Vol. 178, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

July 31, 1991 Pages 586-592

DEMONSTRATION OF A HETEROGENEOUS TRANSCRIPTION PATTERN

OF THYROGLOBULIN mRNA IN HUMAN THYROID TISSUES

F. Bettaux*, M. Noel*, Y. Malthiery” and P.Fragu*

l iNSERM U66, lnstitut Gustave Roussy, 39 rue Camille Desmoulins, 94805 Villejuif, France

“Laboratoire de Biochimie Medicate, CNRS UA 178 et INSERM U38, faculte de medecine, 27 boulevard Jean Moulin, 13385 Marseille Cedex 5, France

Received May 23, 1991

Previous reports on human thyroglobulin (hTg) modifications in thyroid carcinomas prompted us to study hTg mRNA in thyroid adenomas and carcinomas. The quantifica- tion of hTg mRNA showed a decrease in its levels of expression in both pathological conditions which differed by a factor of 2 between adenomas and carcinomas. Furthermore, PCR was used to analyse the characteristics of hTg mRNA by amplifying 4 regions of the hTg mRNA. When applied to 2 normal, 17 benign and 13 malignant pathological tissue specimens, PCR showed no modification in the size of Tg mRNA. However, abnormal sized cDNAs appeared in all tissues with no distinction between the pathologies ; the Restriction Fragment Length Polymorphism study of these cDNAs suggests the existence of alternate splicing patterns in thyroglobulin mRNAs. 0 1991 Academic Pre**, Inc.

Thyroglobulin (Tg), the main protein produced by the thyroid gland is the

precursor of thyroid hormones (T3 and T4). It is a homodimeric glycoprotein with a Mr of

660 000 daltons, and the corresponding mRNA (8500 nucleotides) has a sedimenta-

tion coefficient of 33s [1,2]. Previous studies have suggested a biochemical difference

in hTg (human thyroglobulin) in differentiated thyroid carcinomas and in normal thyroid tissues : carcinoma hTg has a lower capacity for iodination in vivo [3] and in vitro [4]

and differs from normal hTg in its affinity to polyclonal [4] and monoclonal antibodies [5-

7]. The understanding of the molecular bases of such differences is therefore of inte- rest. Should evidence be found to corroborate these differences then they would be a

useful means of differentiation between normal and neoplastic cold nodules at dia- gnosis.

It is now established that alternate splicing takes place in Grave’s disease for the thyroperoxydase messenger [8]. Differential splicing of Tg mRNA has also been obser-

ved in animal [9-l l] and human [12] thyroid tissues. We thought that the hTg modifica-

tions in differentiated thyroid carcinomas might be the result of abnormal splicing or de-

. . AbbrevlS . hTg , human Thyroglobulin; PCR , polymerase chain reaction; RFLP , restriction fragment length polymorphism.

0006-291X/91 $1.50 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved. 586

Vol. 178, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

letion events in hTg mRNA. In order to test this hypothesis, specific hTg mRNA regions of normal, benign and malignant human thyroid tissues were amplified by polymerase

chain reaction (PCR).

MATERIALS AND METHODS

Tissues. Thyroid operative tissue specimens were obtained from 17 patients with thy- roid adenomas and from 13 with differentiated thyroid cancer (papillary form). Samples were immediately frozen in liquid nitrogen. Normal tissues specimens was obtained from histologically normal perinodular tissue surrounding a solitary hypofunctioning thyroid nodule.

Probes and oligonucleotides. The probes used were the 10 overlapping recombi- nant plasmids which correspond to the total hTg mRNA sequence. These clones label- led Ml, M2, M3, M4, M5, Bl, B2, 83, 84 were inserted in pBR322 or pUC19 as pre- viously described [13]. In northern blot and dot blot experiments, Ml, a 1750 bp cDNA corresponding to the 5’ part of hTg mRNA and M4 a 1400 bp cDNA corresponding to the 3’ part of hTg mRNA were used as specific probes. Six pairs of 20-mer oligonucleotides corresponding to both hTg cDNA strands were used for PCR : CY for the complementary strand of mRNA (CYO = 842-823, CYl = 2064-2045, CY2 = 3203- 31 84, CY3 = 4840-4821, CY4 = 6509-6490, CY5 = 8305-8286) and DY for the direct strand (DYO = l-20, DYl = 843-862, DY2 = 20652084, DY3 = 3204-3223, DY4 = 4841-4860, DY5 = 8305-8286). The corresponding PCR products were designated as shown in figure 2.

Isolation of RNA. RNA was extracted from human tissues and purified by LiCl pre- cipitation according to the method of Chomczynski et al [14].

Quantification of RNA by dot hybridization. Aliquots of total RNA (5 pg) were denatured by glyoxal, spotted on ‘GeneScreen’ membranes (NEN) and hybridized ac- cording to the method of Chomczynski et al 1141. The [32P]cDNA probes were labelled according to the nick translation system (BRL). Quantification was obtained by densi- tometric scanning of the autoradiogram and the transcript level was indicated after standardisation in relation to the level of expression of the 2 normal tissue specimens.

Synthesis of first strand cDNA. [15] Total RNA (10 pg) was mixed, in a total vo- lume of 20 ~1, with a solution containing 400 ng CY primer, 1 mM of each of the four deoxynucleotides, 5 units of RNAsin (Boehringer-Mannheim), 5mM Tris-HCI (pH 8.2) 20 mM NaCI, 2 mM DTT, 1.2 mM MgCl2, and 15 units of AMV reverse transcriptase (Boehringer-Mannheim). The reaction was allowed to proceed for 30 minutes at 42°C and was immediately used as a substrate for PCR.

Polymerase Chain Reaction. The PCR was performed by modifying the method originally described by Saiki et al 1161. cDNA (20 pl) was added to 80 ~1 of a solution containing 400 ng DY primer, 62.5 mM KCI, 12.5 mM Tris-HCI (pH 8.3) 3.125 mM MgC12, 0.1% gelatin (w/v) and 2.5 units of Native Taq polymerase (Cetus). To amplify the cDNA sequences, we performed 30 cycles : 1 min at 92’C to denature the double- stranded DNA, 1 min at 55°C to allow annealing of primers to template and 4 min at 70°C for primer extension. And incubation at 70°C for 10 min to ensure that the final extension step had been completed, after the last cycle.

Southern blot. Amplified DNA products were analysed on agarose gel (1%) and transferred on to nitrocellulose membranes. Hybridization of nitrocellulose membranes was performed in a hybridization solution containing 40% formamide according to the procedure described by Maniatis [17].

587

Vol. 178, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

RESULTS AND DISCUSSION

hTg mRNA quantification on dot blot. The rate of expression of Tg mRNA in the

17 human thyroid adenomas was 36.4% f 20.9% (mean k Sd) of normal thyroid tissue

value, it was more decreased for the 13 carcinomas : 17.9% f 8.9% (mean + Sd) These results confirm and extend the previous study in which hTg mRNA was measured by in

situ hybridization in 7 differentiated thyroid cancers [la]. The decrease in the amount of

hTg mRNA might be due to a shorter half-life of mRNA in thyroid adenomas and carcinomas, but we cannot exclude the possibility that oncogenes might have been

implicated in the decrease in Tg gene expression in human thyroid carcinomas. This

has been observed in previous studies during v-mos oncogene transfection of FRTL5

(Fisher Rat Thyrocyte Line 5) [19-211, in which lTF1 and Tg expression were diminished.

Study of hTg mRNA by PCR. Comparative studies of hTg mRNA by northern blot

(fig.1) did not show significant differences in the size of mRNA derived from tissues of various origins. As the majority of exons present in the Tg gene were very small in size

(loo-150 bp) [22], it is possible that the method failed to detect the absence of a particularly minute exon. A PCR analysis of hTg mRNA sequence organization was

conducted to explore this eventuality. hTg mRNA was divided into 6 regions varying in

length from 842 to 1796 bp (fig.2) but only 4 of these could be amplified by PCR. As PCR is subject to artefacts, amplified fragments were analysed by southern

blots followed by hybridization with hTg probes to verify the specificity of the observed

bands. A comparative analysis of the PCR products from the 4 regions studied, indica-

8.5

28

18

12345678

(lanes FiQ. 1 . Northern blot analysis of Tg mRNA from normal human thyroid

l-2), liver (lane 3), thyroid adenomas (lanes 4-6) and thyroid carcinomas (lanes 7-8). Total RNA was isolated and subjected to northern analysis as described ih section 2.

588

Vol. 178, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

TgmRNA 1 84484

amplifiiation by PCR

, %i2 ~ 2&X , 3203 ~ 48iO , kos ~ &ok

cDNAs W 842 bp :!:2 bp :%9 bp >< %O bp ><

Fig. 2 . Products of hTg mRNA amplified by PCR.

ted that normal cDNAs were present in all pathological tissues studied (figures 3 and

4). These results suggest that there is no ostensible modification of the hTg mRNA due

to the malignant transformation of the cell. However, additional PCR products were

found to hybridize with Tg probes (table 1). These additional cDNAs could have been

the result of either alternate splicing or artefactual events. We chose the specific region

(TgO cDNA) because of the large number of exons which increase the likelihood of ob-

serving alternate splicing events compared to the 3 other regions which either compri-

sed fewer exons (Tgl, Tg2), or in which the exon limits were unknown (Tg4). Moreover,

PCR confirmed the presence of 3 abnormal cDNAs in the TgO region. The products of

TgO amplification yielded 3 cDNAs able to hybridize with Tg probes, measuring 1750,

842 and 530 bp (figure 3), 2 of which (1750 and 530 bp) were in excess of the 842 bp

fragment corresponding to normal hTg mRNA. After migration on agarose gel and

staining with ethidium bromide, each cDNA was separately cryoeluated and

reamplified by PCR. The products of these second amplifications were analysed by

agarose migration. The reamplificetion of the selected bands produced bands of the

same size with the exception of the 1750 bp band which yielded the previously ob-

served band measuring 842 bp and a new band measuring 740 bp (figure 5). This

could be explained as the formation of an heteroduplex. Furthermore, the two cDNAs

measuring 740 and 530 bp were studied by Restriction Fragment Length

1 23456789 1234 5 8 7 8 9 10 11 12

1750 bp

842 bp

530 bp

0 3

Fig. 3 . Southern blot analysis of TgO products generated by PCR ampli- fication of single strand cDNAs derived from human normal thyroid (lane l), thyroid adenomas (lanes 2-5) and thyroid carcinomas (lanes 6-9). The normal cDNAs are noted in bold.

Fig. 4 . Southern blot analysis of Tg2 products generated by PCR ampli- fication of single strand cDNAs derived from human normal thyroid (lane l-2), thyroid adenomas (lanes 3-7) and thyroid carcinomas (lanes 6-12). The normal cDNAs are noted in bold.

589

Vol. 178, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Table 1 sizes of normal (bold) and abnormal hTg cDNAs amplified by PCR

TN Tgl TN Tg4 cDNAs cDNAs cDNAs cDNAs

sizes of the 1750 bp 3100 bp cDNAs 842 bp 1222 bp 1139 bp 1800 bp hybridizing 740 bp 820 bp 840 bp the Tg 530 bp 700 bp probes 400 bp

Polymorphism (RFLP). We chose a single restriction site to exclusively represent each of the 7 exons (figure 6). No restriction cutting by the Bgll enzyme (specific for exon 3)

was observed for the 740 bp cDNA, suggesting a deletion event. No restriction cuttings were observed for the 530 bp cDNA by the Pvull enzyme (specific for exon 4) nor by the

Saul enzyme (specific for exon 6), suggesting deletion events here also. These unequivocally abnormal cDNAs (740 bp and 530 bp) seem to be respectively the

products of alternate splicing of exon 3 and of exon 4 and 6 (figure 6). These deletions concerned moderate sized exons framed by large-sized introns. The relative size of the

1750 bp band of higher cDNA could be explained as the formation of a heteroduplex

composed of an 842 bp strand and a complementary strand measuring 740 bp. Its abnormal migration would result from the non-matching of exon 3 and the subsequent

formation of a hairpin causing a delay in migration.

Fig. 5 . Analysis on agarose 1% gel after staining in ethidium bromide of 1750 bp TgO cDNA (lane 1) and 842 bp TgO cDNA (lane 2). These cDNAs were cryoe- luated and reamplified by PCR before migration on agarose.

Vol. 178, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

740 bp 100 cDNA

530 bp TgO cDNA

DYO CYO

Fig. 6 . Alternate splicing’s hypothesis for the two TgO cDNAs sized 740 and 530 bp.

CONCLUSIONS

These preliminary results show evidence of alternate splicing of Tg mRNA and

confirm previously reported data (1 O-l 2). They also suggest that the Tg population may

be heterogeneous, even in normal thyroid tissue. As deleted regions do not encode ty-

rosine residues, known to be hormonogenic, they do not provide sufficient evidence to

explain the previously observed biochemical modifications of Tg. The question of the diminished expression of hTg mRNA persists and further studies are needed to eluci-

date the causes and consequences of these phenomena.

ACKNOWLEDGMENTS

We would like to thank Mrs,J. Jeusset for technical assistance and Mr.M. Barrois

for helpful discussion. We are grateful to Mrs.Lorna Saint Ange and to Mrs. Ingrid

Kuchenthal for their assistance in revising the manuscript. This study was supported by

a grant from C.N.A.M.T.S. (no 998349/257).

REFERENCES

1. Malthiery, Y., Lissitsky, S., (1987) Eur. J. Biochem., 165, 491-498. 2. Mercken, L., Simons, M.J., Swillens, S., Massaer, M., Vassar? G., (1985) Nature, 316, 647-651. 3. Schneider, A.B., Ikekubo, K. Kuma, K., (1983) J. Clin. Endocrinol. Metab., 57, 1251-l 256. 4. Kim, P.S., Dunn, A.D. Dunn, J.T., (1988) J. Clin. Endocrinol. Metab., 67, 161-168. 5. de Micco, C., Ruf, J., Carayon, P., Chrestian, M.-A., Henry, J.-F., Toga, M., (1987) Cancer, 59, 471-476. 6. Kohno, Y., Tarutani, O., Sakata, S., Nakajima, H., (1985) J. Clin. Endocrinol. Metab., 61, 343-350. 7. Kusakabe, K., Mori, E., Kano, K., Ohta, Y., Fujimoto, Y., Mori, T., (1989) Japan J. Exp. Med., 59, 37-42. 8. Zanelli, E., Henry, M., Charvet, B., Malthiery, Y. (1990) Biochem. Biophys. Res. Comm., 170,735-741. 9. Graves, P.N., Davies, T.F., (1990) Mol. Endocrinology, 4, 155-161.

& Mercken, L., Simons, M.J., Brocas, H., Vassart, G., (1989) Biochimie, 71, 223-

591

Vol. 178, No. 2, 1991 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

11. Ricketts. M.H., Pohl, V., de Martinoff, G., Boyd, C.D., Bester, A.J., Van Jaarsveld, P.P., Vassart G., (1985) EMBO J., 4,731-737. 12. Malthi&y,Y., Charvet, B., Henry, M., (1989) C. R. Sot. Biol., 183, 33-39. 13. Malthiery, Y., Marriq, C., Berg&Lefranc, J.-L., Franc, J.-L., Henry, M., Lejeune, P.- J., Ruf, J., Lissitzky, S., (1989) Biochimie, 71, 195210. 14. Chomczynski P., Sacchi N., (1987) Anal. Biochem. 162, 156-159. 15. Doherty, P.J., Huesca-Contreras, M., Dosch, H.M., Pan, S., (1989) Anal. Biochem., 177, 7-l 0. 16. Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B., Erlich, H.A., (1988) Science, 239, 487-491. 17. Maniatis, T., Fritsch, E.F., Sambrook, J., (1982) in Molecular cloning : a labora- tory manual, Cold Spring Harbor Laboratory. 18. Berg&Lefranc, J.L., Cartouzou, G., de Mica, C., Fragu, P., Lissitzky, S., (1985) Cancer, 56, 345-350. 19. Fusco, A., Portella, G., di Fiore, P.P., Berlingieri, M.T., di Lauro, R., Schneider, A.B., Vecchio, G., (1985) J. Virol., 56,284-292. 20. Avvedimento, V.E., Musti, A., Fusco, A., Bonapace, M.J., di Lauro, R., (1988) Proc. Natl. Acad. Sci. USA, 85, 1744-1748. 21. Berlingieri, M.T., Musti, A.M., Avvedimento, V.E., di Lauro, R., di Fiore, P.P., Fusco, A., (1989) Exp. Cell Res., 183, 277-283. 22. Parma, J., Christophe, D., Pohl, V., Vassar-t, G., (1987) J. Mol. Biol., 196, 769-779.

592