Mucosal Lymphatic Vessels of the Esophagus Distant From

Okajimas Folia Anat. Jpn., 88(2): 43–47, August, 2011

Mucosal lymphatic vessels of the esophagus distant from the cancer margin: morphometrical analysis using 27 surgically removed specimens of squamous cell carcinoma located in the upper or middle thoracic esophagus

Hidefumi Nishimori1, Shogo Hayashi2*, Munekazu Naito3, Gen Murakami4, Masahiro Fujita5, Masao Hosokawa1

1Division of Surgery, Keiyu-kai Sapporo Hospital, Sapporo, Hokkaido, Japan 2Medical Education Center, Aichi Medical University, Nagakute, Aichi, Japan 3Department of Anatomy, Tokyo Medical University, Shinjuku, Tokyo, Japan 4Iwamizawa Kojin-kai Hospital, Iwamizawa, Hokkaido, Japan 5Keiyu-kai Institute of Clinical and Surgical Pathology, Sapporo, Hokkaido, Japan

– Received for Publication, March 7, 2011 –

Key Words: lymphatic vessel, esophageal cancer, morphometric study, immnohistochemistry

Abstract: Purpose: To clarify the configuration of the esophageal mucosal lymphatics distant from cancer using D2-40 immunohistochemistry. Methods: D2-40 immunohistochemistry for human lymphatic epithelium was performed at sites about 10 cm anal from the pathologically examined margin of upper or mid- thoracic squamous cell carcinoma (27 patients). We measured the entire length of mucosal lymphatic vessels within a ×10 objective field (1.2 mm along the muscularis mucosae). results: The present morphometrical study demonstrated significant individual differences in the amount of mucosal lymphatic vessels, within a range of more than 10-fold (8.4 mm– 0.8 mm within an objective field). However, the difference in length of the mucosal lymphatic epithelium did not correlate with either N-factor, T-factor including cancer depth or prognosis. Conclusion: A higher density of pre-existing mucosal lymphatic vessels may not always be correlated with larger numbers of nodal metastases. Lymphatic proliferation or dilation induced by cancer seems to occur irrespective of whether pre-existing vessels are rich or sparse.

Introduction lymphatic vessel morphology are present in the esophagus distant from the cancer, although systemic changes in Lymphatic vessel invasion has been shown to be a lymphatics other than those of the esophagus have been strong and independent prognostic factor in cases of suggested in cancer patients.8, 9) esophageal squamous cell cancer.1,2,3) In contrast to the Recently, using donated cadavers without cancer, Yajin identification of lymphatic cancer invasion using conven- et al.10) demonstrated individual differences of more than tional staining, recently developed antibodies such as D2- 2-fold in the “length or number of the mucosal lymphatics 40 (anti-human podoplanin antibody) have allowed more per 1-mm length of the esophageal epithelium”. This led acurate identification of gastrointestinal mucosal lym them to suggest that, before the onset of cancer pathology, phatic vessels including those of the esophagus.4) In this the “pre-existing” mucosal lymphatic vessels exhibit sig- context, many research groups have focused on individ nificant individual differences. On the basis of their studies, ual differences in lymphatic vessel density among pa- we hypothesized that, even in cancer patients, the mor- tients with esophageal squamous cell carcinoma.5,6,7) These phology of lymphatics distant from the cancer site is likely studies have revealed that lymphatic density is significantly to reflect any “pre-existing” morphology. How much of a correlated with nodal involvement, and thus possibly with distance from the cancer site is necessary? In the present prognosis. However, to our knowledge, no previous study study, we hypothesized that a distance of 10 cm anal from has investigated whether certain individual differences in the pathologically estimated cancer margin would be free

* Correspondence to Shogo Hayashi, MD, PhD, Medical Education Center, Aichi University School of Medicine, Nagakute, Aichi 480-1195 Japan. Phone: +81 561 62 3311 ext. 1447. Fax: +81 561 63 1037. E-mail: [email protected] 44 H. Nishimori et al. of cancer invasion because of the dominant upward longi- “upper or mid” thoracic esophagus. The sites examined by tudinal flow along the esophagus.10,11,12,13) Consequently, to D2-40 immunohistochemistry were located 10 cm anal to assess the anatomical background of cancer prognosis, we the pathologically examined cancer margin. This distance conducted the present morphometrical study to analyze was easily estimated because, in the hospital that had mucosal lymphatic vessels in the esophagus around 10 cm treated the present patients, pathological sections had been anal from squamous cell carcinoma occupying the upper routinely prepared along the “longitudinal axis” of the or middle thoracic esophagus. esophagus (Fig. 1A). Conversely, for the present aim, cross-sectional histology was considered better because, in previous studies, 5,6,7,10) the morphometry had been Materials and Methods performed on cross-sections. Therefore, the longitudinal paraffin blocks were melted and the specimens were re- D2-40 immunohistochemistry for human lymphatic embedded to allow transverse sections (6 μm thick) to be epithelium was performed on surgically removed speci- cut. After taking photos using a ×10 objective (at 3 or 4 mens of the “mid or lower” thoracic esophagus from 27 sites along the entire circumferential wall of the esopha- patients with squamous cell carcinoma occupying the gus at the 10 cm anal level; e.g., Fig. 1B), we measured

Figure 1. Materials and methods of the present study. This is an example with a cancer in the mid-thoracic esophagus: the oral part of the removed esophagus is omitted in this figure. Panel A displays the measured site almost 10 cm anal to the cancer (star. In the hospital in which the surgery was performed, the removed esophagus was cut longitudinally, as indicated by dotted lines, to show the longitudinal extension of the cancer. From the long belt-like paraffin embedded specimens, we prepared cross section histology with D2-40 immunohistochemistry ( panel B) and traced the vessels ( panel C). Lymph and carcinoma of esophagus 45 the entire length of mucosal lymphatic vessels within an Using cross-section histology with D2-40 immuno objective field (1.2 mm along the muscularis mucosae). histochemistry, the fourth author (G.M.) took photos at On the basis of 164 pathological reports (April 2003– hot spots of lymphatic vessel density (Weidner et al., September 2004) of curative surgeries of esophageal squa- 1991) in each of 3 sections at the 10-cm anal level mous cell carcinoma at Keiyu-kai Sapporo Hospital, we (Fig. 1B). After tracing of the D2-40-positive vessels chose patients with upper or mid-thoracic squamous carci- (Fig. 1C), the third author (M.N.) measured them using noma from whom the esophagus had been surgically the ImageJ program (developed at the U.S. National Insti- removed more than 10 cm anal to the cancer margin. We tutes of Health and available on the internet at http://rsb. tried to eliminate any bias related to T-factor, N-factor and info.nih.gov/ij/ ). The third author also calculated 1) the prognosis. The T and N factors were classified according total number of mucosal lymphatic vessels in an objective to the criteria of the Japanese Society for Esophageal Dis- field (1.2 mm along the muscularis mucosae) and 2) total eases.14) The population examined (Table 1) comprised 23 circumferential length of vessels in an objective field. The males and 4 females ranging in age from 79 to 51 years second author (S.H.) evaluated the measured data using (70 –79 years, 14 patients; 60 – 69, 10; 51–59, 3; other fac- t-analysis, chi-squared analysis and one-way analysis of tors, see footnote of Table 2). None of the 27 patients had variance. The present study was conducted in accordance Barrett’s esophagus. with the provisions of the Declaration of Helsinki, 1995 For D2-40 immunohistochemistry, the primary anti- (as revised in Edinburgh in 2000). The protocol was ap- body (monoclonal anti-human podoplanin; Nichirei D2- proved by the ethics committees of the hospital of the first, 40 (code No. 413451), Tokyo, Japan; 1 : 100 dilution) was fifth and sixth authors (H.N., M.F., M.H.) who performed used after immersion in a ligand activator (Histofine SAB- the surgery or pathological examination. PO Kit, code No. 415211, Nichirei, Tokyo, Japan) with autoclave treatment (105 °C, 10 min). The second antibody (Dako Chem Mate Envision Kit, Dako, Glostrup, Results Denmark) was labeled with horseradish peroxidase, and antigen-antibody reactions were detected using HRP- In all 27 specimens examined, D2-40 successfully catalysis with diaminobenzidine. Counterstaining with stained vessels in any layer of the esophagus. However, hematoxylin was performed on the same samples. the measured lymphatic structures were limited to D2-40- positive vessels on the mucosal side of the muscularis mucosae (Fig. 1B). There was a difference in the density and amount of vessels among 3 or 4 sections at a level almost Table 1. Indivisual patient data analyzed in this study 10 cm anal from the cancer margin. However, the differ- age sex LV length LV number length/number T N outcome ence was restricted to within 80% of the mean in the spec- imen. In this study, we averaged the data obtained from 60 F 8.4 5.7 1.48 3 0 dead the 3– 4 quadrants of the circumferential esophageal wall 77 M 4.7 8.3 0.56 1 2 alive irrespective of the quadrants occupied by the cancer. We 75 M 4.1 17.0 0.24 1 2 dead 72 M 3.2 14.0 0.23 2 0 dead did not measure lymphatic vessels in and around the can- 57 M 3.1 17.0 0.19 3 2 dead cer because cross-sections were not obtained from this 73 M 2.9 10.7 0.27 2 0 dead area. 72 M 2.9 9.7 0.30 3 2 alive The amount of mucosal lymphatic vessels (the total cir- 60 M 2.8 17.3 0.16 2 2 alive cumferential length of vessels in an objective field) ranged 65 M 2.8 22.8 0.12 3 4 dead 71 F 2.5 15.3 0.17 2 1 alive from 8.4 mm to 0.8 mm (mean, 1.22 mm), while the nu- 57 M 2.5 10.3 0.24 3 4 dead merical density (the total number of mucosal lymphatic 73 M 2.3 10.3 0.22 1 1 alive vessels in an objective field) ranged from 22.8 to 5.7 73 M 2.3 14.0 0.16 2 0 alive (mean, 6.50). Thus, the circumferential length of a hypo- 64 M 2.2 11.0 0.20 2 1 alive thetical single vessel on average ranged from 1.48 to 70 F 2.1 16.3 0.13 2 0 alive 67 M 2.1 16.0 0.13 3 2 alive 0.06 mm (mean, 0.27 mm), although the vessel was likely 75 M 1.9 9.3 0.20 1 2 alive to be continuous with another. The circumferential length 63 M 1.9 7.0 0.27 1 0 alive of a single vessel (length/number in Table 1) depended on 68 M 1.8 7.0 0.26 2 0 alive whether the vessel was dilated or collapsed. In 18 of the 79 M 1.8 8.0 0.22 1 2 dead 27 patients, the length of mucosal vessels was estimated to 51 F 1.6 8.3 0.19 3 2 alive 70 M 1.6 11.5 0.14 2 2 dead be between 2.9 and 1.6 mm. 76 M 1.6 7.3 0.21 3 0 dead Correlations of the amount of mucosal lymphatic ves- 76 M 1.3 21.7 0.06 3 2 alive sels with other factors are summarized in Table 2. Despite 66 F 1.2 6.5 0.19 3 2 dead a difference of more than 10-fold in the total circumferen- 64 M 1.0 8.0 0.13 3 1 alive tial length of mucosal vessels, none of the factors exam- 62 M 0.8 7.0 0.12 1 0 alive ined was significantly correlated with either the amount of 46 H. Nishimori et al.

Table 2. Summary of amount of mucosal lymphatics at an objective field (mean) male, 2.34 mm female 1.88 mm (no significant; p = 0.35) dead patients, 2.10 mm alive patients, 2.47 mm (no significant; p = 0.33) sm cancer, 2.17 mm non-sm cancer, 2.60 mm (no significant; p = 0.32) T-1, 2.50 mm T-2, 2.38 mm T-3, 2.01 mm (no significant; p = 0.51) N-0, 2.06 mm N-1, 2.02 mm N-2, 2.43 mm N-4, 2.62 mm (no significant; p = 0.72)

Numbers examined: males (n = 23), females (n = 4); dead patients (n = 15), alive patients (n = 12); submucosal or sm cancer (n = 7), non-sm cancer (n = 20); T-1 (n = 8), T-2 (n = 9), T-3 (n = 10); N-0 (n = 9), N-1 (n = 4), N-2 (n = 12), N-3 (n = 0), N-4 (n = 2).

vessels (LV length in Table 1), the numerical density (LV during fixation. Therefore, the present measured site, even number) and single vessel size (length/number) (Table 2). at 10 cm anal from the cancer margin, seems to display a Conversely, the distribution of each factor was so random hypertrophic morphology under induction by the cancer. (i.e., one of the lowest correlations) that the present sam- Conversely, the present morphology seems be different pling method appeared to have been ideal. from the pre-existing or normal lymphatic morphology. Yajin et al. (2009) estimated the largest individual difference in the “length of mucosal vessels per 1-mm length Discussion of the epithelium” as 100% or 2-fold. In fact, the range for the total length of vessels in an objective field in 20 of the Most of the previous data for esophageal mucosal lym- present 27 patients was 1.6 –3.2 mm (i.e., a 100% differ- phatic vessels have been obtained in and around areas of ence). Therefore, in these 20 patients, the present data cancer using the hot spot method,15) except for the study seem to correspond to the “normal range” of individual by Yajin et al.10) in which various counting methods were variations. However, in the present study, the maximum employed. In cancer patients, the numerical density of and minimum amounts or lengths of the mucosal vessels vessels has ranged from 9.1 to 33.0 when corrected to were quite different: 8.4 vs. 0.8. Thus, despite the degree values within a ×400 objective field. This objective field is of agreement for 20 of the 27 patients (see above), the much smaller (1/16th) than that used in the present study values for the other patients were likely to have been out- (×100).1,5,6) However, the present numbers of lymphatic side the normal range. More strictly, in the other 7 patients vessels were almost identical with the data obtained using with the greatest amounts of vessels (Table 1), the lym- a ×400 objective field. Thus, at a site 10 cm anal to a can- phatic morphology 10 cm anal from the cancer was likely cer, the numerical density may be much smaller than that to be strongly influenced by the cancer itself or related in the cancer site although, to our regret, we have no data pathology such as pneumonia or other types of inflamma- at and around the cancer. Actually, esophageal squamous tion. In fact, most of these 7 patients died within 5 years cancer is often negative for VEGF-C mRNA.16,17,18) The after surgery, despite the fact that some of them had no data by Yajin et al.10) is located in and around the mean of lymph node involvement. the present numerical density. Thus, even at a distant site, Consequently, a higher density or amount of pre- the cancer seems to provide a great individual difference existing mucosal lymphatic vessels may not always be in the density. correlated with larger numbers of nodal metastases. How- Using specimens without cancer, Yajin et al.10) mea- ever, even at a site distant from a cancer, lymphatic vessel sured the total circumferential length of mucosal vessels proliferation or dilation induced by cancer seems to occur seen in a cross-section (not in an objective field). They irrespective of whether pre-existing vessels are rich or also determined the “length of mucosal vessels per 1-mm sparse. length of the epithelium”: 0.64 –1.11 mm (mean, 0.85 mm) at the mid-thoracic level, although they also reported values for other upper parts of the esophagus. Because of References folding of the mucosa, the values obtained by Yajin et al.10) is likely to be much greater than the length of mucosal 1) Brücher BL, Stein HJ, Werner M, Siewert JR. Lymphatic vessel vessels per 1-mm length of the “muscularis mucosa”. Be- invasion is an independent prognostic factor in patients with a pri- cause of the present field along “1.2 mm” of the muscula- mary resected tumor with esophageal squamous cell carcinoma. ris mucosae, to allow comparison with the data obtained Cancer 2001; 92:2228–2233. by Yajin et al.,10) we need a 10/12 of the present data (e.g., 2) Watanabe M, Kuwano H, Araki K, Kawaguchi H, Saeki H, Kita- 2.34 × 10/12 mm in males = 1.95 mm). Notably, it corre- mura K, et al. Prognostic factors in patients with submucosal carci- 10) noma of the oesophagus. Br J Cancer 2000; 83:609– 613. sponds to more than 2-fold of Yajin et al. In this context, 3) Torres CM, Wang HH, Turner JR, Richards W, Sugarbaker D, the effect of mucosal folding is unlikely in the present data Shahsafaei A, et al. Pathologic prognostic factors in oesophageal because the specimens were fully extended with pins when squamous cell carcinoma: a follow-up study of 74 patients with or Lymph and carcinoma of esophagus 47

without preoperative chemoradiation therapy. Mod Pathol 1999; 11) Rice TW, Zuccaro G Jr, Adelstein DJ, Rybicki LA, Blackstone EH, 12:961–968. Goldblum JR. Esophageal carcinoma: depth of tumor invasion is 4) Yonemura Y, Endou Y, Tabachi K, Kawamura T, Yun HY, Kameya predictive of regional lymph node status. Ann Thorac Surg 1998; T, et al. Evaluation of lymphatic invasion in primary gastric cancer 65:787–792. by a new antibody, D2-40. Human Pathol 2006; 37:1193–1199. 12) Kuge K, Murakami G, Mizobuchi S, Hata Y, Aikou T, Sasaguri S. 5) Mori D, Yamasaki F, Shibata M, Tokunaga O. Lateral peritumoral Submucosal territory of the direct lymphatic drainage system to the lymphatic vessel invasion can predict lymph node metastasis in thoracic duct in the human esophagus. J Thoracic Cardiovasc Surg esophageal squamous cell carcinoma. Mod Pathol 2007; 20:694 – 2003; 125:1343–1349. 700. 13) Mizutani M, Murakami G, Nawata S, Hirai I, Kimura W. Anatomy 6) Nakayama Y, Matsumoto K, Nagato M, Inoue Y, Katsuki T, Mina- of right recurrent nerve node: why does early metastasis of esopha- gawa N, et al. Significance of lymphangiogenesis as assessed by geal cancer occurs in it? Surg Radiol Anat 2006; 28:333–338. immunohistochemistry for podoplanin in patients with esophageal 14) Guide Lines for the Clinical and Pathological Studies on Carcino- carcinoma. Anticancer Res 2007; 27:619– 625. ma of the Esophagus. 1999. Japanese Society for Esophageal Dis- 7) Brundler MA, Harrison JA, deSaussure B, dePerrot M, Pepper MS. eases, Kanehara publication, Tokyo. Lymphatic vessel density in the neoplastic progression of Barrett’s 15) Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis oesophagus to adenocarcinoma. J Clin Pathol 2006; 59:191–195. and metastasis-cprrelation in invasive breast carcinoma. N Eng J 8) Ran S, Volk L, Hall K, Flister MJ. Lymphangiogenesis and lym- Med 1991; 324:1–8. phatic metastasis in breast cancer. Pathophysiology 2009; 17:229– 16) Kimura Y, Watanabe M, Ohga T, Saeki H, Kakeji Y, Baba H, et al. 251. Vascular endothelial growth factor C expression correlates with 9) Dunnwald LK, Gralow JR, Ellis GK, Livingston RB, Linden HM, lymphatic involvement and poor prognosis in patients with esopha- Specht JM, et al. Tumor metabolism and blood flow changes by geal squamous cell carcinoma. Oncol Rep 2003; 10:1747–1751. positron emission tomography: relation to survival in patients 17) Ding MX, Lin XQ, Fu XY, Zhang N, Li JC. Expression of vascular treated with neoadjuvant chemotherapy for locally advanced breast endothelial growth factor-C and angiogenesis in esophageal squa- cancer. J Clin Oncol 2008; 26:4449– 4457. mous cell carcinoma. World J Gastroenterol 2006; 12:4582– 4585. 10) Yajin S, Murakami G, Takeuchi H, Hasegawa T, Kitano H. A nor- 18) Matsumoto M, Natsugoe S, Okumura H, Arima H, Yanagita S, mal configuration and interindividual differences in intramural Uchikado Y, et al. Overexpression of vascular endothelial growth lymphatics of the esophagus. J Thorac Cardiovasc Surg 2009; factor-C correlates with lymph node micrometastasis in submucosal 137:1406 –1414. esophageal cancer. J Gastrointest Surg 2006; 10:1016 –1022.