SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

We recently demonstrated that overexpression of the nucleolar GTP-binding protein nucleostemin drives the fraction of genetically-defined tumor cells that exhibit markers and tumorigenic properties of tumor initiating cells. More specifically, cells that constitutively express elevated levels of nucleostemin exhibit increased TWIST expression; expression of genes that induced pluripotent stem cells; enhanced radioresistance; tumor formation, even when small numbers of cells are implanted; and an increased propensity to metastasize. An immunohistochemical analysis of cancer stem cell markers, such as nucleostemin and TWIST has not been conducted in surgically-resected esophageal squamous cell carcinomas after neoadjuvant chemotherapy. In the present study, we examined the expression of CD133, CD44, nucleostemin, guanine nucleotide-binding protein-like 3-like, and TWIST by immunohistochemistry in a series of 54 surgically-resected specimens of esophageal squamous cell carcinomas after neoadjuvant chemotherapy. We identified that high nucleostemin proportion, TWIST intensity, and advanced pathological N stage were significantly correlated with poor relapse-free survival. Together, these observations imply nucleostemin and TWIST as the predictive markers for postoperative recurrence. (Cancer Sci 2012; 103: 233–238)

The incidence of esophageal cancer ranks eighth among all cancers worldwide, and its frequency is increasing.(1) Although esophageal cancers in Asia predominantly exhibit a squamous histology, both squamous and adenocarcinomas are found among European and US patients.(2) The standard therapy for resectable esophageal squamous cell carcinoma is neoadjuvant chemotherapy plus surgery or neoadjuvant chemoradiotherapy plus surgery.(3,4) However, the survival rate is very low, even in neoadjuvant groups (e.g. 43% of the 2-year survival rate in a previous report).(4)

Surgical specimens should provide the most reliable assessment of residual tumors. Possible prediction of postoperative recurrence and the selection of postoperative adjuvant therapies based on the prediction can lead to improved prognosis in all eligible patients. That is, tailor-made therapies would be realized based on predictors of recurrence in surgical specimens.

Tumor-initiating cells (TIC) or cancer stem cells are characterized by the capacity for unlimited self-renewal, are highly tumorigenic,(5) show resistance to chemotherapy and/or radiotherapy,(6–8) and might contribute to metastasis.(9–12) Although some cancer stem cells express particular cell surface markers, such as CD133 and CD44, the expression of such markers is not exclusive to TIC, and no universal cancer stem cell markers have been identified. The nucleolar GTP-binding protein nucleostemin is expressed at high levels in embryonic stem cells,(13,14) and nucleostemin has been proposed as a marker for TIC in highly-aggressive brain tumors.(15) Recently, we demonstrated that the expression of nucleostemin increases the fraction of tumorigenic human cells that exhibit TIC properties, and overexpression of nucleostemin increased TWIST expression, enhanced radioresistance, and exhibited an increased propensity to metastasize.(16) Moreover, TWIST has been reported to have the main role in epithelial mesenchymal transition (EMT),(17) and accumulating evidence indicates that TIC phenotype and EMT are largely associated.(11,16)

In the present study, we examined surgical specimens of esophageal squamous cell carcinoma after neoadjuvant chemotherapy for the expression of markers suggested to be potential cancer stem cell markers. In addition to clinicopathological factors, we also examined factors that play an important molecular role in metastasis to determine whether these markers could be a predictive marker of recurrence.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

Study population.  This clinical–biological correlative study was performed retrospectively in a consecutive series of 54 patients who underwent neoadjuvant chemotherapy of 5-fluorouracil (5-FU; 800 mg/m2, i.v., days 1–5) and cisplatin (CDDP; 80 mg/m2, i.v., day 1), followed by surgery for clinical stage II/III (Union Internationale Contre le Cancer TNM classification, 6th edition) primary squamous cell carcinoma of the esophagus at National Cancer Center Hospital between April 2002 and August 2008.(18) Pathological response to neoadjuvant chemotherapy was evaluated based on the extent of necrosis or disappearance of tumor cells in the total amount of tumor cells estimated in this study, according to Guidelines for Clinical and Pathological Studies on Carcinoma of the Esophagus (10th edition) by The Japan Esophageal Society.(19) Four major categories, Grades 0–3, and two minor ones for Grade 1, are provided as follows: Grade 0, no change or neither necrosis nor cellular or structural change can be seen throughout the lesion; Grade 1a, necrosis or disappearance of the tumor is present in less than one-third of the whole lesion, or only cellular or structural changes are visible in variable amounts; Grade 1b, necrosis or disappearance of the tumor is present in no more than two-thirds of the whole lesion; Grade 2, necrosis or disappearance of the tumor is present in more than two-thirds of the whole lesion, but viable tumor cells still remain; and Grade 3, entire lesion shows necrosis and/or is replaced by fibrosis, with or without granulomatous changes, and no viable tumor cells are observed. Recurrence-free survival (RFS) was defined as the period from the starting date of neoadjuvant chemotherapy to the date on which recurrence after surgery was confirmed. All patients provided informed consent prior to the collection and analysis of tissue samples. Clinical and pathological information on all patients was obtained from their medical records. This study was approved by the Institutional Review Board of the National Cancer Center Hospital, Tokyo, Japan.

Immunohistochemistry.  The surgically-resected specimens and biopsy specimens before neoadjuvant chemotherapy were fixed in 10% formalin. The entire primary esophageal tumor in each surgically-resected specimen was cut into 5-mm sections, and representative sections and biopsy specimens were embedded in paraffin for histological and immunohistochemical examination. The sections were cut into 3-μm slices, and hematoxylin–eosin staining was used for the histological evaluation. The most representative slices from each patient were selected for the immunohistochemical staining of CD133, CD44, nucleostemin, guanine nucleotide-binding protein-like 3-like (GNL3L), and TWIST, deparaffinized in xylene, dehydrated in a graded methanol series, and immersed in a 0.3% hydrogen peroxide solution in methanol for 30 min to inhibit endogenous peroxidase activity. After washing with PBS, the slides were placed in citrate buffer at pH 6.0. For antigen retrieval, the slides for CD133 and nucleostemin were heated at 121°C for 10 min in an autoclave. The slides for CD44, GNL3L, and TWIST were heated at 95°C for 40 min in a hot bath. The slides were allowed to cool for 30 min at room temperature, and were then washed three times with PBS. Each slide was then incubated overnight at 4°C in blocking buffer with primary antibodies against CD133 (130-092-395; Miltenyi Biotec, Auburn, CA, USA) at a dilution ratio of 1:100, against CD44 variant 3-10 (BBA10; R&D Systems, Minneapolis, MN, USA) at a dilution ratio of 1:1000, against nucleostemin (A300-600A; Bethyl Laboratories, Montgomery, TX, USA) at a dilution ratio of 1:250, against GNL3L (we grew this antibody in our laboratory) at a dilution ratio of 1:600, and against TWIST (sc-15393; Santa Cruz Biotechnology, Santa Cruz, CA, USA) at a dilution ratio of 1:100. The slides were washed three times with PBS, incubated for 30 min at room temperature using the EnVision system (Dako A/S, Glostrup, Denmark), and then developed with diaminobenzidine (Dako, Glostrup, Denmark). After three washes with diluted water, all slides were counterstained with Mayer’s hematoxylin solution.

Evaluation of immunohistochemistry.  Immunostaining was evaluated on whole standard tissue sections of primary esophageal tumors in 54 surgically-resected specimens and 51 biopsy specimens. Ten medium-power fields per section were screened for all sides. Tumors in the surgically-resected specimens were given a quantitative score of 1–100% as an average of the 10 fields screened by two independent clinical doctors who had no prior knowledge of the prognosis or other clinicopathological variables. Biopsy specimens were determined as positive when more than a single stained cell was observed, and were determined as negative when none of the cells were stained.

Positivity for CD133 and CD44 was defined as membranous staining. Positivity for nucleostemin was defined as nucleolus staining. Positivity for GNL3L was defined as cytoplasmic staining. Positivity for TWIST was defined as cytoplasmic staining, and its intensity was evaluated at the following levels: 1, faint; 2, moderate; and 3, strong, in agreement between the two doctors.

Statistical analysis.  Immunohistochemical expression and clinicopathological features were evaluated using the Wilcoxon test and Fisher’s exact test, as appropriate. Conditional logistic regression models were used for estimating the odds ratio and its 95% confidence interval (CI) to evaluate the association of each variable with the response to neoadjuvant chemotherapy. The Cox proportional hazard model was used in the multivariate analysis. Differences with a P-value ≤ 0.05 were considered significant. All statistical analyses were carried out using SAS software (version 9.1.3; Cary, NC, USA).

Cell culture.  The human esophageal cancer cell lines KYSE 180 and KYSE270 were maintained in RPMI-1640 and Ham’s F12 medium (1/1) supplemented with 2% heat-inactivated FBS. The human esophageal cancer cell line TE-8 was maintained in RPMI-1640 supplemented with 10% heat-inactivated FBS.

Immunoblotting.  Cells were lysed in a RIPA buffer containing 1% NP-40, 1 mM EDTA, 50 mM Tris–HCl (pH 7.4), and 150 mM NaCl. After sonication, lysates were cleared of insoluble material by centrifugation at 21 000 g at 4°C for 15 min. Proteins (100 or 200 μg) were subjected to SDS-PAGE in 12% polyacrylamide gels and immunoblotting. The following antibodies were used: anti-β-actin (AC-15; Sigma-Aldrich, St Louis, MO, USA), anti-nucleostemin (A300-600A; Bethyl Laboratories, Montgomery, TX, USA), and anti-human TWIST (Bio Matrix Research, Chiba, Japan).

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

Clinicopathological findings and immunohistochemical features.  The clinicopathological findings and immunohistochemical features of the 54 patients are listed in Table 1. Fifty-two cases had two courses of 5-FU + CDDP as neoadjuvant chemotherapy, and two cases had only one course of the neoadjuvant chemotherapy due to the progression of the disease after the first course. CD133 expression was detected in 44 of the 54 surgically-resected tumors (81.5%) (Fig. 1A). CD44, nucleostemin, GNL3L, and TWIST were detected in all 54 surgically-resected tumors (Fig. 1B-E). The proportion of CD44-positive cells was 100% in all tumors, and the proportion of TWIST-positive cells was 100% in 52 tumors. TWIST intensity was homogeneous in all tumors, and was evaluated as strong in three tumors, moderate in 17 tumors, and faint in 34 tumors. Detailed morphological study of the surface area and invasive area revealed an equal distribution of positive cells with CD133, CD44, nucleostemin, GNL3L, and TWIST expression throughout the tumors (data not shown). Each expression of CD133, CD44, nucleostemin, GNL3L, and TWIST was not related to histological type, and these expressions were not related to each other (Table 2,3).

Table 1.   Clinicopathological findings
Variablesn%
  1. †Union Internationale Contre le Cancer TNM classification.

  2. ‡Japanese Classification of Esophageal Carcinoma. cM, clinical metastasis; cN, clinical nodes; cT, clinical tumor; GNL3L, guanine nucleotide-binding protein-like 3-like; M/D, moderately differentiated; P/D, poorly differentiated; pM, pathological metastasis; pN, pathological nodes; pT, pathological tumor; SD, standard deviation; Sq, squamous cell carcinoma; W/D, well differentiated.

Sex
 Male4787.0
 Female713.0
Age (years)
 Median62 
 Range36–74 
Performance status
 03361.1
 12138.9
Histological type
 Sq (P/D)2037.0
 Sq (M/D+W/D)3463.0
cT†
 247.4
 35092.6
cN†
 0814.8
 14685.2
cM†
 054100.0
cStage†
 IIA713.0
 IIB59.3
 III4277.8
Pathological response to chemotherapy
 01731.5
 1a2240.7
 1b916.7
 2611.1
pT‡
 1611.1
 21018.5
 33564.8
 435.6
pN‡
 01222.2
 1713.0
 22240.7
 3713.0
 4611.1
pM‡
 054100.0
pStage‡
 I23.7
 II1120.4
 III3361.1
 IVA814.8
 Mean (%)SD
Immunohistochemistry in surgically-resected specimens
 CD13333.835.0
 CD44100.00.0
 Nucleostemin95.111.0
 GNL3L91.622.3
 TWIST99.24.5
TWIST intensity
 13463.0
 21731.5
 335.6
image

Figure 1.  Immunohistochemistry in surgically-resected specimens. (A) CD133; (B) CD44; (C) nucleostemin; (D) guanine nucleotide-binding protein-like 3-like; (E) TWIST expression with faint intensity (i), TWIST expression with moderate intensity (ii), and TWIST expression with strong intensity (iii) (×400).

Download figure to PowerPoint

Table 2.   Immunohistochemical findings in surgically-resected specimens: correlation between immunohistochemistry and histological type
VariablesMean (%), SDP-value*
Sq (P/D)Sq (M/D + W/D)
  1. *Wilcoxon test was performed.

  2. **Fisher’s exact test was performed.

  3. GNL3L, guanine nucleotide-binding protein-like 3-like; M/D, moderately differentiated; P/D, poorly differentiated; SD, standard deviation; Sq, squamous cell carcinoma; W/D, well differentiated.

 CD13327.333.937.735.60.11
 CD44100.00.0100.00.01.00
 Nucleostemin91.515.597.26.60.33
 GNL3L89.725.692.720.50.89
 TWIST100.00.098.75.70.29
 Count, %P-value**
Sq (P/D)Sq (M/D + W/D)
TWIST intensity
 11338.22161.80.53
 2529.41270.6
 3266.7133.3
Table 3.   Immunohistochemical findings in surgically-resected specimens: sample correlation matrix (Spearman correlation coefficients)
VariablesCD133CD44NucleosteminGNL3LTWIST
  1. GNL3L, guanine nucleotide-binding protein-like 3-like; M/D, moderately differentiated; P/D, poorly differentiated; SD, standard deviation; Sq, squamous cell carcinoma; W/D, well differentiated.

CD133 1.00 0.00−0.03−0.12
CD44
Nucleostemin 0.00 1.00−0.08−0.10
GNL3L−0.03−0.08 1.00−0.08
TWIST−0.12−0.10−0.08 1.00

Biopsy specimens of 51 of the 54 patients were evaluated. Thirty-eight biopsy specimens of the 51 patients were >0% positive for CD133 (74.5%), and 41 surgically-resected tumors from the 51 patients were >0% positive for CD133 (80.4%). Six patients whose biopsy specimens were evaluated as negative for CD133 were re-evaluated as positive after the evaluation of surgically-resected tumors, and three patients whose biopsy specimens were evaluated as positive for CD133 were re-evaluated as negative in the surgically-resected tumors. CD44, nucleostemin, GNL3L, and TWIST were positive in all biopsy specimens, while the staining intensity differed in each specimen. TWIST intensity was not evaluated in biopsy specimens because of the small tissue sample size.

Immunohistochemical features and clinical outcomes.  The correlation between the pathological response to neoadjuvant chemotherapy and the expression of each molecular markers in the surgically-resected specimens are shown in Table 4. In the multivariate analysis, adjusted for histological type and clinical T stage, a high CD133 proportion (>0%) was significantly correlated with poor response (= 0.018). The proportion of nucleostemin and GNL3L, and the intensity of TWIST were not correlated with the pathological response.

Table 4.   Pathological response to chemotherapy ratio (no. patients 1b + 2)/(no. patients 0 + 1a)
VariablesUnivariate odds ratio95% CIP-valueMultivariate odds ratio95% CIP-value
  1. †Union Internationale Contre le Cancer TNM classification. CI, confidence interval; cT, clinical tumor; GNL3L, guanine nucleotide-binding protein-like 3-like; M/D, moderately differentiated; P/D, poorly differentiated; Sq, squamous cell carcinoma; W/D, well differentiated.

CD133 (%)
 >0 vs =00.10(0.02, 0.45)0.0030.11(0.02, 0.68)0.018
Nucleostemin (%)
 80 vs <800.74(0.12, 4.55)0.7481.40(0.12, 16.1)0.787
GNL3L (%)
 50 vs <501.60(0.16, 15.6)0.6861.01(0.08, 13.1)0.994
TWIST intensity
 2, 3 vs 10.80(0.23, 2.8)0.7270.91(0.2, 4.13)0.904
cT†
 3 vs 20.35(0.04, 2.75)0.3190.61(0.04, 8.6)0.715
Histological type
 Sq (M/D + W/D) vs Sq (P/D)0.39(0.11, 1.32)0.1300.85(0.17, 4.31)0.844

Interestingly, patients with high nucleostemin proportion (>80%), strong TWIST intensity, and advanced pathological N stage were significantly correlated with poor RFS in the multivariate analysis adjusted for pathological T and N stages (= 0.025, 0.04, and 0.020, respectively) (Table 5). Poor performance status was significantly correlated with poor overall survival (OS) (= 0.012) (Table 6).

Table 5.   Relapse-free survival and clinicopathological variables
VariablesRisk Ratio95% CIUnivariate P-valueMultivariate P-value
  1. †Japanese Classification of Esophageal Carcinoma. CI, confidence interval; GNL3L, guanine nucleotide-binding protein-like 3-like; M/D, moderately differentiated; P/D, poorly differentiated; pN, pathological nodes; pT, pathological tumor; Sq, squamous cell carcinoma; W/D, well differentiated.

CD133 (%)
 >0 vs =03.920.9516.130.0500.058
Nucleosemin (%)
 ≥80 vs <806.701.2735.490.2200.025
GNL3L (%)
 ≥50 vs <504.880.6039.620.1940.138
TWIST intensity
 2, 3 vs 12.611.185.800.0130.018
Histological type
 Sq (M/D+W/D) vs Sq (P/D)0.480.181.280.2650.143
pT†
 3, 4 vs 1, 22.250.736.940.0140.159
pN†
 2–4 vs 0, 13.001.197.590.0200.020
Table 6.   Overall survival and clinicopathological variables
VariablesRisk Ratio95% CIUnivariate P-valueMultivariate P-value
  1. †Japanese Classification of Esophageal Carcinoma. CI, confidence interval; GNL3L, guanine nucleotide-binding protein-like 3-like; M/D, moderately differentiated; P/D, poorly differentiated; pN, pathological nodes; PS, performance status; pT, pathological tumor; Sq, squamous cell carcinoma; W/D, well differentiated.

CD133 (%)
 >0 vs =04.480.5536.670.0350.162
Nucleostemin (%)
 ≥80 vs <804.290.4937.760.1450.190
GNL3L (%)
 ≥50 vs <503.690.4331.940.3700.235
TWIST intensity
 2, 3 vs 11.830.774.340.0920.171
PS
 1 vs 03.061.287.360.0120.012
Histological type
 Sq (M/D+W/D) vs Sq (P/D)1.560.494.970.0070.455
pT†
 3, 4 vs 1, 22.090.557.910.0260.279
pN†
 2–4 vs 0, 12.350.836.670.0470.110

Correlation between nucleostemin and TWIST.  We recently reported that the cells overexpressing nucleostemin exhibited increased expression of TWIST, master regulator of EMT, and induced tumor metastasis.(16) In this study, we confirmed that the ectopic expression of nucleostemin induced the upregulation of TWIST expression in the human esophageal cancer cell line, KYSE180 (Fig. 2A). Moreover, we found that the endogenous nucleostemin expression level correlated with the TWIST expression level in esophageal cancer cell lines (Fig. 2B).

image

Figure 2.  Correlation between nucleostemin expression and TWIST expression. (A) Effects of the overexpression of nucleostemin (NS) on TWIST expression in KYSE180. (B) Correlation between nucleostemin expression and TWIST expression in esophageal cancer cell lines KYSE180, KYSE270, and TE-8. IB, immunoblotting.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

The present study demonstrated the significant expression of CD133 in a high percentage of surgically-resected esophageal squamous cell carcinoma specimens after neoadjuvant chemotherapy of patients who showed poor response to neoadjuvant chemotherapy. As the specimens from the biopsy do not necessarily exhibit a representative biological nature of the whole tumor tissues, and the esophageal cancers usually show strong heterogeneity, we need to be careful about the evaluation of the biopsy samples before neoadjuvant chemotherapy.

All of the specimens (100%) were evaluated as positive for CD44 staining, with no difference in staining pattern and intensity. Further evaluation of a possible correlation with chemotherapeutic effect or recurrence was insignificant. While this study found that nucleostemin is expressed in a high percentage of esophageal squamous cell carcinomas before and after chemotherapy, it is of great interest that the statistical analysis revealed a correlation between a proportion of positive cells after chemotherapy and recurrence. While all specimens exhibited positive staining for TWIST, if we carefully evaluated the staining intensity of TWIST, the statistical analysis would reveal that there is a significant correlation between staining intensity and recurrence rate following surgery. However, no molecular markers were correlated with OS.

Nucleostemin is a nucleolar GTP-binding protein, and recent work suggests that nucleostemin is involved in the regulation of cell proliferation in stem cells and in at least several types of cancer cells.(13) Recently Tamase et al.(15) reported that nucleostemin expression is essential for maintaining TIC in glioblastoma, and nucleostemin expression is closely related to the aggressive clinical outcome of glioblastoma cells. Moreover, we recently reported that the expression of nucleostemin drives the formation of cancer stem cells.(16) Specifically, cells expressing nucleostemin form tumors after the implantation of a small number of cells exhibit markers associated with TIC; induce the expression of genes that suffice to program induced pluripotent stem (iPS) cells; show activation of EMT, as gauged by TWIST and SNAIL expression; and show an increased propensity to metastasize. Our data suggest that the expression level of nucleostemin is correlated with clinical prognosis in esophageal cancer patients. It is noteworthy that this is the first clinical attempt to examine the clinical impact of the cancer stem cell factor(s) of nucleostemin in esophageal cancer.

Recent evidence indicates that increased single transducer and activator of transcription 3 (STAT3) signaling regulates the expression of the master regulator TWIST to induce EMT and metastasis.(14) Indeed, we recently showed that nucleostemin expression activates EMT and induces metastasis.(16) Specifically, we found that cells expressing nucleostemin exhibited increased expression of the tyrosine phosphorylated form of STAT3 and higher levels of TWIST. In addition, we found that tumorigenic cells expressing nucleostemin exhibited an increased capacity for metastasis. A study on esophageal squamous cell carcinoma reported that the TWIST expression rate was 80% or higher, and recurrence following surgery was high in patients with strong staining intensity.(20,21) Our data are consistent with these previous reports, and furthermore, it is intriguing as to whether the TWIST expression in residual cancer cells of surgical specimens after neoadjuvant chemotherapy was correlated to the subsequent recurrence.

In the present study, high nucleostemin expression and strong TWIST intensity in surgically-resected specimens of esophageal squamous cell carcinoma after neoadjuvant chemotherapy were significantly correlated with the possibility of postoperative recurrence. In addition, we found that endogenous nucleostemin expression level correlated with the endogenous TWIST expression level in esophageal cancer cell lines. These results suggest that nucleostemin and TWIST might serve as predictive markers for postoperative recurrence. Therapeutic strategies are expected to be developed, which would improve treatment outcomes in advanced esophageal squamous cell carcinoma with poor prognosis, by identifying the group at high risk for recurrence based on these factors. These findings imply that recurrence is related to the remaining fraction of chemo-resistant cancer stem cells after neoadjuvant chemotherapy.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

This work was supported in part by a funding program for Next Generation World-Leading Researchers (Next Program), Kato Memorial Bioscience Foundation and Japan Advanced Molecular Imaging Program (J-AMP) (all to KM). NO was supported by a research fellow of the Japan Society for the Promotion of Science. We thank Ms. Mami Yasukawa and Professor Kiyoko Fukami for their discussions and technical assistance.

Disclosure Statement

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References

The authors declare no financial or commercial conflict of interest.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Disclosure Statement
  8. References