Prince Henrys Institute of Medical Research Clayton 3168, Australia
We read with interest the paper by Ligtenberg et al. (1) in which they screened unsuccessfully for mutations in the FSH-receptor gene in ovarian granulosa cell tumors. Several authors have argued that activating mutations of the FSH-receptor might have a role in the pathogenesis of granulosa cell tumors (1, 2, 3). The hypothesis is based in part on the reports of an activating mutation of the LH and TSH receptors in Leydig cell adenomas and thyroid tumors, respectively (1, 2, 3). In our original analysis of the FSH receptor in granulosa cell tumors (2), we had restricted our screen to exon 10, which encodes the transmembrane domains and the intracellular domain. At the time that we conducted the study, activating mutations of G protein-coupled receptors had been reported only in exon 10, predominantly in the 3rd intracellular loop and the 4th transmembrane domain. Kotlar et al. (3) had restricted their search for FSH receptor mutations to this subregion of exon 10. Several recent studies of activating mutations in G protein-coupled receptors have identified activating mutations occurring in the extracellular domain, including mutations of the TSH receptor (4, 5) and the calcium receptor (6). In view of these reports we elected to re-examine the status of the FSH receptor in our panel of granulosa cell tumors. The extracellular domain of the receptor is encoded by nine exons (1). To circumvent the need for nine separate PCR reactions for each tumor (1), we used a coupled reverse transcriptase-polymerase chain reaction (RT-PCR) approach. This has the advantage that only expressed receptor alleles are examined and thus any stromal tissue will not obscure tumor-associated mutations; clearly, for an activating mutation to be relevant it must be expressed. The disadvantage is that archival material cannot be used, and thus the number of tumors available is limited. We examined RNA from six granulosa cell tumors using three overlapping amplicons spanning the extracellular domain.
The 23-mer oligonucleotide primers were at positions 48 and 507, 480 and 898, and 848 and 1257 where the numbers correspond to the published sequence as previously (2). The resulting amplicons were subject to direct sequencing using dye-termination PCR reactions on an ABI 377 Prism automated DNA sequencer (Perkin Elmer, Foster, CA) in the Wellcome Trust Sequencing Centre at Monash Medical Centre (Victoria, Australia).
No FSH receptor mutations were detected in RNA from these 6 granulosa cell tumors. This finding compliments that of Ligtenberg et al. (1), who were similarly unable to find evidence of a mutation in genomic DNA from 23 granulosa cell tumors. They analyzed the tumor DNA for single strand conformation polymorphisms, an approach to mutation detection that while very sensitive, may miss a percentage of mutations. Direct sequencing is arguably the most specific approach to mutation detection, but it may lack sensitivity when searching for heterozygous mutations in DNA derived from a mixed cell population. Our study therefore compliments that of Ligtenberg et al. (1) in that, while it lacks the power of their study (23 tumors), the use of RT-PCR with direct sequencing enhances the sensitivity.
In the absence of mutations of the FSH receptor and indeed of the
associated G-proteins (1, 7), the molecular pathogenesis of ovarian
granulosa cell tumors remains to be explained. It is plausible that a
com ponent of the FSH-receptor signaling pathway downstream
of Gs harbors the postulated activating mutations.
Footnotes
Address correspondence to: Dr. Peter J. Fuller, Prince Henrys Institute of Medical Research, P.O. Box 5152, Clayton 3168, Victoria, Australia. Email: peter.fuller@med.monash.edu.au
Received June 16, 1999.
References