Dear Editor:
A recent report (1) on the histocompatibility within the mouse 129/SvEv strain contains a number of misleading errors and deficiencies.
Rosenberg and coworkers (1) report that two strains of 129/ Sv//Ev mice, one homozygous for a null mutation of the IFN-γ ligand gene (GKO), and the other homozygous for a null mutation of this ligand's receptor gene (RGKO), are histoincompatible in skin grafts with the 129/Sv//EvTac (Taconic Farms, Germantown, NY) strain used as a control. The authors imply that all of these three strains were homozygous for the c allele of the glucose phosphate isomerase-1 structural gene, Gpi-ls (which codes for the C electrophoretic form of the enzyme). For example, they state that the receptor mutation was bred onto the 129/Sv//EvTac strain, which are Gpi-lsc/Gpi-lsc. Therefore, Rosenberg and coworkers suggest that the histoincompatibility observed was probably derived from differences at other chromosomal regions and conclude that it is not possible to compare the effects of mutations made within the 129/Sv//Ev strain, e.g., RGKO vs GKO mice, at least for studies relevant to histocompatible loci.
However, we, who supplied the RGKO strain to the authors, did not breed the receptor mutation onto the 129/Sv//Tac strain. Indeed, RGKO mice, derived from gene targeting using AB.l embryonic stem (ES) cells (2), are Gpi-lsa/Gpi-lsa (our determinations). Also, in deriving the other strain, GKO, while we mated chimeras to 129/Sv//EvTac mice, we subsequently selected only Gpi-lsa/Gpi-lsa animals for propagation. In this case, the Gpi-lsa allele was derived from the AB.l cell clone that carried the targeted mutation of the ligand gene (3). Contrary to a previous report (4), AB.l ES cells are not Gpi-lsc/Gpi-lsc, but Gpi-lsa/Gpi-lsa. At least this is the genotype of the clone just mentioned, and is the genotype of another source of AB.l ES cells obtained directly from the laboratory that derived this cell line (our determinations). Therefore, it would seem quite possible that the histoincompatibility observed was a result of genetic differences between the Gpi-lsc (129/Sv//EvTac) and Gpi-lsa (GKO and RGKO) chromosomal regions, respectively. Indeed, it is this specific difference that is considered to be the major, if not the only, reason why 129/Sv//Ev Gpi-lsc/Gpi-lsc mice reject skin grafts from all other 129 substrains (4). This is all the more likely given that the Gpi-lsc region is derived from wild mice (5), and from the observation that all 129/Sv substrains of the agouti type that are Gpi-lsa/Gpi-lsa are histocompatible despite some genetic differences outside of the Gpi-ls region (4). Therefore, we would predict that GKO and RGKO mice are histocompatible and therefore can be compared with confidence. Unfortunately, Rosenberg and coworkers (1) did not perform grafts between these two lines.
While the authors raise questions as to the feral source of the Gpi-lsc allele, there should be no doubt on this issue. That the allele has been reported to be derived from 101 mice (4), which are in fact Gpi-lsa/Gpi-lsa, is likely a confusion with the source of a modifier of Gpi-ls expression levels, Gpi-lt (where t designates “temporal”). This modifier exists in three allelic forms, as does the structural gene and the c allele present in 101 mice (6).
Another concern of the Rosenberg (1) study is the significance attached to a chronic rejection of a graft between two littermates of control 129/Sv//Tac mice at F16 (after 16 generations of brother × sister matings). Although the authors are technically correct in stating that 129/Sv//EvTac mice did not constitute an inbred strain at F16 (a strain is considered inbred at F20, or ∼99.7% loci fixed), the strain was equivalent to inbred status at this stage. In construction of the 129/Sv//EvTac strain, the parents in the first mating (Fo) were already highly related with at least 90% of loci identical between the two mice, a very conservative estimate based on the reports of Simpson et al. (4) and Threadgill et al. (7). In addition, the Gpi-lsc locus was fixed at F8. Thus, the degree of homozy-gosity achieved at F16 would have been at least 99.9%, exceeding the level of a newly formed inbred strain. The probability then of obtaining a chronic graft rejection due to genetic incompatibility as described would be extremely low, and such a result should be considered with reservation.
- Copyright © 1999 by The American Association of Immunologists, Inc.
References
Genetic Variation Among 129 Substrains: Practical Consequences
Reply.
The data are incontrovertible: the IFN-γ knockout (KO) and IFN-γR KO strains reject skin from 129/SvEvTac mice, and 129/SvEvTac mice reject skin from the KO strains (1). The KO strains thus differ from putative control mice in minor-histocom-patibility Ag expression and therefore cannot be used in a straightforward fashion to assess the effects of the loss of IFN-γ activity on rejection responses to defined histocompatibility loci. None of the comments in Dr. Cantin’s letter significantly impact on this, the critical point of our paper. We do acknowledge and apologize for the misstatement in our manuscript regarding the origin of the IFN-γR KO mice: as corrected by Dr. Cantin, founder animals were not backcrossed to the 129/SvEvTac strain, as were the ligand KOs. This observation does not impact on our findings in any significant way. In fact, this correction indicates that the receptor KO mice are even less genetically compatible with 129/SvEvTac than previously realized. Moreover, our study did not address histocompatibility between the two KO strains. If this is a point of contention, histocompatibility studies can be performed to address the issue.
The source of the genetic disparity or disparities between the KO mice and 129/SvEvTac “controls” is the focus of Dr. Cantin’s letter and was raised in the Discussion of our manuscript. We pointed out that genetic variability of mice derived by injection of 129/SvEvTac blastocysts with embryonic stem (ES) cell line donors could emanate from the ES cell line as well as from the 129/SvEvTac line. Because the AB.l ES line was designated as Gpi-1c by Simpson et al. (2), we thought that the Gpi locus from the ES line did not contribute to the observed alloreactivity, but we raised the point that minor-histocompatibility loci linked to the Gpi locus might be disparate and contribute to rejection responses. These loci are yet uncharacterized in the KO and control strains. Dr. Cantin now states that by his assessment (unpublished data) the AB.l ES line is Gpi-1a, contradicting the finding of Simpson et al. (2). At the time we wrote the manuscript, Dr. Cantin had not assessed the Gpi phenotype of AB.l cells. This issue should be resolved by these investigators.
Regarding genetic variability in the 129/SvEvTac line, we made three points: 1) the ligand KO was established in this strain prior to attainment of homozygosity at the Gpi-1 locus, and thus Gpi differences could serve as a source of alloreactivity (confirmed by Dr. Cantin); 2) 129/SvEvTac differs from all other 129 substrains (except for 129/SvEvGpi-lcHprtb-m2) at the Pgm locus (Pgm-1b), which is not linked to Gpi; thus, this locus and loci closely linked to it that encode retroviral markers (3) might well generate minor-histocompatibility alloantigens recognized by the KO strains; and 3) our finding of a chronic rejection response of a skin graft from one 129/SvEvTac female to another raised the possibility that all genetic loci were not yet fixed in this line at the time these studies were performed, at or prior to F16. We are puzzled by Dr. Cantin’s contention that this strain was inbred at F16. Neither data nor a reference were provided that support this assertion.
Last, we are in agreement with Dr. Cantin regarding inconsistencies in the Simpson et al. manuscript, which we previously pointed out (1). Clarification is required for the finding that the AB.l ES line as typed by Simpson was Gpi-1c despite its derivation from a Gpi-1a bearing founder line (2). Furthermore, Simpson et al. attributed the origin of Gpi-1c to a mating between 101 strain mice with 129 strain progenitors. As correctly stated by Dr. Cantin, 101 strain mice are Gpi-1a (4), and thus the Gpi-1c contribution is likely from a feral source.
- Copyright © 1999 by The American Association of Immunologists, Inc.
In this issue
