A proposed explanation for female predominance in alveolar soft part sarcoma

Noninactivation of X; Autosome translocation fusion gene?

Abstract

BACKGROUND

Alveolar soft-part sarcoma (ASPS) is a rare malignant soft tissue tumor with both clinically and morphologically distinct features. It often involves the extremities of adolescents and young adults and shows a predilection for females. Recently, ASPS was found to have a nonreciprocal der(17)t(X;17) translocation with the corresponding fusion gene located in chromosome 17. Because females have an extra X-chromosome, their likelihood of developing an X;autosome translocation is theoretically double that of males, and thus, this extra X-chromosome is a likely explanation for female predominance of ASPS.

METHODS

The authors used data from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) registry program, which included 87 ASPS cases (33 males and 54 females), and published ASPS cases, which included 317 cases (121 males and 196 females), to test our hypothesis. The authors compared the observed proportion of female cases with that expected under the two X-chromosomes–double-risk hypothesis including the consideration of X-inactivation status.

RESULTS

The hypothesis that the fusion gene is not subject to X-inactivation is supported by data (P = 0.6, 0.24, and 0.20 for SEER cases, published cases, and their combination, respectively). In contrast, the competing hypothesis that the fusion gene is subject to X-inactivation is rejected (P = 0.007, < 0.00001, and < 0.00001 for SEER cases, published cases, and their combination, respectively).

CONCLUSIONS

Therefore, the authors found a statistical association between the female predominance observed in ASPS and female possession of an extra X-chromosome/noninactivation of the ASPS X;autosome translocation fusion gene. Cancer 2005. © 2005 American Cancer Society.

Alveolar soft-part sarcoma (ASPS), first described by Christopherson et al. in 1952,1 is a rare malignant soft tissue tumor with several unique features. Morphologically, the organoid nests of large esosinophilic cells are arranged in a highly characteristic pseudoalveolar pattern. ASPS also shows distinctive intracytoplasmic rhomboid to polygonal crystals at the ultrastructural level. A recent study showed that these crystals contain monocarboxylate transporter 1 and CD147.2 Clinically, ASPS most often involves the extremities and, less frequently, head and neck of adolescents and young adult and, unlike most other sarcomas, shows an intriguing female predominance.3 A 2:1 female to male ratio was originally reported by Christopherson et al. in 1952,1 and many other follow-up series have shown the same female predominance. This tumor is indolent, but patients often develop distant metastases and die. Despite numerous immunohistochemical, ultrastructural, and molecular studies, two major questions remain unanswered, one regarding the histogenesis of ASPS and the other, the female predominance of ASPS. The authors address the latter question in this study.

ASPS was found to have a recurrent nonreciprocal der(17)t(X;17)(p11;q25) translocation.4, 5 More recently, this translocation was found to involve ASPL gene on chromosome 17 and TFE3 gene on X-chromosome in all 13 cases tested.6, 7 The corresponding ASPL-TFE3 fusion gene with the N-terminal portion of TFE3 gene being fused with C-terminal portion of ASPL gene is located in chromosome 17.7 Because an ASPL-TFE3 fusion transcript was detected in all 13 ASPS cases tested, it is reasonable to speculate that this nonreciprocal der(17)t(X;17)(p11;q25) translocation, or more specifically ASPL-TFE3 fusion gene, is the critical event in the development of ASPS. TFE3 gene encodes a transcription factor and is broadly expressed. ASPL-TFE3 fusion gene retains the TFE3 DNA-binding domain and, thus, may cause transcriptional deregulation in tumorigenesis.7

Although the likelihood females developing an X;autosome translocation is theoretically double that of males because of female possession of an extra X-chromosome, a likely explanation for the female predominance seen in ASPS, the final disease phenotype will certainly be influenced by the X-inactivation process, a mechanism to compensate this extra X-chromosome in females. When an X;autosome translocation occurs and involves the inactivated X-chromosome, the X portion in the derivative autosome can theoretically either maintain its original inactivation status or become active. According to the current X-inactivation theory, X inactivation is initiated and maintained by the presence of in cis, the X-inactivation center located at Xq13.2.8 The process is likely to be mediated by transcripts from X-inactive–specific transcript gene (XIST) gene.9 Thus, any X segment lacking XIST is expected to be active regardless of its original inactivation status. More specifically, although the Xp segment in the derivative autosome can originally come from an inactivated X-chromosome, it will become active, as it has been physically disconnected from XIST. That the X portion translocated onto the derivative autosome is rarely inactivated has been well documented.10, 11 Further, studies on hypomelanosis of Ito (HI) provide a real example of this phenomenon. HI is a heterogeneous disorder and is characterized by skin hypopigmentation and central nervous system abnormalities. A subset of HI patients have a structurally balanced X;autosome translocation. Many studies now indicate that phenotypes in these females are caused by the mosaic functional disomy of Xp distal to the breakpoint, rather than disruption of X-linked genes.12–14

According to a recent study by Carral et al.,15TFE3 gene is subject to X inactivation in females. However, in ASPS, the lesional ASPL-TFE3 fusion gene, which is located in chromosome 17 and physically separated from the XIST gene, will hence not be subject to X inactivation. Therefore, the authors can conceivably hypothesize that the female predominance seen with ASPS is due to both the possession of an extra X-chromosome in females and the oncogenic X;autosome translocation fusion gene not being subject to X inactivation. The authors used two large data sets to test this hypothesis at the population level.

MATERIALS AND METHODS

We identified all ASPS cases from the population-based, National Cancer Institute-sponsored Surveillance, Epidemiology, and End Results (SEER) registries.16 Eighty-seven ASPS cases (ICDO morphology code 9581) were diagnosed in 11 SEER registry regions (9 registries from 1973–91 and 11 registries from 1992–99) (Table 1). Because the observed male and female population counts within the SEER registries differ within each age group, the authors age-adjusted the observed male and female case counts in each age group by applying sex- and age-specific incidence rates for ASPS to a standard population that was obtained by summing the male and female populations at risk over the entire time period in each age group and dividing by 2. We then summed the age-specific numbers to obtain a total adjusted count of cases for males and for females from which an observed proportion of female cases was calculated.

Table 1. Alveolar Soft Part Sarcoma Incidence in Surveillance, Epidemiology, and End Results Registries, 1973–1999

Age Group (yrs)	Males		Females		Adjusted No. male casesa	Adjusted No. female casesa
	No. of cases	Age-specific incidence rate/million	No. of cases	Age-specific incidence rate/million
a Adjusted numbers of cases were obtained by applying sex-specific and age-specific incidence rates to a standard population obtained by summing the male and female populations at risk over the entire time period in each age group and dividing by 2.
0–4	0	0.00	2	0.08	0	2.1
5–9	2	0.08	4	0.16	2.0	4.1
10–14	5	0.19	3	0.12	4.9	3.1
15–19	0	0.00	7	0.26	0	7.2
20–24	8	0.28	10	0.36	7.9	10.1
25–29	6	0.20	15	0.51	6.0	15.1
30–34	1	0.03	5	0.17	1.0	5.0
35–39	5	0.19	4	0.15	5.0	4.0
40–44	1	0.04	1	0.04	1.0	1.0
45–49	0	0.00	1	0.05	0	1.0
50–54	1	0.06	1	0.06	1.0	1.0
55–59	2	0.13	0	0.00	2.1	0
60–64	2	0.15	0	0.00	2.1	0
65–69	0	0.00	0	0.00	0	0
70–74	0	0.00	1	0.08	0	0.9
≥75	0	0.00	0	0.00	0	0
Total	33		54		33.0	54.6

The authors also compiled all then-available published studies of ASPS with adequate information by searching MEDLINE® (a registered name of the National Library of Medicine; URL:http://www.ncbi.nlm.nih.gov/pubmed; search terms: alveolar soft part sarcoma; date of search: May 21, 2002). This compilation includes U.S. cases before 1973 and non-U.S. cases that appeared in English-language publications before 2002 (Table 2).4, 5, 17–133 U.S. cases after 1973 were excluded because some would be expected to also be in the SEER database. As these published cases are primarily from case reports and case series, the authors did not make adjustments for the possible unbalanced gender distribution among different age groups.

Table 2. Published Alveolar Soft Part Sarcoma Cases

No. of males	No. of females	Country	Year of diagnosis or publication	References
19	34	U.S.	Before 1964	17
1			1956	18
1	1	Japan	1957	19
	1	India	1963	20
1			1963	21
1	1	Mexico	1966	22
1		U.S.	1966	23
	1		1967	24
1		Jamaica	1968	25
1		Malaysia	1969	26
	4	Norway	1969	27
	1	U.S.	1969	28
	1	Malaysia	1970	29
	1	U.K.	1970	30
1		Israel	1971	31
	1	U.S.	1971	32
	1	U.S.	1971	33
1		U.S.	1972	34
	1	U.S.	1972	35
3	1	U.S.	1972	36
	1	India	1973	37
1	4	Japan	1975–1995	38
2	1	Japan	1975–1997	39
	2	Netherlands	1976	40
	1	India	1977	41
	1	Norway	1977	42
2		U.K.	1977	43
	1	India	1978	44
1		Japan	1978	45
	1	Netherlands	1978	46
	2	Nigeria	1978	47
	1	U.K.	1978	48
1		India	1979	49
	1	Japan	1979	50
	1	Peru	1979	51
6	13	Italy	1979–1999	52
	1	Japan	1980	53
	1	U.K.	1981	54
	1	Germany	1982	55
	1	Singapore	1982	56
4	2	U.K.	before 1982	57
1		Yugoslavia	1982	58
	1	Germany	1983	59
1	1	Japan	1984	60
1		U.K.	1984	61
	1	Taiwan	1984	62
	1	Japan	1984	63
1		France	1985	64
	1	India	1985	65
	1	Nigeria	1985	66
	3	Japan	1985–1993	67
1		Japan	1986	68
1		Canada	1988	69
	1	Hong Kong	1988	70
	1	Hong Kong	1988	71
	1	Italy	1988	72
1		Netherlands	1988	73
4	6	Sweden	1988	74
1		U.K.	1988	75
	1	Italy	1988	76
	1	Italy	1989	77
	1	Norway	1989	78
	1	Turkey	1989	79
7	3	Japan	1989	80
	1	Ireland	1990	81
	1	Italy	1990	82
1		Japan	1990	83
	1	U.K.	1990	84
8	4	Japan	1965–1988	85
1		Japan	1990	86
1	4	Japan	1990	87
1		S. Africa	1990	88
2	5	Finland	1990	89
	1	Japan	1991	90
2	6	Poland	1991	91
	1	Switzerland	1991	92
	1	Canada	1992	93
3	3	Hong Kong	1992	94
	1	Japan	1992	95
	1	U.K.	1992	96
	1	Belgium	1993	97
	1	France	1993	98
8	2	Japan	1993	99
1		Japan	1993	100
	1	Japan	1993	101
1		Singapore	1993	102
1		U.K.	1993	103
1	8	U.K.	1993	104
1		Canada	1994	105
	1	Taiwan	1994	106
	2	U.K.	1994	107
1		Canada	1995	108
	1	Denmark	1995	109
4	6	Japan	1995	110
	1	Netherlands	1995	111
	1	Taiwan	1995	112
	1	India	1996	113
	1	Italy	1996	114
	1	Taiwan	1997	115
	1	Taiwan	1997	116
1		U.K.	1997	117
	1	Belgium	1998	4
	1	Brazil	1998	118
4	5	Canada	1998	119
	2	Germany	1998	120
	1	Japan	1998	121
1		Germany	1999	122
1		Japan	1999	5
	1	Japan	1999	123
	1	U.K.	1999	124
	1	Japan	2000	125
	1	Japan	2000	126
3	7	Korea	2000	127
4		Germany	2000	128
	2	Hong Kong	2001	129
2	2	Hungary	2001	130
	1	France	2001	131
	1	France	2001	132
1		Japan	2001	133

The expected male and female proportions of ASPS are calculated according to our previous publication on synovial sarcoma.134 Briefly, should a chromosomal translocation develop in a cell, the conditional probabilities for such a translocation being an ASPL-TFE3 translocation and the resultant fusion gene being functional is:

• 

in which S, G, and C are target gene size, total human genome size, and constant, respectively. The I (in IS)is used to factor in the number of S_TFE3 genes and/or the functional status of ASPL-TFE3 fusion gene. Therefore, I = 1 for males or for females under the hypothesis that the fusion gene is subject to X-inactivation, i.e., only half of the ASPL-TFE3 fusion genes will be functional due to the X-inactivation. For females under the hypothesis that the fusion gene is not subject to X-inactivation, I=2. If we assume C_f = C_m, then the expected proportion of male cases will be

• 

The total human genome sizes G_m and G_f for males and females, respectively, were obtained from the most recent publication of the Human Genome Project.135

A Z test was used to determine whether the observed proportion of female cases (standardized to the average person-years distribution for the SEER data) differed significantly from the expected proportion based on the two–X-chromosomes double-risk hypothesis with the consideration of X-inactivation status. The SEER cases and the cases derived from the literature were analyzed separately and then combined.

RESULTS

Table 1 provides age-specific and sex-specific incidence rates of ASPS and the number of ASPS cases by gender for 9 SEER registries from 1973–91 and for 11 SEER registries from 1992–99. The observed proportion of female cases standardized to the average population distribution covered by SEER during these years is 62.3%. Under the hypothesis that the fusion gene is not subject to X-inactivation, the expected proportion for female cases is 65%, which does not differ from the observed proportion (Z = 0.53 and P = 0.6) (Table 3). In contrast, under the competing hypothesis that the fusion gene is subject to X-inactivation, the expected portion of female cases calculated from Equation 2 is 48%, which differs significantly from the observed proportion (Z = 2.7; P = 0.007) (Table 3).

Table 3. Alveolar Soft Part Sarcoma Cases and Z Test Results for Competing Hypotheses Regarding X Inactivation.

Data set	Males	Females	Total	P values
				No X-inactivation	X-inactivation
a Adjusted number of SEER (Surveillance, Epidemiology, and End Result) registry cases. (Adjusted numbers of cases were obtained by applying sex-specific and age-specific incidence rates to a standard population obtained by summing the male and female populations at risk over the entire time period in each age group and dividing by 2.)
SEER casesa	33	54.6	87.6	0.60	0.007
Published cases	121	196	317	0.24	<0.00001
Total	154	250.6	404.6	0.20	<0.00001

For the compilation of published cases, the observed proportion of female cases is calculated as 61.8%, which does not differ from the expected proportion under the hypothesis that the fusion gene is not subject to X-inactivation (Z = 1.18; P = 0.24) (Table 3). Under the competing hypothesis that the fusion gene is subject to X-inactivation, the expected proportion of female cases differs significantly from that the observed (Z = 4.93; P < 0.00001) (Table 3).

When both SEER cases and compiled published cases are combined, a similar observed proportion, 61.9%, of female cases is obtained. The null hypothesis that the fusion gene is not subject to X-inactivation is supported (P = 0.20), whereas the competing hypothesis is rejected (P < 0.00001) (Table 3).

DISCUSSION

Based on the biology of X-inactivation, autosomal location of the ASPL-TFE3 fusion gene, and female predominance seen in ASPS, the authors hypothesized that the increased risk of ASPS in females is due to their possession of an extra X-chromosome and the ASPL-TFE3 fusion gene being not subject to X-inactivation. The authors tested this hypothesis and found that both SEER registry data and published cases are consistent with this hypothesis. The SEER cases are population based, whereas cases from literature compilation are not, and thus the latter may be subject to bias. Nevertheless, the observed proportion of female cases in both data sets was similar. The 2 data sets together have more than 400 ASPS cases, which is the largest compilation of cases to date. The combination of these two complementary data sets likely represents a true picture of ASPS in terms of the sex-specific proportion of cases. Similar results were obtained from analyzing both data sets separately and in combination, thus providing further support for this argument. In the current study, the authors were unable to examine results by race and geographic region because of limited numbers of cases and/or incomplete information.

ASPS and synovial sarcoma (SS) share many similarities. They both are malignant soft-tissue sarcomas with unique morphology and unknown histogenesis, and both ASPS and SS show a characteristic X;autosome translocation, ASPS at t(X;17) and the others at t(X;18). However, some key differences between the two tumors exist. In contrast to ASPS, SS does not show a female predominance, and its SYT-SXX fusion gene is located in the derivative X-chromosome, and, thus, it is subject to X-inactivation. Based on the model in the current study, the X-inactivation is probably a key factor in determining the fate of an oncogenic X;autosome translocation fusion gene. Further, using SEER data on SS, the authors previously provided population-based evidence to support our hypothesis that no increased risk of SS in females is due to t(X;18) fusion gene being subject to X-inactivation.134

Interestingly, a balanced t(X;17)(p11.2; q11) has been reported in a subset of young patients with renal cell carcinoma.6, 136, 137 The same fusion transcript ASPL-TFE3 as that of ASPS was identified in these patients. Unlike in ASPS, a balanced or reciprocal translocation is present in all cases tested by fluorescence in situ hybridization.6 A reciprocal fusion product, TFE3-ASPL, was found in all three males, but only in two out of five females, by reverse transcriptase-polymerase chain reaction (RT-PCR). The existence of female patients with identifiable ASPL-TFE3 fusion product but no TFE3-ASPL fusion product appears initially confusing but, actually, is consistent with and supportive of our model. According to our model, the TFE3-ASPL fusion gene at the der(X) is subject to X-inactivation; thus a fusion transcript can only be identified in those females with activated der(X) but not in those females with inactivated der(X). In contrast, ASPL-TFE3 fusion gene on the der(17) is not subject to X-inactivation, and, therefore, is functional.

The results of the present study should be interpreted with caution because what the authors found is just a statistical association or population-based evidence awaiting future molecular evaluation. There are many tumors that have unequal sex ratios but no known genetic involvement of X-inactivation. Our model is only applicable to very few tumors with a defined X;autosome translocation under a certain precondition. In addition, our model is also limited because other alternate non–X chromosome related hypotheses, such as hormonal, occupational, or social factors, can theoretically explain the female predominance in ASPS.

In summary, using two complementary data sets, the authors have shown a statistical association between the increased risks of ASPS in females, i.e., female predominance and their possession of an extra X-chromosome and ASPS t(X;17) translocation fusion gene being not subject to X-inactivation. However, evaluation of this hypothesis at the molecular level is warranted and may explain how a chromosomal translocation can modify the gender distribution of a sarcoma.