Differential expression of THOC1 and ALY mRNP biogenesis/export factors in human cancers

cDNA probes were obtained by PCR of cDNA clones purchased from I.M.A.G.E. consortium. Mouse monoclonal anti-THOC1 and anti-ALY were purchased from Abcam, and secondary reagents were purchased from Dako for immunohistochemistry analysis (IHC).

Breast cancer cell lines (MCF7, SKBr-3 and T47-D), and MCF10-A were purchased from American Type Culture Collection and propagated according to the conditions recommended by the vendor.

1 mm tissue cores were obtained from archival paraffin blocks of tumors resected at Hospital Universitario Virgen del Rocío, and then arrayed into recipient paraffin blocks. Pathological diagnoses are provided in Additional file 1. The collection and use of the human material was approved by the local ethics committee.

Total RNA was isolated from breast cell lines using Purescript RNA isolation kit (Gentra System), cDNA was synthesized with transcriptor first-strand cDNA synthesis kit (Roche) and Real-time PCR was performed with SYBR green dye in the 7500 Real Time PCR system of Applied Biosystems by following the manufacturer's instructions. THO-UAP56 and ALY specific primers were designed using Primer Express software (the sequence is available upon request). The HPRT1 gene was used as an endogenous control after evaluating different housekeeping genes in several breast cancer cell lines and identifying HPRT as the most stable internal control.

Immunohistochemical analyses (IHC) were performed on tissue microarray four-micron sections. Monoclonal antibodies against THOC1 and ALY were used. Tissues were microwaved in 1 mM EDTA buffer (for ALY) or treated with 4N HCl and trypsin for THOC1. Primary antibody incubation was overnight at 4°C. Secondary reagents (Dako) were applied according to manufacturer's protocol. Slides were then counterstained with hematoxylin and mounted in DPX. Sections where the primary antibody was omitted were used as negative controls. Immunostaining was evaluated at least on 10 microscopic fields at × 200 by two independent pathologists and unequivocal immunostained cells were counted. Immunohistochemical expression was scored as strong (>50% of the carcinoma cells stained), weak (10%-50%) or negative (<10%). Digital images were acquired using a Zeiss microscope.

A two-tailed paired t-test was used to determine whether the differences in THOC1, ALY and hSPT4 cDNA levels in tumor versus normal samples were statistically significant. To determine the association between the expression levels of ALY with a high grade tumor, a chi-squared test was performed.

First the mRNA levels of THOC1 and ALY in tumor tissues by hybridization of a Cancer Profiling Array II (Clontech) were evaluated. THOC1, was taken as representative of factors that act at the transcription level such as THO-UAP56, and ALY, as a related factor that serves as adaptor in mRNA export. We also analyzed the expression of hSpt4, the subunit of the human transcription elongation factor DSIF, [18], which participates early in transcription and is not related to mRNA export. The array contains normalized cDNAs from tumor and matched normal tissues from a variety of organs (154 cDNAs pair samples derived from 19 different tissues and 3 to 11 patients by tissue). The results of the hybridization and the relative expression in tumor samples calculated as the ratio of tumor versus normal signal (T/N) are shown in Figure 1A and Additional file 2. We considered those cases in which the values of T/N ratio of cDNAs were outside the interval 0.5-1.5 (Figure 1B). THOC1 was ubiquitously expressed in all normal tissues (Additional file 2) and a significant increase was found in ovary, stomach and lung tumors (marked with *; for more details see Additional file 3) with the highest levels of overexpression in ovary (4 fold higher than those of the matched normal samples) (Figure 1A). In addition, a statistically significant reduction of THOC1 transcripts was observed in thyroid, testis and skin tumors, even though in the later the differences were lower (Figure 1). We have compared our results with the SAGE (Serial Analyses of Gene Expression) data of THOC1 available at the human Gene Compendium GeneCard http://​www.​genecards.​org. The GeneCard data show changes in THOC1 mRNA levels in tumor versus normal tissues, in almost all samples analyzed: higher levels in lymphoid node, breast, lung, colon, liver, pancreas, prostate and skin and lower in thyroid, cortex and cerebellum. In contrast our analyses report significant changes in only a subset of the tissues analyzed, as indicated above.

The expression pattern of ALY was analyzed with the same cDNA array. Hybridization with an ALY specific probe revealed a general and strong signal for ALY transcripts in all normal tissues tested, with the highest levels in testis (Additional file 2). ALY mRNA was increased in a wide variety of tumors, whereas it was not significantly reduced in any tissue (Figure 1, Additional file 3). SAGE analyses show also that ALY mRNA levels are higher in a number of tissues, except in thyroid and cortex tissues in which a strong reduction is observed.

Expression of the transcription factor hSPT4 in tumors was similar to those of normal matched tissues in almost all the different organs studied (except vulva, rectum and thyroid).

From the analyses of cDNA arrays we can conclude that there is a differential pattern of expression of THOC1 and ALY in tumor tissues. However, given the number of comparisons made (see Additional file 3), we cannot discard that in some cases the comparison might reach the chosen level of significance by random. Therefore, the analyses of cDNA arrays indicate that THOC1 is differentially expressed in some tumors, both upregulated and downregulated, whereas ALY is overexpressed in a broad range of tumors and only a few significant changes were observed for hSPT4.

It has been described that THOC1 is overexpressed in breast cancer [12]; however, we only detected 3 out of 10 tumor samples with mRNA levels higher than the corresponding normal samples. We decided to study in more detail the expression of THO and other mRNP factors in breast tissues. We evaluated the expression of the components of the human THO complex, THOC1, THOC2, THOC3, THOC5, THOC6 and THOC7, as well as the expression levels of UAP56 and ALY in breast cancer cell lines (Figure 2). mRNA levels were analyzed by real-time-PCR in the malignant breast tumor cell lines MCF7, SKBr3 and T47-D and compared to those of the breast immortalized cell line MCF10-A. An increase in the expression of all components of the THO complex was observed in the breast cancer cell lines, the expression being stronger in SKBr3 and T47D, the most aggressive of the breast tumor cell lines analyzed (Figure 2A). Interestingly, THOC6 mRNA levels were the highest. Although preliminary, these data could indicate that in some cases the stoichiometry of THO is altered and suggest that THOC6 could play additional roles apart from the THO complex [19, 20]. RT-PCR of UAP56 and ALY indicated that these genes were also overexpressed in breast cancer cell lines. The study was extended by performing immunostaining analysis of THOC1 and ALY in normal and tumor breast tissues in order to see if the protein levels of these factors had also increased. Specific antibodies were used to immunostain an array containing 43 breast cancer samples and normal breast controls. A weak signal for THOC1 protein was detected in the cytoplasm of normal ductal epithelial cells and a strong nuclear signal in 80% of the tumor samples analyzed (Figure 2B). Immunohistochemical analysis using anti- ALY revealed a strong expression of this protein in the nucleus of epithelial glands in normal breast tissues. This nuclear signal was observed with at least the same intensity in 70% of the breast tumor samples analyzed (Figure 2B). Nevertheless, given the high expression levels of the ALY protein the sensitivity of this array to detect an increase in ALY expression is limited. These data are in agreement with the overexpression of hHPR1/THOC1 previously reported in breast cancer cell lines [12] and could indicate an up-regulation of the expression of ALY in breast cancer, at least at the mRNA level.

Next, we decided to extend the comparative analysis between THOC1 and ALY, with the analysis by immunostaining of the protein levels and distribution of these factors in tumor samples (Figure 3 and Figure 4). For this purpose, a custom tissue array containing a set of paraffin-embedded samples of different tumor and normal tissues was created. The array contained a representation of tumor specimens of those organs in which, according to the results of the analysis of the cDNA tumor array, THOC1 and ALY could be expected to be deregulated. It was done with a total of 121 samples including approximately 3 normal tissue controls and an average of 9 samples covering low and high-grade tumors for each organ analyzed (for more details see Additional file 1). As can be seen in Figure 3, THOC1 staining revealed that the gene was expressed in all normal tissues with the highest expression in skin and pancreas. The lower signal was detected in stomach, bladder and testis, and the weakest staining was observed in ovary, colon, lung and thyroid. An increase in THOC1 protein signal in the nucleus was detected in ovarian, colon and lung tumors (Figure 3A and 3B). Our results suggest that THOC1 could be relevant in ovarian tumors, in which THOC1 expression levels were high in almost all the carcinomas analyzed (Figure 3A and Figure 1). The increase in the expression of THOC1 protein has been confirmed by western blot in some ovarian and lung normal and tumor samples (Additional file 4). A loss of THOC1 protein staining was observed in testis and skin cancers, in up to 75% of the samples. In these tumors the nuclear signal disappeared and a low immunoreactivity was detected in the cytoplasm (Figure 3). In general, the protein data, although IHC is not a real quantitative method, is in agreement with the results obtained by hybridization of the cDNA array in all these tissues, except for stomach and thyroid tumors that showed a protein signal very similar to that of normal samples.

Tissue microarrays provide the possibility to immunohistochemically stain a large number and variety of normal and cancer tissues, as it is the case of the human protein atlas [17]. We have compared our IHC results with those obtained in the human protein atlas. In the case of THOC1 the comparison does not provide reliable results, because there are two protein atlas IHC analyses (HPA019687 and HPA019096) with different conclusions due to the different protein levels detected in the tissues samples. Whereas in the first few cases of breast cancer, malignant lymphomas and melanomas were considered moderately positive, in the second, this was only the case in few cases of lung, stomach, pancreatic and testicular tumors. On the other hand, the THOC1 overexpression that we report in breast and lung tumors is in agreement with the protein atlas data and other previous studies [12, 15]. In contrast to the atlas IHC data, we observe THOC1 overexpression in ovarian tumors and a decrease in THOC1 expression in testis and skin tumor tissues (Figure 3 and Figure 1). Altogether, these results strengthen the importance of additional studies to establish a reliable association with tumorogenicity and THO expression.

Immunohistochemical analyses were performed with specific antibody anti-ALY, (Figure 4), a strong nuclear signal for the ALY protein being detected in almost all the normal tissues, in contrast to the tissue differential THOC1 immunostaining.

Interestingly, ALY signal was shown to have an equal intensity in half of the tumor samples as in normal tissues (in 46 out of 94 tumors) (Figure 4A and Table 1). The intensity of the immunostaining with anti-ALY was reduced in colon, stomach, thyroid, testis and pancreas tumor samples with a cytoplasmic and lower signal than the control samples in almost 50% of the tumors analyzed (Figure 4B).

When we compare our ALY IHC array with those of the protein atlas IHC data (reports CAB016281 and HPA019799, which show close to 90% of coincidence), it becomes evident that whereas the Protein Atlas shows that ALY is highly expressed in normal tissues and in a wide range of tumors, in our study half of the tumors showed a lower overall signal and an additional cytoplasmic distribution (Figure 4A and Table 1).

A priori, the reduction of the ALY protein contrasts with the up-regulation of ALY mRNA observed after hybridization of the cDNA array (Figure 1). However, when the ALY protein signal was analyzed in the different tumors, it was clear that a significant reduction of ALY protein levels occurs only in the high-grade tumors (a weak or negative staining in 35% of low-grade tumors -20 out of 58- versus 72% -26 out of 36- in high-grade tumors) (Figure 4A and Table 1). This datum suggests that ALY protein levels are down-regulated in the most dedifferentiated phases of the tumor. The ALY protein atlas IHC data shows that only a small proportion of tumor samples (13-23%) had a lower signal than normal tissues and do not provide sufficient data to make conclusions on the relationship between ALY expression and tumor grade.

To gain further insight into the physiological relevance of this reduction, ALY expression was analyzed in different normal tissues with high cell proliferation rate or cells that differentiate into a wide variety of cell types. As is shown in Figure 5, a strong nuclear signal corresponding to the ALY protein was detected in cells of the trophoblast in a normal placenta, in the basal epithelium and lymphoid follicles of normal tonsil and in the germinal cells of testis. A positive signal was observed in dividing cells consistent with the high expression observed in the cDNA array.