucagon-expressing PETs of undetermined malignant potential were added.
              Absent/weak9918.914.97715.47.8
              Moderate/strong1237.424.30.0061121.39.30.03
              Absent6415.913.64814.78.6
              Present4727.718.90.0014017.97.30.06
          Somatostatin
              Absent/weak9621.217.27515.27.5
              Moderate/strong810.15.90.076619.87.50.12
          Pancreatic polypeptide
              Absent/weak9420.617.27115.17.5
              Moderate/strong918.214.70.83918.27.50.23
          Gastrin
              Absent/weak10020.517.17915.67.6
              Moderate/strong41611.50.832138.40.73
      VEGF-A expression
          Absent/weak2621.314.61813.16.2
          Moderate/strong7820.317.50.396316.37.90.09
      VEGF-C expression
          Absent/weak8317.214.56214.17.1
          Moderate/strong2134.118.7<0.0011920.47<0.001
      bFGF expression in tumor cells
          Absent/weak6618.214.14915.28
          Moderate/strong3824.620.20.183116.27.10.38
      bFGF expression in endothelial cells
          Absent/weak6020.815.564214.66.4
          Moderate/strong4420.318.620.63816.68.70.37
      * Mann-Whitney U-test, if not otherwise specified.
      Kruskal-Wallis test.
      Seven glucagon-expressing PETs of undetermined malignant potential were added.
      § Malignant tumors only.
      When the malignant PETs were divided into groups consisting of tumors that infiltrated surrounding organs without showing other signs of malignancy (n = 7; largest diameter, 7.5 ± 3 cm) and tumors that were angioinvasive/metastatic (n = 58; largest diameter, 5.17 ± 2.77 cm), no tumors in the group consisting of grossly invasive tumors revealed more than 20 iLVD, whereas in the second group 26 of 58 (44.8%) tumors had iLVD values >20 (P = 0.022, Fisher's exact test).
      Peritumoral LVD values were significantly higher in malignant than in benign PETs (Mann-Whitney U-test, P = 0.004), in functionally inactive versus active PETs (Kruskal-Wallis test, P = 0.001), in PETs showing high VEGF-C expression versus PETs with no or low VEGF-C expression (Mann-Whitney U-test, P = <.001), and in PETs with no or low insulin expression versus moderate-to-strong insulin expression (Mann-Whitney U-test, P = <.001). Multivariate linear regression analysis revealed that functional inactivity (r2 = 0.19; P < 0.001; B value, 6.6; 95% confidence intervals, 3.5 to 9.6), moderate-to-strong VEGF-C expression (r2 = 0.065; P = 0.011; B value, 4.7; 95% confidence intervals, 1.1 to 8.3), and large tumor size (r2 = 0.039; P = 0.041; B value, 1.6; 95% confidence intervals, 0.07 to 3.2) increased pLVD independently. pLVD was not related to lymph vessel invasion, lymph node metastasis, and angioinvasive/metastatic phenotype (Mann-Whitney U-test, P = 0.312). To detect and quantify proliferating lymph vessels, 27 PETs having high iLVD (>25/0.75 mm2) were investigated with podoplanin and PCNA immunohistochemistry (Figure 1d). Double-labeled lymph vessels were found in 11 of 27 tumors. The proliferation rate ranged from 0.7 to 3%. We confirmed the co-expression of PCNA and podoplanin by immunofluorescent labeling in eight of nine PETs showing high iLVD (Figure 1e).

      VEGF-C, VEGF-D, VEGF, and bFGF Expression

      Immunohistochemically, mainly tumor cells and some peritumoral pancreatic islets exhibited VEGF-C expression (Figure 1, f and g). Twenty-one of one hundred four PETs (20.1%) expressed VEGF-C moderately or strongly. Table 5 summarizes the associations between VEGF-C and tumor features. VEGF-C expression was positively correlated with malignant phenotype (Fisher's exact test, P = 0.002), moderate to strong glucagon expression (Fisher's exact test, P = 0.001), and moderate to strong pancreatic polypeptide expression (Fisher's exact test, P = 0.016) in tumor cells and inversely correlated with insulin expression (Fisher's exact test, P = 0.001). The correlation between VEGF-C and glucagon expression remained strong (Fisher's exact test, P < 0.001) when the tumors were divided into glucagon-negative and -positive subgroups or when seven additional probably malignant PETs were also included in the analysis. No association was found between VEGF-C expression and functional status (Fisher's exact test, P = 0.07), lymph vessel invasion (Fisher's exact test, P = 0.16), or lymph node metastasis (Fisher's exact test, P = 0.39).
      Table 5VEGF-C Expression Related to Clinicopathological Features
      VEGF-C, absent/weak No. of tumorsVEGF-C, moderate/strongP value
      Fisher's exact test.
      Biological behavior
          Benign all37 (35.6%)2 (5.9%)
          Malignant all46 (44.2%)19 (18.3%)0.002
      Functionally
          active49 (47.1%)8 (7.7%)
          inactive34 (32.7%)13 (12.5%)0.07
      Lymph vessel invasion
      Malignant tumors only.
          absent34 (52.3%)11 (16.9%)
          present12 (18.5%)8 (12.3%)0.16
      Lymph node metastasis
      Malignant tumors only.
          absent28 (43.1%)13 (20%)
          present18 (27.7%)6 (9.2%)0.39
      Hormone expression in tumor cells
          Insulin
              Absent/weak49 (47.1%)20 (19.2%)
              Moderate/strong34 (32.7%)1 (1%)0.001
          Glucagon (n = 104)
              Absent/weak81 (77.9%)15 (14.4%)
              Moderate/strong2 (1.9%)6 (5.8%)=0.001
              Absent59 (56.7%)5 (4.8%)
              Present24 (23.1%)16 (15.4%)<0.001
          Glucagon (n = 111)
      Seven glucagon-expressing PETs of undetermined malignant potential were added.
              Absent/weak82 (73.9%)17 (15.3%)
              Moderate/strong3 (2.7%)9 (8.1%)<0.001
              Absent59 (53.2%)5 (4.5%)
              Present26 (23.4%)21 (18.9%)<0.001
          Somatostatin
              Absent/weak77 (74%)19 (18.3%)
              Moderate/strong6 (5.8%)2 (1.9%)0.508
          Pancreatic polypeptide
              Absent/weak78 (75.7%)16 (15.5%)
              Moderate/strong4 (3.9%)5 (4.9%)0.016
          Gastrin
              Absent/weak79 (76%)21 (20.2%)
              Moderate/strong4 (3.8%)0.4
      * Fisher's exact test.
      Seven glucagon-expressing PETs of undetermined malignant potential were added.
      Malignant tumors only.
      Multivariate logistic regression analysis showed that moderate to strong glucagon expression (P = 0.001; B value, 28.4; 95% confidence interval, 4.1 to 194.5), moderate to strong pancreatic polypeptide expression (P = 0.007; B value, 10.2; 95% confidence interval, 1.87 to 56.2), weak glucagon expression (P = 0.029; B value, 4.4; 95% confidence interval, 1.1 to 16.8), and malignant phenotype (P = 0.052; B value, 7.1; 95% confidence interval, 1 to 52.2) were independent cause variables of VEGF-C expression, but insulin expression turned out not to be independent. Immunofluorescent double staining of glucagon and VEGF-C revealed cellular co-expression of glucagon and VEGF-C in 8 of 10 investigated PETs. However, not all VEGF-C-expressing tumor cells were positive for glucagon (Figure 1h). Low-grade VEGF-D expression was detectable in 3 of the investigated 31 PETs showing iLVD >25/0.75 mm2, whereas the other tumors did not express VEGF-D (Figure 1, i and j).
      VEGF was expressed in tumor cells (Figure 1k), in some endothelial cells and macrophages, and in islets of the peritumoral pancreas. The majority of the PETs (78 of 104, 75%) expressed VEGF in more than 10% of the tumor cells. bFGF expression was seen in the cytoplasm and nuclei of tumor cells (moderate to strong in 38 of 104 PETs) (Figure 1l). After preadsorption with specific blocking peptide no signal was detected. Endothelial cells showed strong to moderate bFGF expression in 44 of 104 tumors, as stromal cells and peritumoral fibroblasts often did.

      Quantitative PCR for VEGF-C and VEGF-D mRNA Detection

      RNA extracted from paraffin-embedded tissue blocks is known to be considerably degraded.
      • Godfrey TE
      • Kim SH
      • Chavira M
      • Ruff DW
      • Warren RS
      • Gray JW
      • Jensen RH
      Quantitative mRNA expression analysis from formalin-fixed, paraffin-embedded tissues using 5′ nuclease quantitative reverse transcription-polymerase chain reaction.
      To estimate the degree of degradation, we compared the expression of the housekeeping gene GAPDH in six freshly cryoconserved PET tissues with that in 29 paraffin-embedded tissue blocks. The GAPDH copy number was 10 times lower in paraffin embedded tissues (mean, 1.79 attomol/100 ng; SD, 2.56 attomol/100 ng) than in snap-frozen tissue (mean, 16.7 attomol/100 ng; SD, 25.8 attomol/100 ng). This reduction of detectable mRNA copy numbers in paraffin-embedded material was lower than previously reported
      • Godfrey TE
      • Kim SH
      • Chavira M
      • Ruff DW
      • Warren RS
      • Gray JW
      • Jensen RH
      Quantitative mRNA expression analysis from formalin-fixed, paraffin-embedded tissues using 5′ nuclease quantitative reverse transcription-polymerase chain reaction.
      (one-thirtieth signal intensity when paraffin-embedded material was compared with fresh tissue RNA) and indicated good RNA recovery and amplification efficacy in our assay. The application of appropriate proteinase K predigestion, usage of primers resulting in short amplicons (<140 bp), and application of a dilution series of gel-purified external standards for the PCR reaction ensured high sensitivity and reliability of quantification.
      Three independent PCR runs detecting VEGF-C mRNA were analyzed using Spearman's rank correlation analysis. The resulting correlation coefficients were 0.72, 0.8, and 0.91. The median VEGF-C mRNA expression was 13 copies/1 attomol GAPDH (mean, 2080 copies; SD, 1070; range, 0.07 to 6410 copies; mean without one outlier, 80 copies/1 attomol GAPDH). The mean mRNA levels of three PCR runs correlated with VEGF-C immunohistochemical results (mean of PCR results divided into quartiles and tested with Kendall-Tau b test, correlation coefficient 0.482, P = 0.003) and showed an association with iLVD (quartiles of mean iLVD compared with quartiles of mean PCR results, Kendall-Tau b test, correlation coefficient 0.355, P = 0.023). VEGF-D mRNA expression was below the detection limit in 27 tumors. Only in two tumors were low levels of VEGF-D measurable (0.16 to 0.84 copies/1 attomol GAPDH).

      Discussion

      We found that both benign and malignant PETs exhibited higher iLVD values than normal pancreata. As the strongest evidence for intratumoral lymphangiogenesis, we detected proliferating lymphatic endothelial cells in PETs with high iLVD. Angioinvasive/metastatic tumors more frequently revealed higher iLVD values than grossly infiltrating PETs. Moreover, high iLVD correlated with lymph vessel invasion in malignant PETs and with VEGF-C expression. Interestingly, VEGF-C expression was associated with glucagon and pancreatic polypeptide production in tumor cells.
      For the specific immunohistochemical recognition of lymph vessels, we used an antiserum against podoplanin, a marker of lymphatic endothelial cells that was shown to co-localize with LYVE-1, a CD44-related hyaluronan receptor expressed mainly in lymphatic endothelial cells
      • Banerji S
      • Ni J
      • Wang SX
      • Clasper S
      • Su J
      • Tammi R
      • Jones M
      • Jackson DG
      LYVE-1, a new homologue of the CD44 glycoprotein, is a lymph-specific receptor for hyaluronan.
      • Prevo R
      • Banerji S
      • Ferguson DJ
      • Clasper S
      • Jackson DG
      Mouse LYVE-1 is an endocytic receptor for hyaluronan in lymphatic endothelium.
      . Apart from lymphatic endothelial cells, podoplanin was also found to be expressed in glomerular podocytes and occasionally in tumor-associated myofibroblasts.
      • Schoppmann SF
      • Birner P
      • Stockl J
      • Kalt R
      • Ullrich R
      • Caucig C
      • Kriehuber E
      • Nagy K
      • Alitalo K
      • Kerjaschki D
      Tumor-associated macrophages express lymphatic endothelial growth factors and are related to peritumoral lymphangiogenesis.
      In our study, the latter finding was encountered in three PETs, which were therefore excluded from further evaluation. The podoplanin antiserum identified lymph vessels in normal pancreatic tissue and in most (90%) of the PETs. The lymphatic channels encountered in the PETs were mainly small and scattered and differed markedly from the wide lymph vessels of the normal pancreas. Their density was significantly higher than in normal pancreatic tissue. Both findings suggested that the detected lymphatics in PET were newly developed vessels rather than entrapped pre-existing vessels. To find further evidence for this assumption, we used double labeling for podoplanin and the PCNA. Using this method we detected proliferating lymphatic endothelial cells in 11 of 27 PETs showing high iLVD. The proliferation rate of lymph vessels ranged between 0.7% and 3%, which can be regarded as low-grade lymphangiogenesis compared to data from angiogenesis research. Evidence for proliferating lymph vessels has so far only been reported in melanomas
      • Straume O
      • Jackson DG
      • Akslen LA
      Independent prognostic impact of lymphatic vessel density and presence of low-grade lymphangiogenesis in cutaneous melanoma.
      • Dadras SS
      • Paul T
      • Bertoncini J
      • Brown LF
      • Muzikansky A
      • Jackson DG
      • Ellwanger U
      • Garbe C
      • Mihm MC
      • Detmar M
      Tumor lymphangiogenesis: a novel prognostic indicator for cutaneous melanoma metastasis and survival.
      and squamous cell carcinomas of the head and neck region.
      • Beasley NJ
      • Prevo R
      • Banerji S
      • Leek RD
      • Moore J
      • van Trappen P
      • Cox G
      • Harris AL
      • Jackson DG
      Intratumoral lymphangiogenesis and lymph node metastasis in head and neck cancer.
      Thus PETs represent the third human tumor type showing intratumoral lymphangiogenesis.
      • Maula SM
      • Luukkaa M
      • Grenman R
      • Jackson D
      • Jalkanen S
      • Ristamaki R
      Intratumoral lymphatics are essential for the metastatic spread and prognosis in squamous cell carcinomas of the head and neck region.
      The iLVD values detected in the 104 PETs in our series ranged from 0 to 87/0.75 mm2. When these values were correlated with the tumors’ biological behavior, it was found that they did not discriminate between benign and malignant PETs. Even small insulinomas commonly had high iLVD values. These findings suggest that intratumoral lymphangiogenesis per se may be unrelated to malignant growth. However, an association of high iLVD with aggressive behavior was found when malignant PETs were divided into ones that were grossly invasive only and ones that were angioinvasive and/or metastatic. Grossly invasive tumors infiltrated surrounding organs without showing other signs of malignancy. Although in our series they were significantly larger than angioinvasive or metastatic tumors, their iLVD values did not exceed 20/0.75 mm2. In contrast, 26 of 58 (44.8%) angioinvasive/metastatic tumors revealed iLVD values >20 (P = 0.022). This may indicate that induction of lymphangiogenesis increases the angioinvasive/metastatic capacity of PETs. In line with this suggestion is the observation that tumor cells that invaded lymphatic spaces were significantly more often detected in PETs with a high iLVD than in PETs with low iLVD. This finding supports the hypothesis that an increase in tumor-related lymph vessels may contribute to tumor progression. The fact that high iLVD values were not associated with lymph node metastasis may have its explanation in the techniques used for surgical removal of the tumors. During the years in which a considerable number of the PETs in our series were collected, they did not necessarily include systematic lymphadenectomy. Alternatively, one may speculate that intratumoral lymph vessels had only limited access to the draining lymphatic system of the peritumoral tissue or that penetration of lymph vessels and the formation of lymph node metastases represent two different biological processes.
      Although pLVD was increased compared to normal pancreas, pLVD was not related to iLVD, lymph vessel invasion, lymph node status, and angioinvasive/metastatic phenotype. These results were in accordance with the observation that peritumoral lymph vessel invasion was only seen in one malignant PET in our series. The findings may be explained by the expansive growth of malignant PETs, which leads to tumor demarcation and often to pseudoencapsulation, diminishing the accessibility of tumor cells to peritumoral lymphatics. In cervical carcinomas
      • Schoppmann SF
      • Birner P
      • Stockl J
      • Kalt R
      • Ullrich R
      • Caucig C
      • Kriehuber E
      • Nagy K
      • Alitalo K
      • Kerjaschki D
      Tumor-associated macrophages express lymphatic endothelial growth factors and are related to peritumoral lymphangiogenesis.
      high pLVD correlated with lymphatic tumor progression, while in head and neck carcinomas
      • Beasley NJ
      • Prevo R
      • Banerji S
      • Leek RD
      • Moore J
      • van Trappen P
      • Cox G
      • Harris AL
      • Jackson DG
      Intratumoral lymphangiogenesis and lymph node metastasis in head and neck cancer.
      and melanomas
      • Straume O
      • Jackson DG
      • Akslen LA
      Independent prognostic impact of lymphatic vessel density and presence of low-grade lymphangiogenesis in cutaneous melanoma.
      high pLVD turned out to be a favorable prognostic factor.
      Twenty percent of the PETs (21 of 104) showed moderate to strong VEGF-C expression. These immunohistochemical results were confirmed in 29 PETs by quantification of VEGF-C mRNA from formalin-fixed paraffin-embedded tumor tissue blocks using quantitative real-time RT-PCR. Although VEGF-C mRNA was detected in virtually all PETs using this highly sensitive assay, we failed to detect VEGF-D mRNA.
      PETs that exhibited moderate to strong VEGF-C expression showed a closely correlated glucagon expression of the same intensity, both in functioning (ie, glucagonomas) and nonfunctioning glucagon-expressing tumors. Even in PETs in which glucagon-expressing cells accounted for only a minor cell population of the tumor cells there was associated VEGF-C expression. In a multivariate analysis testing cause variables of VEGF-C expression, moderate to strong glucagon expression was the strongest indicator, followed by moderate to strong pancreatic polypeptide expression and weak glucagon expression. Moreover, immunofluorescent double labeling of glucagon and VEGF-C revealed cellular co-localization in 8 of 10 PETs. Because VEGF-C is also detected in the glucagon cells of the normal islets, the results in the PETs suggest that this so far unexplained relationship between VEGF-C and the normal glucagon cells is retained in neoplastic conditions. At present, there is no plausible explanation for the association between pancreatic polypeptide expression and VEGF-C expression, apart from the observation that glucagonomas frequently co-express pancreatic polypeptide.
      • Warner TF
      • Block M
      • Hafez GR
      • Mack E
      • Lloyd RV
      • Bloom SR
      Glucagonomas. Ultrastructure and immunocytochemistry.
      Although there is ample experimental evidence of an association between VEGF-C expression and lymphangiogenesis,
      • Makinen T
      • Veikkola T
      • Mustjoki S
      • Karpanen T
      • Catimel B
      • Nice EC
      • Wise L
      • Mercer A
      • Kowalski H
      • Kerjaschki D
      • Stacker SA
      • Achen MG
      • Alitalo K
      Isolated lymphatic endothelial cells transduce growth, survival and migratory signals via the VEGF-C/D receptor VEGFR-3.
      • Veikkola T
      • Jussila L
      • Makinen T
      • Karpanen T
      • Jeltsch M
      • Petrova TV
      • Kubo H
      • Thurston G
      • McDonald DM
      • Achen MG
      • Stacker SA
      • Alitalo K
      Signalling via vascular endothelial growth factor receptor-3 is sufficient for lymphangiogenesis in transgenic mice.
      • Enholm B
      • Karpanen T
      • Jeltsch M
      • Kubo H
      • Stenback F
      • Prevo R
      • Jackson DG
      • Yla-Herttuala S
      • Alitalo K
      Adenoviral expression of vascular endothelial growth factor-C induces lymphangiogenesis in the skin.
      • Skobe M
      • Hawighorst T
      • Jackson DG
      • Prevo R
      • Janes L
      • Velasco P
      • Riccardi L
      • Alitalo K
      • Claffey K
      • Detmar M
      Induction of tumor lymphangiogenesis by VEGF-C promotes breast cancer metastasis.
      • Mandriota SJ
      • Jussila L
      • Jeltsch M
      • Compagni A
      • Baetens D
      • Prevo R
      • Banerji S
      • Huarte J
      • Montesano R
      • Jackson DG
      • Orci L
      • Alitalo K
      • Christofori G
      • Pepper MS
      Vascular endothelial growth factor-C-mediated lymphangiogenesis promotes tumour metastasis.
      this relationship is less conspicuous in human tumors. A correlation between VEGF-C expression and high LVD was recently reported in a small series of oral squamous cell carcinomas
      • Sedivy R
      • Beck-Mannagetta J
      • Haverkampf C
      • Battistutti W
      • Honigschnabl S
      Expression of vascular endothelial growth factor-C correlates with the lymphatic microvessel density and the nodal status in oral squamous cell cancer.
      but has been shown to be absent in melanomas
      • Straume O
      • Jackson DG
      • Akslen LA
      Independent prognostic impact of lymphatic vessel density and presence of low-grade lymphangiogenesis in cutaneous melanoma.
      • Dadras SS
      • Paul T
      • Bertoncini J
      • Brown LF
      • Muzikansky A
      • Jackson DG
      • Ellwanger U
      • Garbe C
      • Mihm MC
      • Detmar M
      Tumor lymphangiogenesis: a novel prognostic indicator for cutaneous melanoma metastasis and survival.
      and in squamous cell carcinomas of the head and neck.
      • Beasley NJ
      • Prevo R
      • Banerji S
      • Leek RD
      • Moore J
      • van Trappen P
      • Cox G
      • Harris AL
      • Jackson DG
      Intratumoral lymphangiogenesis and lymph node metastasis in head and neck cancer.
      • Maula SM
      • Luukkaa M
      • Grenman R
      • Jackson D
      • Jalkanen S
      • Ristamaki R
      Intratumoral lymphatics are essential for the metastatic spread and prognosis in squamous cell carcinomas of the head and neck region.
      In cervical carcinoma peritumoral LVD was correlated with the density of VEGF-C-expressing peritumoral macrophages.
      • Schoppmann SF
      • Birner P
      • Stockl J
      • Kalt R
      • Ullrich R
      • Caucig C
      • Kriehuber E
      • Nagy K
      • Alitalo K
      • Kerjaschki D
      Tumor-associated macrophages express lymphatic endothelial growth factors and are related to peritumoral lymphangiogenesis.
      Our study revealed a clear association between VEGF-C expression and high intratumoral LVD. This finding demonstrates for the first time in a large tumor series that the production of a lymphangiogenic cytokine is coupled with intratumoral lymphangiogenesis. However, statistical analysis using a multivariate regression model showed that VEGF-C (together with glucagon) accounted for high iLVD in only approximately one-fourth of the tumors (r2 = 0.27). Because the expression of the other cytokines implicated in lymphangiogenesis, VEGF-D, bFGF, and VEGF, was not associated with iLVD, the presence of additional, possibly so far unknown, factors has to be assumed. Alternatively, posttranslational processing of known cytokines may also play a role, rendering angiogenic factors more lymphangiogenic.
      In summary, we demonstrated intratumoral lymphangiogenesis in PETs. This process is, at least in part, mediated by VEGF-C expression and seems to be independent of other lymphangiogenic factors such as VEGF-D, VEGF, and bFGF. The observations that high iLVD was associated with lymph vessel invasion and with angioinvasive/metastatic tumor features indicate that lymphangiogenesis may promote the malignant progression of PETs. Our data on a large series of human PETs confirm and extend the previous reports on RipTag/RipVEGF-C transgenic mice, in which VEGF-C-induced lymphangiogenesis rendered locally invasive islet cell tumors lymphangioinvasive and metastatic.
      • Mandriota SJ
      • Jussila L
      • Jeltsch M
      • Compagni A
      • Baetens D
      • Prevo R
      • Banerji S
      • Huarte J
      • Montesano R
      • Jackson DG
      • Orci L
      • Alitalo K
      • Christofori G
      • Pepper MS
      Vascular endothelial growth factor-C-mediated lymphangiogenesis promotes tumour metastasis.

      Acknowledgements

      We thank Maike Pacena, Anja Paulus, Birgit Rohlf, and Sonja Vollbehr for the excellent technical assistance; Kay Dege for editing the manuscript; and J. Hedderich for help with the statistics.

      References

        • Beasley NJ
        • Prevo R
        • Banerji S
        • Leek RD
        • Moore J
        • van Trappen P
        • Cox G
        • Harris AL
        • Jackson DG
        Intratumoral lymphangiogenesis and lymph node metastasis in head and neck cancer.
        Cancer Res. 2002; 62: 1315-1320
        • Maula SM
        • Luukkaa M
        • Grenman R
        • Jackson D
        • Jalkanen S
        • Ristamaki R
        Intratumoral lymphatics are essential for the metastatic spread and prognosis in squamous cell carcinomas of the head and neck region.
        Cancer Res. 2003; 63: 1920-1926
        • Straume O
        • Jackson DG
        • Akslen LA
        Independent prognostic impact of lymphatic vessel density and presence of low-grade lymphangiogenesis in cutaneous melanoma.
        Clin Cancer Res. 2003; 9: 250-256
        • Dadras SS
        • Paul T
        • Bertoncini J
        • Brown LF
        • Muzikansky A
        • Jackson DG
        • Ellwanger U
        • Garbe C
        • Mihm MC
        • Detmar M
        Tumor lymphangiogenesis: a novel prognostic indicator for cutaneous melanoma metastasis and survival.
        Am J Pathol. 2003; 162: 1951-1960
        • Birner P
        • Schindl M
        • Obermair A
        • Plank C
        • Breitenecker G
        • Kowalski H
        • Oberhuber G
        Lymphatic microvessel density in epithelial ovarian cancer: its impact on prognosis.
        Anticancer Res. 2000; 20: 2981-2985
        • Mouta Carreira C
        • Nasser SM
        • di Tomaso E
        • Padera TP
        • Boucher Y
        • Tomarev SI
        • Jain RK
        lYVE-1 is not restricted to the lymph vessels: expression in normal liver blood sinusoids and down-regulation in human liver cancer and cirrhosis.
        Cancer Res. 2001; 61: 8079-8084
        • Schoppmann SF
        • Birner P
        • Studer P
        • Breiteneder-Geleff S
        Lymphatic microvessel density and lymphovascular invasion assessed by anti-podoplanin immunostaining in human breast cancer.
        Anticancer Res. 2001; 21: 2351-2355
        • Williams CS
        • Leek RD
        • Robson AM
        • Banerji S
        • Prevo R
        • Harris AL
        • Jackson DG
        Absence of lymphangiogenesis and intratumoural lymph vessels in human metastatic breast cancer.
        J Pathol. 2003; 200: 195-206
        • Schoppmann SF
        • Birner P
        • Stockl J
        • Kalt R
        • Ullrich R
        • Caucig C
        • Kriehuber E
        • Nagy K
        • Alitalo K
        • Kerjaschki D
        Tumor-associated macrophages express lymphatic endothelial growth factors and are related to peritumoral lymphangiogenesis.
        Am J Pathol. 2002; 161: 947-956
        • Joukov V
        • Pajusola K
        • Kaipainen A
        • Chilov D
        • Lahtinen I
        • Kukk E
        • Saksela O
        • Kalkkinen N
        • Alitalo K
        A novel vascular endothelial growth factor, VEGF-C, is a ligand for the Flt4 (VEGFR-3) and KDR (VEGFR-2) receptor tyrosine kinases.
        EMBO J. 1996; 15: 290-298
        • Joukov V
        • Sorsa T
        • Kumar V
        • Jeltsch M
        • Claesson-Welsh L
        • Cao Y
        • Saksela O
        • Kalkkinen N
        • Alitalo K
        Proteolytic processing regulates receptor specificity and activity of VEGF-C.
        EMBO J. 1997; 16: 3898-3911
        • Achen MG
        • Jeltsch M
        • Kukk E
        • Makinen T
        • Vitali A
        • Wilks AF
        • Alitalo K
        • Stacker SA
        Vascular endothelial growth factor D (VEGF-D) is a ligand for the tyrosine kinases VEGF receptor 2 (Flk1) and VEGF receptor 3 (Flt4).
        Proc Natl Acad Sci USA. 1998; 95: 548-553
        • Stacker SA
        • Stenvers K
        • Caesar C
        • Vitali A
        • Domagala T
        • Nice E
        • Roufail S
        • Simpson RJ
        • Moritz R
        • Karpanen T
        • Alitalo K
        • Achen MG
        Biosynthesis of vascular endothelial growth factor-D involves proteolytic processing which generates non-covalent homodimers.
        J Biol Chem. 1999; 274: 32127-32136
        • Kukk E
        • Lymboussaki A
        • Taira S
        • Kaipainen A
        • Jeltsch M
        • Joukov V
        • Alitalo K
        VEGF-C receptor binding and pattern of expression with VEGFR-3 suggests a role in lymphatic vascular development.
        Development. 1996; 122: 3829-3837
        • Jeltsch M
        • Kaipainen A
        • Joukov V
        • Meng X
        • Lakso M
        • Rauvala H
        • Swartz M
        • Fukumura D
        • Jain RK
        • Alitalo K
        Hyperplasia of lymphatic vessels in VEGF-C transgenic mice.
        Science. 1997; 276: 1423-1425
        • Duff SE
        • Li C
        • Jeziorska M
        • Kumar S
        • Saunders MP
        • Sherlock D
        • O'Dwyer ST
        • Jayson GC
        Vascular endothelial growth factors C and D and lymphangiogenesis in gastrointestinal tract malignancy.
        Br J Cancer. 2003; 89: 426-430
        • Valtola R
        • Salven P
        • Heikkila P
        • Taipale J
        • Joensuu H
        • Rehn M
        • Pihlajaniemi T
        • Weich H
        • deWaal R
        • Alitalo K
        VEGFR-3 and its ligand VEGF-C are associated with angiogenesis in breast cancer.
        Am J Pathol. 1999; 154: 1381-1390
        • Yang W
        • Klos K
        • Yang Y
        • Smith TL
        • Shi D
        • Yu D
        ErbB2 overexpression correlates with increased expression of vascular endothelial growth factors A, C, and D in human breast carcinoma.
        Cancer. 2002; 94: 2855-2861
        • Nakamura Y
        • Yasuoka H
        • Tsujimoto M
        • Yang Q
        • Tsukiyama A
        • Imabun S
        • Nakahara M
        • Nakao K
        • Nakamura M
        • Mori I
        • Kakudo K
        Clinicopathological significance of vascular endothelial growth factor-C in breast carcinoma with long-term follow-up.
        Mod Pathol. 2003; 16: 309-314
        • White JD
        • Hewett PW
        • Kosuge D
        • McCulloch T
        • Enholm BC
        • Carmichael J
        • Murray JC
        Vascular endothelial growth factor-D expression is an independent prognostic marker for survival in colorectal carcinoma.
        Cancer Res. 2002; 62: 1669-1675
        • Hanrahan V
        • Currie MJ
        • Gunningham SP
        • Morrin HR
        • Scott PA
        • Robinson BA
        • Fox SB
        The angiogenic switch for vascular endothelial growth factor (VEGF)-A, VEGF-B, VEGF-C, and VEGF-D in the adenoma-carcinoma sequence during colorectal cancer progression.
        J Pathol. 2003; 200: 183-194
        • Funaki H
        • Nishimura G
        • Harada S
        • Ninomiya I
        • Terada I
        • Fushida S
        • Tani T
        • Fujimura T
        • Kayahara M
        • Shimizu K
        • Ohta T
        • Miwa K
        Expression of vascular endothelial growth factor D is associated with lymph node metastasis in human colorectal carcinoma.
        Oncology. 2003; 64: 416-422
        • Yokoyama Y
        • Charnock-Jones DS
        • Licence D
        • Yanaihara A
        • Hastings JM
        • Holland CM
        • Emoto M
        • Umemoto M
        • Sakamoto T
        • Sato S
        • Mizunuma H
        • Smith SK
        Vascular endothelial growth factor-D is an independent prognostic factor in epithelial ovarian carcinoma.
        Br J Cancer. 2003; 88: 237-244
        • Yokoyama Y
        • Charnock-Jones DS
        • Licence D
        • Yanaihara A
        • Hastings JM
        • Holland CM
        • Emoto M
        • Sakamoto A
        • Sakamoto T
        • Maruyama H
        • Sato S
        • Mizunuma H
        • Smith SK
        Expression of vascular endothelial growth factor (VEGF)-D and its receptor, VEGF receptor 3, as a prognostic factor in endometrial carcinoma.
        Clin Cancer Res. 2003; 9: 1361-1369
        • Nakamura Y
        • Yasuoka H
        • Tsujimoto M
        • Yang Q
        • Imabun S
        • Nakahara M
        • Nakao K
        • Nakamura M
        • Mori I
        • Kakudo K
        Prognostic significance of vascular endothelial growth factor D in breast carcinoma with long-term follow-up.
        Clin Cancer Res. 2003; 9: 716-721
        • Debinski W
        • Slagle-Webb B
        • Achen MG
        • Stacker SA
        • Tulchinsky E
        • Gillespie GY
        • Gibo DM
        VEGF-D is an X-linked/AP-1 regulated putative onco-angiogen in human glioblastoma multiforme.
        Mol Med. 2001; 7: 598-608
        • Achen MG
        • Williams RA
        • Minekus MP
        • Thornton GE
        • Stenvers K
        • Rogers PA
        • Lederman F
        • Roufail S
        • Stacker SA
        Localization of vascular endothelial growth factor-D in malignant melanoma suggests a role in tumour angiogenesis.
        J Pathol. 2001; 193: 147-154
        • Partanen TA
        • Arola J
        • Saaristo A
        • Jussila L
        • Ora A
        • Miettinen M
        • Stacker SA
        • Achen MG
        • Alitalo K
        VEGF-C and VEGF-D expression in neuroendocrine cells and their receptor, VEGFR-3, in fenestrated blood vessels in human tissues.
        FASEB J. 2000; 14: 2087-2096
        • Solcia E
        • Kloppel G
        • Sobin LH
        Histological Typing of Endocrine Tumours. Springer, Berlin2000: 56-60
        • Weidner N
        • Semple JP
        • Welch WR
        • Folkman J
        Tumor angiogenesis and metastasis—correlation in invasive breast carcinoma.
        N Engl J Med. 1991; 324: 1-8
        • Godfrey TE
        • Kim SH
        • Chavira M
        • Ruff DW
        • Warren RS
        • Gray JW
        • Jensen RH
        Quantitative mRNA expression analysis from formalin-fixed, paraffin-embedded tissues using 5′ nuclease quantitative reverse transcription-polymerase chain reaction.
        J Mol Diagn. 2000; 2: 84-91
        • Breiteneder-Geleff S
        • Soleiman A
        • Kowalski H
        • Horvat R
        • Amann G
        • Kriehuber E
        • Diem K
        • Weninger W
        • Tschachler E
        • Alitalo K
        • Kerjaschki D
        Angiosarcomas express mixed endothelial phenotypes of blood and lymphatic capillaries: podoplanin as a specific marker for lymphatic endothelium.
        Am J Pathol. 1999; 154: 385-394
        • Banerji S
        • Ni J
        • Wang SX
        • Clasper S
        • Su J
        • Tammi R
        • Jones M
        • Jackson DG
        LYVE-1, a new homologue of the CD44 glycoprotein, is a lymph-specific receptor for hyaluronan.
        J Cell Biol. 1999; 144: 789-801
        • Prevo R
        • Banerji S
        • Ferguson DJ
        • Clasper S
        • Jackson DG
        Mouse LYVE-1 is an endocytic receptor for hyaluronan in lymphatic endothelium.
        J Biol Chem. 2001; 276: 19420-19430
        • Warner TF
        • Block M
        • Hafez GR
        • Mack E
        • Lloyd RV
        • Bloom SR
        Glucagonomas. Ultrastructure and immunocytochemistry.
        Cancer. 1983; 51: 1091-1096
        • Makinen T
        • Veikkola T
        • Mustjoki S
        • Karpanen T
        • Catimel B
        • Nice EC
        • Wise L
        • Mercer A
        • Kowalski H
        • Kerjaschki D
        • Stacker SA
        • Achen MG
        • Alitalo K
        Isolated lymphatic endothelial cells transduce growth, survival and migratory signals via the VEGF-C/D receptor VEGFR-3.
        EMBO J. 2001; 20: 4762-4773
        • Veikkola T
        • Jussila L
        • Makinen T
        • Karpanen T
        • Jeltsch M
        • Petrova TV
        • Kubo H
        • Thurston G
        • McDonald DM
        • Achen MG
        • Stacker SA
        • Alitalo K
        Signalling via vascular endothelial growth factor receptor-3 is sufficient for lymphangiogenesis in transgenic mice.
        EMBO J. 2001; 20: 1223-1231
        • Enholm B
        • Karpanen T
        • Jeltsch M
        • Kubo H
        • Stenback F
        • Prevo R
        • Jackson DG
        • Yla-Herttuala S
        • Alitalo K
        Adenoviral expression of vascular endothelial growth factor-C induces lymphangiogenesis in the skin.
        Circ Res. 2001; 88: 623-629
        • Skobe M
        • Hawighorst T
        • Jackson DG
        • Prevo R
        • Janes L
        • Velasco P
        • Riccardi L
        • Alitalo K
        • Claffey K
        • Detmar M
        Induction of tumor lymphangiogenesis by VEGF-C promotes breast cancer metastasis.
        Nat Med. 2001; 7: 192-198
        • Mandriota SJ
        • Jussila L
        • Jeltsch M
        • Compagni A
        • Baetens D
        • Prevo R
        • Banerji S
        • Huarte J
        • Montesano R
        • Jackson DG
        • Orci L
        • Alitalo K
        • Christofori G
        • Pepper MS
        Vascular endothelial growth factor-C-mediated lymphangiogenesis promotes tumour metastasis.
        EMBO J. 2001; 20: 672-682
        • Sedivy R
        • Beck-Mannagetta J
        • Haverkampf C
        • Battistutti W
        • Honigschnabl S
        Expression of vascular endothelial growth factor-C correlates with the lymphatic microvessel density and the nodal status in oral squamous cell cancer.
        J Oral Pathol Med. 2003; 32: 455-460