Critical Role for TSLC1 Expression in the Growth and Organ Infiltration of Adult T-Cell Leukemia Cells In Vivo

Abstract

Adult T-cell leukemia (ATL) is associated with human T-cell leukemia virus type 1 infection. The tumor suppressor lung cancer 1 (TSLC1) gene was previously identified as a novel cell surface marker for ATL, and this study demonstrated the involvement of TSLC1 expression in tumor growth and organ infiltration of ATL cells. In experiments using NOD/SCID/γc^null mice, both leukemia cell lines and primary ATL cells with high TSLC1 expression caused more tumor formation and aggressive infiltration of various organs of mice. Our results suggest that TSLC1 expression in ATL cells plays an important role in the growth and organ infiltration of ATL cells.

Human T-cell leukemia virus type 1 (HTLV-1) is the causative agent of an aggressive form of CD4⁺ T-cell leukemia termed adult T-cell leukemia (ATL) (7, 14, 18). Carriers of HTLV-1 have been identified in a number of locations throughout the world, including parts of Africa; Papua New Guinea; specific regions in Europe including Romania; parts of South America including northern Brazil, Peru, northern Argentina, and Colombia; and the southern part of Kyushu in Japan (17). Common findings in patients with ATL include enlargement of peripheral lymph nodes, hepatomegaly, splenomegaly, skin infiltration, and hypercalcemia. The Tax gene is a unique viral gene thought to play a central role in HTLV-1-induced transformation. It is responsible for transactivation of the HTLV-1 long terminal repeat (5, 16) and numerous cellular genes involved in T-cell activation and growth, including those encoding interleukin-2 (IL-2) (11) and the α chain of IL-2 receptor (IL-2Rα) (CD25, Tac) (1, 2). The long latency of ATL development suggests that multiple genetic events accumulate in HTLV-1-infected cells; however, the precise molecular mechanisms of ATL leukemogenesis following HTLV-1 infection have not been fully elucidated.

The tumor suppressor lung cancer 1 gene (TSLC1) at chromosome 11q23 has been identified as a tumor suppressor gene in non-small-cell lung cancer (9, 13). In contrast, it was recently found to be highly and ectopically expressed in acute-type ATL cells, most ATL cell lines, and HTLV-1-infected T-cell lines (15). Enforced expression of TSLC1 in ATL-derived ED-40515(−) cells resulted in higher aggregations and binding abilities in a human umbilical vein endothelial cell line (HUVEC). These results suggest that TSLC1 might contribute to tumor growth by enhancing aggregation after infiltration and migration outside blood vessels. Since the role of TSLC1 overexpression in the course of tumor growth and organ infiltration of ATL cells remains to be fully elucidated, we investigated the direct involvement of TSLC1 in the growth and infiltration of leukemia cells using C57BL/6J and NOD-SCID/γc^null (NOG) mice (4, 8).

In order to analyze the tumorigenicity of TSLC1 expression in leukemia cells, a murine IL-2-independent T-lymphoma cell line (EL4) injected into the intraperitoneum of syngeneic C57BL/6J mice was used as a model for ATL. EL4 cells were transfected with a pcDNA3 expression plasmid containing TSLC1, and transformant cells were selected by a limiting-dilution method in the presence of G-418. We also used EL4 cells expressing a green fluorescent protein-Tax fusion protein (EL4/GAX) (6) and parental EL4 (EL4/p) as a control. Expression of Tax protein in EL4 cells, a 38-kDa band of Tax protein in HUT102 cells, and a 64-kDa band of green fluorescent protein-Tax fusion protein in EL4/GAX cells were all detected by Western blot analysis (Fig. 1A). Expression of a TSLC1 protein in EL4/TSLC1 cells was also shown on Western blot analysis with KK1, an ATL cell line expressing TSLC1 (12) (Fig. 1B). In an in vitro cell growth assay, 2 × 10⁴ cells were incubated, and their growth was analyzed by direct counting with trypan blue dye staining. EL4 and EL4/TSLC1 cells showed nearly identical proliferation profiles in vitro, while Tax-expressing EL4 cells proliferated more slowly (Fig. 1C). This difference in cell growth might be caused by different expression vectors. In an in vivo growth assay, 2 × 10⁶ cells of each cell line were injected into the peritoneal cavity of C57BL/6J mice: eight mice for EL4 cells as controls, 13 mice for EL4/TSLC1 cells, and eight mice for EL4/GAX cells. All of the mice died of tumor invasion of various organs with ascitic fluids in 40 to 120 days. The median survival time of the control mice injected with EL4 cells or EL4/GAX cells was 72 days. The mice with EL4/TSLC1 cells, however, died within 60 days, with a median survival time of 41 days (Fig. 1D). The phenotypes of the control mice and the EL4/TSLC1 mice were almost identical with invasion of tumors into various organs. Organ metastasis of tumor cells in three EL4/TSLC1-inoculated mice, two EL4-inoculated mice, and one EL4/GAX-inoculated mouse was analyzed and evaluated with hematoxylin-eosin staining. The liver was one of the major sites of metastasis in all three of the EL4/TSLC1-inoculated mice by histopathological analysis but not in the two EL4-inoculated mice or the EL4/GAX-inoculated mouse (Fig. 1E). These results support the role of TSLC1 overexpression in T-lymphoma cells as one of an aggressive factor in the development of leukemia/lymphoma.

FIG. 1.

Transplantation of EL4 T-cell lymphoma cells expressing TSLC1 shortened the life span of syngeneic mice. (A) Expression of Tax protein in HUT102, EL4, EL4/GAX, and EL4/TSLC1 cells was detected by Western blot analysis. Expression of β-actin protein (ACTB) was used as a loading control. (B) Expression of TSLC1 protein in KK1, EL4, EL4/GAX, and EL4/TSLC1 cells was detected by Western blot analysis. Expression of β-actin protein (ACTB) was used as a loading control. (C) Cell numbers in a growth curve are shown for an average of three independent counts, and standard deviations are indicated as error bars. (D) Survival curves of C57BL/6 mice inoculated in the abdominal cavity with EL4, EL4/GAX, or EL4/TSLC1 cells. Cumulative survival rates were calculated by the Kaplan-Meier method and compared using a log-rank test. (E) Liver sections from all mice were stained with hematoxylin-eosin. The regions of liver metastasis (arrow) were seen in liver sections from mice inoculated with EL4/TSLC1 cells but not shown in the liver sections from the mice inoculated with EL4 or EL4/GAX cells. Magnification, ×100; bars, 400 μm.

In order to investigate the possibility that overexpression of TSLC1 promotes tumor growth and/or infiltration in vivo, ATL-derived ED-40515(−) cells (10) were injected into NOG mice. Since expression of TSLC1 in ED-40515(−) cells is severely reduced by promoter methylation, they were transfected with either a TSLC1 expression plasmid (pcDNA3/TSLC1) or a mock plasmid (pcDNA3/Neo). ED/TSLC1 and ED/Neo cells were identified by selection with G-418. High levels of TSLC1 expression were verified in the ED/TSLC1 cells, but not in the ED/Neo cells, by Western blot analysis (Fig. 2A). The ED/TSLC1, ED/Neo, and ED-40515(−) cell lines all showed the same proliferation profile in vitro (Fig. 2B). Cells (10 × 10⁶) were inoculated subcutaneously into the postauricular region of NOG mice, which permitted the observation of tumor growth macroscopically and the measurement of tumor size over a relatively short time (3). The ED/TSLC1 cell lines caused greater formation of larger tumors than did the ED/Neo and ED-40515(−) cell lines (Fig. 2C). The development of clinical signs of near-death (e.g., piloerection, weight loss, and cachexia) in mice at the time of killing was also more prevalent with the ED/TSLC1 cell line. These results suggest that TSLC1 expression in ATL cells enhances in vivo tumor growth in NOG mice.

FIG. 2.

Involvement of TSLC1 expression in tumor growth and infiltration of leukemia cells in NOG mice. (A) Expression of TSLC1 in KK1, ED-40515(−), ED/Neo, or ED/TSLC1 cell lines was detected by Western blot analysis. Expression of β-actin protein (ACTB) was used as a loading control. (B) Cell growth curves of ED-40515(−), ED/Neo, and ED/TSLC1 cell lines are shown for an average of three independent counts, and standard deviations are indicated as error bars. (C) Tumor volumes of mice inoculated subcutaneously with ED/TSLC1, ED/Neo, or ED-40515(−) cells after 21 days are shown as the means ± standard errors of the means for five mice in each group. Statistical analysis was done with a Student t test. (D and E) The pictures shown were derived from gross photographs of the sacrificed mice at 1 month after intravenous inoculation of ED/Neo (D) or ED/TSLC1 (E) cells. (F) Immunohistochemical staining for TSLC1 protein in liver metastases of the mice inoculated intravenously with ED/TSLC1 cells is shown. An arrow indicates a tumor mass with strong staining with a rabbit anti-TSLC1 antibody; however, the same mass shows no staining with rabbit immunoglobulin (Ig) as a negative control. Magnification, ×100; bars, 400 μm.

Since the mice died within 4 weeks after subcutaneous inoculation of leukemia cells due to heavy tumor burden, 2 × 10⁶ ED/TSLC1 or ED/Neo cells were intravenously injected into six NOG mice in order to investigate their capacity for invasion of various organs. After 1 month, we sacrificed the mice to determine the extent of organ invasion. Macroscopically, all of the mice injected with ED/TSLC1 cells (six/six) showed severe liver invasion with swelling of the ovaries. None of the mice injected with ED/Neo cells showed liver invasion, but they did show ovarian involvement (Fig. 2D and E). Microscopically, all of the mice inoculated with ED/TSLC1 cells showed severe and massive liver and lung invasions. On the other hand, only one of six mice inoculated with ED/Neo cells showed a large amount of liver metastasis (Table 1). TSLC1 expression in tumor cells infiltrating the liver was confirmed by immunohistochemical staining (Fig. 2F). Thus, overexpression of TSLC1 in ATL cells might enhance organ invasion, and particularly invasion of the liver and lung.

TABLE 1.

Invasion scores of mice inoculated with ED/Neo or ED/TSLC1 cells

Cell line and mouse	Invasion score for organ by observation:
	Macroscopica					Microscopicb
	Liver	Kidney	Lung	Ovary	Spleen	Liver	Kidney	Lung	Ovary	Spleen
ED/TSLC1
T1	3+	−	+/−	1+	−	3+	−	2+	2+	−
T2	3+	−	−	1+	−	3+	−	2+	2+	−
T3	3+	−	+/−	2+	−	3+	−	2+	2+	−
T4	3+	−	−	1+	−	3+	−	2+	2+	−
T5	2+	−	−	2+	−	3+	−	2+	3+	−
T6	3+	−	+/−	1+	−	3+	−	+/−	2+	−
ED/Neo
N1	−	−	−	2+	−	2+	−	+/−	3+	−
N2	+/−	−	−	1+	−	+/−	−	−	2+	−
N3	−	−	−	2+	−	−	−	+/−	2+	−
N4	−	−	−	1+	−	−	−	−	2+	−
N5	−	−	−	1+	−	NDc	ND	ND	ND	ND
N6	−	−	−	1+	−	ND	ND	ND	ND	ND

^aSubjective invasion scores by macroscopic observation were as follows: −, no invasion; +/−, less than 10% invasion in the organ; 1+, 10 to 30% invasion in the organ; 2+, 30 to 70% invasion in the organ; 3+, over 70% invasion in the organ.

^bSubjective invasion scores by microscopic observation were as follows: −, no invasion; +/−, less than 1% leukemia cells in the section; 1+, less than 10% leukemia cells in the section; 2+, 10 to 30% leukemia cells in the section; 3+, over 30% leukemia cells in the section

^cND, not done.

Next, we examined whether primary ATL cells with various levels of expression of TSLC1 could efficiently grow and infiltrate various organs in NOG mice. TSLC1-positive primary ATL cells (2 × 10⁷) from five acute-type and five chronic-type ATL patients were inoculated subcutaneously into the postauricular region of NOG mice (Table 2). All of the mice developed clinical signs of near-death (e.g., piloerection, weight loss, and cachexia) 6 to 8 weeks after inoculation, in addition to the enlargement of the lymph nodes, spleen, lungs, and liver. Microscopically, ATL cells invaded various organs of all ATL-bearing NOG mice to different degrees. Based on results of immunohistochemical staining for TSLC1, all invading leukemia cells expressed TSLC1 protein, compared with no TSLC1 expression in these organs in control NOG mice (Table 2 and Fig. 3A). The dispersion diagram for the levels of invasion and the levels of TSLC1 expression in the leukemia cells showed a correlation coefficient of 0.714, suggesting that there was a moderate correlation between invasive capability and the level of TSLC1 expression (Fig. 3B). Thus, TSLC1 could aid in the formation of a rapidly growing large tumor and massive infiltration of ATL cells into various organs in NOG mice. Since TSLC1 is expressed in various types of ATL cells, including smoldering and chronic types, it might be a promising target for the development of a new anti-ATL therapy. The NOG mouse model system described in the present study could provide a novel means by which to understand and investigate the further importance of TSLC1 in ATL progression.

TABLE 2.

Clinical characteristics of patients and pathological findings of organ invasion^a

Patient no.	Age (yr)/sex	Clinical characteristic				Invasion score in NOG miceb				TSLC1 expression scorec
		Diagnosis (ATL type)	WBC (109/liter)	Lymphocytes (%)	Atypical cells (%)	Liver	Lung	Spleen	Lymph node
1	73/M	Chronic	7.8	59	47	3+	3+	3+	ND	3+
2	59/F	Chronic	9.0	75	40	3+	2+	2+	1+	2+
3	66/F	Chronic	29.4	49	75	3+	3+	3+	ND	3+
4	44/F	Chronic	22.6	51	45	3+	2+	2+	2+	2+
5	43/F	Chronic	18.6	63	43	3+	3+	3+	ND	2+
6	54/M	Acute	192.8	65	91	1+	2+	ND	ND	1+
7	58/M	Acute	67.3	71	80	3+	3+	3+	ND	2+
8	65/F	Acute	29.4	25	60	3+	2+	ND	3+	3+
9	68/M	Acute	30.0	79	81	3+	1+	1+	2+	2+
10	66/F	Acute	10.2	38	51	3+	3+	3+	ND	3+

^aAbbreviations: M, male; F, female; WBC, white blood cells; ND, not done.

^bSubjective invasion scores were as follows; 0, no invasion; 1+, less than 10% leukemia cells in the section; 2+, 10 to 30% leukemia cells in the section; 3+, over 30% leukemia cells in the section.

^cSubjective scores of TSLC1 expression in pathological immunostaining were as follows; −, no staining; 1+, faint staining in less than 10% of invasive leukemia cells; 2+, weak to moderate staining in 30 to 70% of invasive leukemia cells; 3+, intense staining in more than 70% of invasive leukemia cells.

FIG. 3.

Growth and infiltration of primary ATL cells in various organs of NOG mice based on TSLC1 expression. (A) Immunohistochemical staining of various organs of NOG mice inoculated with leukemia cells from patient 6, 8, 9, or 10 is shown with the use of rabbit anti-TSLC1 antibody or rabbit immunoglobulin (Ig) as a negative control. Sections from patients 8 and 10 showed severe invasion (invasion score, 3) and dense staining for TSLC1 (expression score, 3), while sections from patients 6 and 9 showed mild invasion (invasion score, 1) and light staining for TSLC1 (expression score, 1). Liver and lung sections from control NOG mice were used as negative controls, and a lymph node from an ATL patient was used as a positive control. Magnification, ×400; bars, 100 μm. (B) The diagram of dispersion between mean values of each invasion score and scores for TSLC1 expression in each NOG mouse inoculated with primary ATL cells showed moderate correlation (R = 0.714).