Novel Splicing of an IGF2 Polymorphic Region in Human Adrenocortical Carcinomas
Steven G. Gray,* Magnus Kjellman, t # Catharina Larsson, t and Tomas J. Ekström*,1
* Laboratory for Molecular Development and Tumour Biology, Experimental Alcohol and Drug Addiction Research Section, Department of Clinical Neuroscience, Karolinska Institute, CMM, L8 01, S-171 76 Stockholm, Sweden; t Department of Molecular Medicine, Karolinska Hospital, CMM, L8 01, S-171 76 Stockholm, Sweden; and }Department of Surgery, Karolinska Hospital, S-171 76 Stockholm, Sweden
Received September 24, 1997
The human IGF2 gene lies on chromosome 11p15.5 and encodes for a mitogenic peptide. IGF2 is often overexpressed in many tumours including adrenal carcinomas. In this study while screening 12 adreno- cortical carcinomas for heterozygosity at the Apa I and (CA)„ repeat polymorphisms we observed a novel splic- ing event in two samples which showed both an allelic expression imbalance and preferential splicing for one of the alleles. Further examination revealed that the splicing was not confined to one particular site. Three of such splice products were isolated and cloned. Us- ing RNase protection analysis the presence of this splicing event was demonstrated for both adrenocorti- cal carcinoma samples and also in a Hep3B cell line. This suggested that the event may be occurring in all the samples. The presence of this splicing was then confirmed in all 12 adrenocortical carcinoma samples by PCR. These data suggest that the splicing event may be a general feature for IGF2 transcripts. @ 1997 Academic Press
Human Insulin-like growth factor 2 (IGF2) encodes for a mitogenic peptide, lies on chromosome 11p15.5, and is transcribed in a tissue and developmental spe- cific fashion (1). Overexpression of IGF2 has been ob- served in several tumours including Wilm’s Tumour (2,3), colon carcinoma (4), liposarcoma (4), liver cancer (5,6), ovarian cancer (7), hepatoblastoma (8), gastroin- testinal tumours (9), and renal cell carcinomas (10). High expression of IGF2 has also been reported for various adrenal tumours (11-15), but little has been
1 To whom correspondence should be addressed. Fax: +46 8 517 74615. E-mail: Tomas.Ekstrom@cmm.ki.se.
The abbreviations used are: IGF2, insulin-like growth factor 2: RPA, RNase Protection analysis: PCR, polymerase chain reaction.
done to investigate the imprinting status and promoter usage. The human IGF2 gene comprises 9 exons span- ning a region of about thirty kilobases. Expression of IGF2 is driven in a tissue and developmental fashion by four promoters (P1-P4). Only exons 7, 8 and the first 237 nucleotides of exon 9 are translated, the remaining exons being utilized by the four promoters to provide distinct 5’- untranslated regions (leaders) (16). We set out to investigate IGF2 expression in a set of adrenocor- tical carcinomas. Here we report the discovery of an allele preferential splicing event within exon 9, the splice product of which is present in all the samples.
MATERIALS AND METHODS
Tumour samples. This study includes 12 primary sporadic adre- nocortical carcinomas. The tumours were classified as malignant based on histopathological findings, tumour size and clinical out- come.
Nucleic acid isolation, PCR amplification, and analysis of IGF2 polymorphic regions. Genomic DNA was prepared by repeated phe- nol/chloroform extractions of tissue cells lysed with 0.5% SDS and proteinase K (200 µg/ml final concentration). Total RNA was pre- pared as described previously (17), and tested for the Apa I and (CA)n polymorphisms according to published methods (18,19). The following primers were used for PCR amplification:
Apa I; Primer 2 5’-CTTGGACTTTGAGTCAAATTGG-3’. Primer 3 5’-GGTCGTGCCAATTACATTTCA-3’. (CA)n; Primer 8 5’-GAGTATGAAATACGTAGGGGC-3’.
Primer 9 5’-GCCTGATCCATACAGATATCG-3’.
Amplification conditions for the Apa I and (CA)n specific amplifica- tions were as previously published (18,19). The conditions for the amplification of the IGF2 splice were as follows: approximately 100ng of template (or 1 ul of cDNA as appropriate) and 0.48pM of each primer in the presence of 1.5mM MgCl2, 0.2mM dNTPs and 1unit of Taq DNA Polymerase with supplied buffer in a total volume of 50ul. Cycling conditions were 95 ℃ for 5 minutes followed by 3 cycles of (1 minute at 94 ℃, 1 minute at 62 ℃, 1 minute at 72 ℃) followed by 32 cycles of (1 minute at 94 ℃, 1 minute at 60 ℃, 1 minute at
72 °℃) with a final extension at 72 ℃ for 10 minutes. To ensure complete digestion occurred with Apa I, the PCR products were spiked with 100 ng of uncut pBluescript SK II- (Stratagene) prior to digestion.
Generation of cDNA by reverse transcription with MuMLV RT-ase. The primers used to generate cDNA in this study were:
| Primer 10 | 5'-GCATCTCTGTCATGGTGGAAAG-3'. |
| Primer 11 | 5'-GTAAGGTGTATCGGGAATG-3'. |
| Primer 12 | 5'-CCTTCCAGGAGCACACCA-3'. |
Generation of cDNA from RNA was carried out as follows: 10µg of total RNA was digested with RQ1 DNase (Promega) for 1.5 hours at 37 ℃, phenol extracted, and ethanol precipitated. The RNA was redissolved in 15ul of sterile water, and 5ul removed for use as a negative DNA control in the PCR amplifications after cDNA genera- tion. Primer was added to the remainder of the RNA at a final concen- tration of 0.75 pM, the sample was then heated to 85 ℃ for 5 minutes, cooled rapidly on ice and then reverse transcribed with MMuLV- RTase according to the manufacturers instructions (Life Technolo- gies).
Generation of cDNA by reverse transcription with rTth polymerase. First strand cDNA synthesis using rTth polymerase was carried out
b
123 bp Ladder
a/b
a/a
b/b
b/b
Uncut
a/b
Cut plasmid
Uncut plasmid
a
+ Uncut PCR
Apal
2
13
8
9
10
11
12
Cut PCR
(CA)n
Undigested
123 bp ladder
Tumour 5
Tumour 8
d
Probe
Tumour 5
Tumour 5
Tumour 8
Tumour 8
C
Allele 1
Bluescript
C
Allele 2
PCR →
GCGC
DNA RNA DNA RNA
according to the manufacturers instructions (Perkin Elmer) except that all volumes were doubled. The primer used was primer 12 (see above). A “Hot-Start” procedure was used whereby the sample and primer were incubated at 70 ℃ for 5 minutes and then the tempera- ture was dropped to 62 ℃ for a further 5 minutes, before the tran- scriptase mix was added. Reverse transcription was then allowed to proceed at 62 ℃ for a further 30 minutes.
Isolation, cloning, and sequencing. The splice product was run out in 1% agarose and isolated using a QIAEX II Gel Extraction Kit (Qiagen). The PCR product was blunted using T4 DNA Polymerase (Life Technologies) and cloned into EcoRV (Promega) digested Blue- script SKII- (Stratagene). Sequencing was carried out with T7 Sequenase according to manufacturers instructions (United States Biochemical).
Preparation of probe and RNase protection analysis. T3 and T7 RNA polymerases were purchased (Life Technologies) and RNA probes were prepared from the cloned splice according to the protocol provided in the RPA II Kit (Ambion). When incorporating radioactiv- ity into the probe, radioactive 32P-UTP with a specific activity of 800uCi/mmol was used. When generating cold probe equal amounts (2.5mM) of all cold rNTPs were added. RNase protection was carried out according to the protocol given with the RPA II Kit (Ambion).
RESULTS AND DISCUSSION
Polymorphism analysis of IGF2 in adrenal car- cinomas. Twelve adrenocortical carcinomas were screened for heterozygosity using the Apa I and (CA)n repeat polymorphisms contained within the 3’- un- translated region of exon 9 (Fig.1a). When the Apa I polymorphism was examined a surprisingly high pro- portion of the samples, 8 out of 12 samples were hetero- zygous and showed biallelic expression (for details see Table 2). An example of the typical digestion pattern observed for all three allelic possibilities is shown (Fig. 1b). However, when the (CA)n repeat polymorphism was used to examine informity at the DNA and RNA level, an apparent anomaly was observed for 2 samples. In these samples, if Apa I was used to examine expres- sion from cDNA, expression was biallelic (Fig.1c). The expression from both alleles is significantly skewed with most of the expression coming from the allele con- taining the Apa I site. Conversely, if the (CA)n repeat polymorphism was used, only monoallelic expression, was observed (Fig. 1d). This was an unexpected result as a small but significant expression from the lowly expressed allele would have been expected because of the Apa I result. This anomaly was also observed if the (CA), repeat was amplified from cDNA and then tested.2 No detectable P1 expression was observed for these samples,3 indicating that the IGF2 expression was being directed from promoters P2, P3 and P4. Us- ing exon specific RNase protection analysis (RPA) this was shown to be the case, with highest levels of expres- sion coming from promoter P3.4
2 S. G. Gray and T. J. Ekström, unpublished data.
3 S. G. Gray and T. J. Ekström, unpublished data.
4 S. G. Gray and T. J. Ekström, unpublished data.
| Analysis of allele frequency | |||
|---|---|---|---|
| Samples | Allele types | ||
| A/A | A/B | B/B | |
| Normal | 17 | 13 | 1 |
| Tumour | 9 | 7 | 1 |
| Overall | 26 | 20 | 2 |
Note. Examination of the Apa I polymorphism was carried out as described in Materials and Methods. Blood samples from 31 normal individuals and 17 individuals who developed adrenocortical carcino- mas were examined to assess the Apa I allele frequency. The relative frequency of heterozygosity within the population is 20/48 = 42%.
Examination of the Apa I polymorphism within the swedish population. Because the level of heterozygos- ity for the Apa I polymorphism was so high, it was decided to examine the allele frequency within the pop- ulation to see if this was significant. The Apa I polymor- phism was examined using blood DNA from 31 normal individuals, and from 17 individuals who developed adrenocortical carcinomas. The overall results for each group and when combined as a whole are shown in Table1. Essentially, the original finding that eight of the twelve (67%) tumour samples were heterozygous for the Apa I polymorphism is insignificant, because the overall level of heterozygosity drops to 42% when all the samples are considered.
Demonstration of novel splicing around the IGF2 (CA)n region. To explain the observed discrepancy be- tween the two polymorphisms (i.e., biallelic versus mo- noallelic expression in the two samples), we hypothe- sized that the (CA), repeat from one allele was being preferentially spliced out of its RNA. To examine this possibility, cDNA was subjected to PCR amplification using primers 2 and 10. The expected product from such an amplification should cover the region encom- passing both polymorphisms (Fig. 1a). If the (CA), re- peat was being spliced out, a smaller amplification product in addition to the full length amplification product would be expected (Fig. 1a). Both amplification products were observed suggesting that the (CA)n re- peat was being spliced out of the RNA (Fig. 2a). The PCR products were then digested with Apa I. Surpris- ingly, both products digested indicating that both al- leles are spliced (Fig. 2b).
Cloning and detection of the splice in adrenal carcino- mas and Hep3B cells. The smaller amplification prod- uct was then cloned and sequenced. The data obtained revealed the presence of a region within the splice with a high homology to that of the primer used to generate
123 bp ladder
a
Genomic
CDNA
Neg
b
Genomic
CDNA
SK -
SK
Unspliced 1339 bp
Unspliced
Spliced 193 bp
Spliced
+
+
a
b
Hep3B
Tumour 2
Tumour 8
Control
Splice 2-10
TCCTCACTCC
CTTTCCACCATGACAGAGATGC
Probe →
+993
+2072
Splice 2-11 a
CTTACATCTT
GTAAGGGCTATGTGGAATG
Spliced
+1071
+2171
Splice 2-11 b
Unspliced
TACTTTATGC
GTAAGGGCTATGTGGAATG
+1104
+2171
| Sample | Apa I polymorphism genomic | cDNA | (CA)„ polymorphism genomic | cDNA | 2-10 splice cDNA | |
|---|---|---|---|---|---|---|
| 1 | Uninformative (*) | A/A | — | Uninformative | — | Present |
| 2 | Informative | A/B | Biallelic | Uninformative | — | Present |
| 3 | Uninformative (*) | A/A | — | Uninformative | — | Present |
| 4 | Informative | A/B | Biallelic | Uninformative | — | Present |
| 5 | Informative | A/B | Biallelic | Informative | monoallelic | Present |
| 6 | Informative | A/B | Biallelic | Uninformative | — | Present |
| 7 | Informative | A/B | Biallelic | Uninformative | — | Present |
| 8 | Informative | A/B | Biallelic | Informative | monoallelic | Present |
| 9 | Informative | A/B | Biallelic | Uninformative | — | Present |
| 10 | Uninformative | B/B | — | Uninformative | — | Present |
| 11 | Uninformative (*) | A/A | — | Uninformative | — | Present |
| 12 | Informative | A/B | Biallelic | Uninformative | — | Present |
Note. Twelve adrenocortical carcinomas were tested for the Apal and (CA), repeat polymorphisms at the Genomic and RNA level. The ApaI polymorphism was examined by PCR amplification of the relevant region (genomic or cDNA) using primer set 2, 13, followed by restriction analysis. The completion of digestion was verified by including an internal digestion control in each sample. Unless indicated (*), the results for the Apal polymorphism are for at least three separate experiments. A, indicates an allele containing the Apal site, B, indicates an allele which does not contain the Apal recognition site. To examine the (CA), polymorphism, PCR amplification of the relevant region from genomic DNA was prepared using primer set 8, 9, followed by RPA analysis on both the PCR product and on RNA using a (CA)n specific probe. The presence of the 2-10 splice was determined from cDNA by PCR amplification using primer set 2, 10.
the cDNA. In order to determine whether or not the smaller PCR product was due to mis-priming during cDNA generation a primer further downstream, primer 11, was designed and used to make new cDNA (Fig. 1a). This cDNA was subjected to the same amplification strategy as before. Upon reamplification two products of the expected size were observed. The smaller ampli- fication product was isolated, cloned and sequenced. The same sequence was obtained as before indicating that the smaller splice product was not an artifact of the cDNA generation process but a genuine splice per se. When primers 2 and 11 were used for amplification, two other splice products were obtained. These were isolated cloned and sequenced. The sequences of the IGF2 splice products are shown (Fig. 3a). A feature of these splices is that they all splice directly to the downstream primer, which might suggest that second- ary structure within the mRNA transcript is interfer- ing with cDNA generation. In order to reduce this pos- sibility, first strand cDNA was generated with primer 12 (Fig. 1a), at 62 ℃ using a thermostable recombinant polymerase rTth (20). PCR amplification using both of the previous primer sets on this cDNA gave products of the expected size for each primer set. When isolated, cloned and sequenced two of the prior splice products were obtained indicating that these were correctly spliced and not artifacts due to secondary structure. We have observed both the Apa I restriction site and the C-T transition in the cloned splice from primer set 2 and 105, confirming that both alleles are being
spliced. The clone containing the 2-10 splice was then used as a probe to test for the presence of the splice product in total adrenocortical carcinoma RNA by RNase protection analysis. The samples showed the presence of the fully protected splice product. In addi- tion, the splice product was also observed in Hep3B cells (Fig. 3b). This raised the possibility that the splic- ing event may in fact be a general phenomenom. All 12 adrenocortical carcinomas were then tested for the presence of this splice by PCR amplification using primers 2 and 10 on cDNA. Surprisingly, the splice was observed to be present in all 12 adrenocortical carci- noma samples. The overall results of these investiga- tions are presented in Table 2.
The functional significance of this splicing is not known. The fact that it occurs in all samples indicates that it may have a strong functional significance. Of interest is the fact that splicing occurs in both alleles, but with a preference for one allele in particular. The low expression allele is completely spliced out for the (CA), repeat region. We have been unable to detect any unspliced transcripts from this allele either by RPA using (CA)n, or by PCR of cDNA using Primers 2 and 10. The highly expressed allele shows both forms. Splic- ing is occurring but only to a proportionally small ex- tent based on the RPA analysis shown in Fig. 3b. It is interesting to note that an allelic bias for the Apa I polymorphism has also been observed for IGF2 not only in breast cancer (21), but also in blood leukocytes (22) raising the possibility that this splicing event may also be occurring under normal conditions . Four factors which point towards the existence of splicing around
5 S. G. Gray and T. J. Ekström, unpublished data.
the (CA), region, and discount the possibility of cDNA artifacts are as follows. Firstly, RPA analysis of total RNA using the (CA), repeat region showed monoallelic expression in two samples which had been shown to have biallelic expression using the Apa I polymor- phism. For the second factor, studies have shown that reverse transcriptases are inhibited by secondary structure, but can be alleviated by preheating the sam- ple RNA prior to reverse transcription (23). Such a step was included in our first strand cDNA synthesis using MMuLV-RTase. The third factor is that by using the thermostable rTth polymerase as a reverse tran- scriptase we were able to produce first strand cDNA at high temperatures. And the fourth, is that using one of the cloned splice products for RPA analysis a fully protected band corresponding to the full length splice product appears.
The position of splicing within exon 9 lies between two regions implicated in the turnover of IGF2 mRNA. However, it has been demonstrated that the (CA)n re- gion can be deleted without affecting the rate of endo- nucleolysis (24), which indicates that the splicing event is not an attempt by the cell to downregulate IGF2 by making the RNA more unstable. Efforts are currently underway to examine the function of the splicing event, and whether or not it is cell or tissue specific.
ACKNOWLEDGMENTS
We thank Professor Rolf Ohlsson for meaningful discussions. This work was supported by The Swedish Cancer Foundation, The Chil- dren’s Cancer Foundation of Sweden and The Swedish Natural Sci- ence Research Council.
REFERENCES
1. Franklin, G. C., Adam, G. I., and Ohlsson, R. (1996) Placenta 17(1), 3-14.
2. Scott, J., Cowell, J., Robertson, M. E., Priestley, L. M., Wadey, R., Hopkins, B., Pritchard, J., Bell, G. I., Rall, L. B., Graham, C. F., et al. (1985) Nature 317(6034), 260-262.
3. Reeve, A. E., Eccles, M. R., Wilkins, R. J., Bell, G. I., and Millow, L. J. (1985) Nature 317(6034), 258-260.
4. Tricoli, J. V., Rall, L. B., Karakousis, C. P., Herrera, L., Petrelli, N. J., Bell, G. I., and Shows, T. B. (1986) Cancer Research 46(12 Pt 1), 6169-6173.
5. Cariani, E., Lasserre, C., Seurin, D., Hamelin, B., Kemeny, F., Franco, D., Czech, M. P., Ullrich, A., and Brechot, C. (1988) Can- cer Research 48(23), 6844-6849.
6. Cariani, E., Seurin, D., Lasserre, C., Franco, D., Binoux, M., and Brechot, C. (1990) Journal of Hepatology 11(2), 226-231.
7. Yun, K., Fukumoto, M., and Jinno, Y. (1996) American Journal of Pathology 148(4), 1081-1087.
8. Li, X., Adam, G., Cui, H., Sandstedt, B., Ohlsson, R., and Ek- strom, T. J. (1995) Oncogene 11(2), 221-229.
9. Wang, S. N., Souza, R. F., Kong, D. H., Yin, J., Smolinski, K. N., Zou, T. T., Frank, T., Young, J., Flanders, K. C., Sugimura, H., Abraham, J. M., and Meltzer, S. J. (1997) Cancer Research 57(13), 2543-2546.
10. Nonomura, N., Nishimura, K., Miki, T., Kanno, N., Kojima, Y., Yokoyama, M., and Okuyama, A. (1997) Cancer Research 57(13), 2575-2577.
11. Gicquel, C., Bertagna, X., Schneid, H., Francillard-Leblond, M., Luton, J. P., Girard, F., and Le Bouc, Y. (1994) Journal of Clini- cal Endocrinology & Metabolism 78(6), 1444-1453.
12. Liu, J., Kahri, A. I., Heikkila, P., Ilvesmaki, V., and Voutilainen, R. (1995) Journal of Clinical Endocrinology & Metabolism 80(2), 492-496.
13. Voutilainen, R., and Miller, W. L. (1987) Proceedings of the Na- tional Academy of Sciences of the United States of America 84(6), 1590-1594.
14. Voutilainen, R., and Miller, W. L. (1988) Dna 7(1), 9-15.
15. Haselbacher, G. K., Irminger, J. C., Zapf, J., Ziegler, W. H., and Humbel, R. E. (1987) Proceedings of the National Academy of Sciences of the United States of America 84(4), 1104-1106.
16. Sussenbach, J. S. (1989) Progress in Growth Factor Research 1(1), 33-48.
17. Chomczynski, P., and Sacchi, N. (1987) Analytical Biochemistry 162(1), 156-159.
18. Tadokoro, K., Fujii, H., Inoue, T., and Yamada, M. (1991) Nucleic Acids Research 19(24), 6967.
19. Ohlsson, R., Nystrom, A., Pfeifer-Ohlsson, S., Tohonen, V., Hed- borg, F., Schofield, P., Flam, F., and Ekstrom, T. J. (1993) Nature Genetics 4(1), 94-97.
20. Myers, T. W., and Gelfand, D. H. (1991) Biochemistry 30(31), 7661-7666.
21. McCann, A. H., Miller, N., A, O. M., Pedersen, I., Keogh, K., Gorey, T., and Dervan, P. A. (1996) Human Molecular Genetics 5(8), 1123-1127.
22. Vafiadis, P., Bennett, S. T., Colle, E., Grabs, R., Goodyer, C. G., and Polychronakos, C. (1996) Journal of Autoimmunity 9(3), 397-403.
23. Brooks, E. M., Sheflin, L. G., and Spaulding, S. W. (1995) Bio- techniques 19(5), 806-12, 814-815.
24. Meinsma, D., Scheper, W., Holthuizen, P. E., Van den Brande, J. L., and Sussenbach, J. S. (1992) Nucleic Acids Research 20(19), 5003-5009.