Editorial Commentary: The Sa Gene

• Editorial
• hypertension, genetic
• genes
• congenic strains
• rats, inbred SHR

The Sa gene was first described in 1991 by Iwai and Inagami¹ in a study to identify genes that were differentially expressed in the kidneys of spontaneously hypertensive rats (SHR) and Wistar-Kyoto rats (WKY). To find such genes they first prepared a cDNA library from 16-week-old SHR. The clones were grown on plates, and 4 replica filter lifts were prepared from each plate. Two filters were hybridized to ³²P-labeled single-strand antisense cDNA fragments from kidneys of 16-week-old SHR, and 2 filters were hybridized with a similar probe prepared from WKY rat kidneys. Clones that gave a different intensity of autoradiographic signal with the SHR and WKY probes were selected for study; one of these was the Sa gene. The designation Sa is apparently arbitrary.

Iwai and Inagami¹ confirmed that the Sa gene was differentially expressed in SHR and WKY kidneys by Northern blot analysis, with SHR expressing markedly more than WKY. They also showed that Dahl salt-sensitive (S) rats expressed the Sa gene more than Dahl salt-resistant (R) rats.¹

Although the technique of differential hybridization is a standard one, its application to hypertensive rat models was innovative, and the Sa gene has generated a lot of interest. Important subsequent results were that (1) the Sa gene is expressed in the kidney proximal tubule,^{2 3} (2) the Sa gene is located on rat chromosome 1 and cosegregates with blood pressure (BP) in both the SHR/WKY^{4 5 6 7} and Dahl rat models,^{8 9 10} and (3) polymorphisms in the Sa gene and differential expression of the Sa gene cosegregate.⁷

The linkage of a candidate gene such as the Sa gene to BP certainly does not prove that the candidate is the causative factor. Such linkage is necessary but not sufficient. Linkage analysis for a quantitative trait such as BP in theory^{11 12} and in practice yields only a very broad segment of chromosome that is likely to include the causative gene (or genes). This region is typically 20 to 40 centimorgans and contains hundreds of genes. To be sure that in fact a BP quantitative trait locus (QTL) exists in the region of interest, the next logical step is the construction of congenic strains.^{13 14} The congenic strain also provides a vehicle for subsequent fine genetic mapping of the QTL.

In forming a congenic strain, a segment of chromosome is moved from one inbred rat strain (donor) to another strain (recipient) by a standard protocol involving a series of back-cross breeding to the recipient strain with selection for donor alleles at marker loci along the chromosomal segment to be moved.¹³ This results in a pair of strains (the congenic and the recipient strain) that are (essentially) genetically identical except that the congenic strain has a defined chromosomal segment from the donor. If the congenic region in fact contains a variant BP QTL allele compared with the recipient strain, then the 2 strains will have different BP, demonstrating the existence of a BP effect of the chromosomal region that was moved.

In this issue of Hypertension, Iwai et al¹⁵ and Frantz et al¹⁶ report the construction of congenic strains involving SHR and WKY chromosomal regions containing the Sa gene. In both cases, the congenic strains demonstrated the existence of a BP QTL in the general vicinity of the Sa gene. This work confirms and extends the results from a previous congenic strain on rat chromosome 1 using SHR and WKY¹⁷ ; an additional report on a chromosome 1 congenic strain also demonstrates a BP QTL using Dahl S and Lewis rats.¹⁰ Thus, there is an impressive consistency in these results.

Although the function of the Sa gene product is unknown, the Sa gene has passed several important tests as a hypertension candidate gene: (1) it has an interesting intermediate phenotype (differential expression between strains in the proximal tubule); (2) it cosegregates with BP; and (3) it is contained in a congenic region influencing BP. An important remaining step is to make congenic strains in which the alleles at the Sa locus are moved with reduced (preferably minimal) flanking DNA to reduce the chance that the real QTL is a locus linked to the Sa locus. This procedure has been applied to the loci for renin and the inducible form of nitric oxide synthase; in both cases, the candidate gene was eliminated as causative for the BP difference between the specific strains used.^{18 19} Thus, some caution is appropriate until fine genetic mapping of the QTL in the region of the Sa gene is accomplished by construction of congenic substrains.

Other data suggest that the Sa gene is not responsible for BP differences. The Milan hypertensive strain (MHS) carries the same Sa allele as the normotensive WKY, and the Milan normotensive strain (MNS) carries the same allele as SHR.²⁰ This conclusion was based on the restriction fragment length patterns of 15 enzymes that yield a polymorphic pattern between the 2 alleles. Also, MNS express a high level of renal Sa mRNA compared with MHS, as would be expected based on the allele in each strain. This reversal of alleles compared with what might be anticipated based on the SHR/WKY and Dahl S/Dahl R comparison in itself has no bearing on interpretation of the linkage and congenic strain data. It is expected theoretically, and found experimentally, that some QTL alleles for increase of a quantitative trait will be carried by a low strain and visa versa; some examples are found in work by Garrett et al.¹⁰ A high or low strain is determined by the net effect over multiple loci, so alleles at any single locus do not determine the ultimate phenotype. Much more insightful is the observation²⁰ in an F₂ population derived from MHS and MNS that alleles at the Sa locus did not cosegregate with BP. This does imply that the Sa alleles that control the large difference in renal expression have nothing to do with BP and that the QTL on chromosome 1 does not have functionally variant alleles in the MHS versus MNS comparison. One could also argue (less convincingly in my view) that ancillary factors (epistatic interaction with the genetic background, environmental factors such as normal dietary NaCl) blunted the response of BP to the Sa alleles in the F₂ (MHS×MNS) study.

Animal work with the Sa gene has stimulated work in humans; the Sa gene is located on human chromosome 16p13.11.^{21 22} In 1 study,²³ there was a difference in allelic frequency of Sa alleles between hypertensive and control groups. In 3 other such association studies,^{24 25 26} no relationship was found; linkage analysis in humans was also negative.²⁶ Thus, the studies on humans fail to provide convincing evidence for an effect of the Sa locus on BP.

In summary, the work of Iwai and Inagami¹ and the subsequent work it generated is science at its best, regardless of the ultimate status of the Sa gene as one of “the” genes causing BP variation. At a very minimum, the Sa gene as a candidate led to the detection and confirmation of a BP QTL on rat chromosome 1. Fine genetic mapping in the animal models using derivatives of the new congenic strains will certainly determine the status of the Sa gene and provide a better QTL localization on rat chromosome 1 for guiding analyses on human chromosome 16.