Age-related Thymic Involution in C57BL/6J X DBA/2J Recombinant Inbred Mice Fits A Negative Exponential Regression Model and Maps to Mouse Chromosome 9
Hui-Chen Hsu1, Huang-Ge Zhang1, 2, Lina Li1, Christopher L. Mullinax3, Lu Lu4, Gary Van Zant5, Robert W. Williams4, David B. Allison3, John D. Mountz1, 2
1The University of Alabama at Birmingham, Department of Medicine, Division of Clinical Immunology and Rheumatology, Birmingham, AL 35294
2Veterans Administration Medical Center, 700 South 19th Street, Birmingham, AL 35233
3Section of Statistical Genetics, Department of Biostatistics, The University of Alabama at Birmingham, AL 35294
4Center for Neuroscience and Department of Anatomy and Neurobiology, University of Tennessee, Memphis, TN 38163
5The Markey Cancer Center, Division of Hematology/Oncology, The University of Kentucky Medical Center, Lexington, KY 40536
Thymic involution is the most dramatic and ubiquitous age-associated change in the immune system of humans as well as mice. However, a comprehensive quantitative analysis of initial thymus size and involution rate has not been mathematically quantitated for different genetic backgrounds of mice, thus genetic linkage analysis of thymic involution has not been possible. Here, we have used a mathematical method to analyze the age-related thymic atrophy in C57BL/6 (B6), DBA/2 (D2), and (B6XD2)F1 mice and have observed that, based on the best value of Akaike’s information criterion (Akaike, 1974), thymic involution could be best fit with a negative exponential curve N(t)=b0*exp(-b1t), where N(t) is the total count or number of thymocytes as a function of age in days (t) (Table 1). This regression model was then applied to 18 strains of C57BL/6 X DBA/2 (BXD) recombinant inbred (RI) strains of mice to determine the initial thymocyte count (b0) and thymic involution slope (-b1) of each strain. Thymic involution could be classified into four types in BXD RI strains (Table 2): type A (initially small/rapid involution), type B (initially small/slow involution), type C (initially large/rapid involution), and type D (initially large/slow involution). To ensure counting accuracy, the total thymocyte count was further compared with the thymus weight and thymus size (Fig. 1). A MapManager QTX Program (Manly et al, 2001) was used to determine the quantitative trait loci (QTL) influencing the age-related thymic involution. The strongest QTL for the rate of thymic involution (b1) was found on chromosome (Chr) 9 at 18 cM with D9Mit20 as the best single marker at 61 cM reaching a linkage suggestive likelihood ratio statistics (LRS) value of 10.7 (p=0.001) (Fig. 2). Using D9Mit20 as a control, a second locus maps to the centromeric portion of Chr 14 with D14Nds1 at 2.5 cM reaches the highest LRS of 14.2 (p=0.00017) (Fig. 2). Since day 0 mice are small in size, the thymocyte count of these mice cannot be used to represent the peak thymocyte count. We therefore calculated the day 60 thymocyte count [N(60)] of each strain based on the negative exponential curve. This also provided a fixed age of thymocyte count from each strain of mice for mapping the maximal thymocyte count. The best QTL associated with N(60) thymocyte count mapped at mouse Chr 3, with D3Mit17 (71.8 cM) as the best marker (LRS=14.2, p=0.00017) (Fig. 3). Another locus reaching a linkage suggestive value was DXMit25 (27.8 cM) (LRS=11.8, p=0.00058) (Fig. 3). In summary, the present mathematical regression model for QTL analysis suggests that age-related thymic involution may be regulated by a relatively small number of genes on mouse Chr 9 influencing the rate of thymic atrophy and by genes on mouse Chr 3 influencing the initial thymocyte count.
[This work is supported by NIH grants R01 AG 16653, N01 AR 6-2224, and CA 20408, and a Birmingham VAMC Merit Review Grant. Dr. Huang-Ge Zhang is a recipient of Arthritis Foundation Investigator Award]
Fig. 1. Age-related thymic involution in BXD RI strains of mice can be characterized into four types based on the original thymus size, and the rate of decline in thymocyte count. Total thymocyte count obtained from all strains of BXD RI mice were plotted against the age of each mouse at the indicated age (at least 3 different age group of and 3 mice per age group per strain) using a negative exponential regression model as shown in Table 1.
Based on the values of b0 and b1, the developmental pattern of thymus over the age of mice was characterized into four types. Mouse strains with a -b1 value greater than the mean value of -b1 value are designated as the strains with early and rapid rate of thymic involution. Mouse strains with a b0 greater than the mean value of b0 value are designated as the strains with initial large thymus size. The size, weight, and the total thymocyte count from each representative strain were shown.
Fig. 2. QTLs associated with the slope (-b1 x 104) of thymic involution as a function of age (day) map to mouse chromosomes (Chrs) 9 and 14. (A) LRS plot shows results of the interval mapping of thymic involution rate at mouse D9Mit20 (without control) (upper left), D14Nds1 (controlled with D9Mit20) (upper middle), and D9Mit20 (controlled with D14Nds1)(upper right). The black line is the LRS value and the red line is an estimate of the additive affects of substituting a single b allele for d allele in the test interval. A permutation test was performed to establish criteria for suggestive (left green line, LRS=10.1) and significant linkage (right green line, LRS=15.8). The LRS of the mapped loci and the likely candidate genes nearby the mapped region of D9Mit20 include several genes that are associated with the regulation of cell cycle such as cell division cycle 25 homolog A (Cdc25a), retinal binding protein 2 (Rbp2), and growth arrest and DNA-damage-inducible 45 beta (Gadd45b). Candidate genes at D9Mit20 region that are associated with the regulation of T cells include cytokine inducible SH2-containing protein (Cish), chemokine (C-C) receptor 4 (Cmkbr4), and TRAF-interacting protein (Trip). Candidate genes in the D14Nds1 region include the protein tyrosine phosphatase receptor type G (Ptprg), retinoic acid receptor, beta (Rarb), interleukin 17B receptor (IL-17br) and interleukin 3 receptor alpha chain (IL-3ra).
Fig. 3. QTLs associated with the day 60 thymocyte count [N(60)] map to mouse Chrs 3 and X. (A) LRS plot shows results of the interval mapping of N(60) value at mouse D3Mit17 (upper left) and DXMit25 (upper right). The black line is the LRS value and the red line is an estimate of the additive affects of substituting a single b allele for d allele in the test interval. A permutation test was performed to establish criteria for suggestive (left green line, LRS=10.4) and significant linkage (right green line, LRS=16.4). The LRS of the mapped loci and the likely candidate genes nearby the mapped regions were shown as a table below. Candidate genes at the D3Mit17 region include epidermal growth factor (Egf), lymphoid enhancer binding factor 1 (Lef-1), nuclear factor of kappa light chain gene enhancer in B-cells 1, p105 (Nfkb1), and the cell cycle regulatory Gadd45a. Candidate genes for the locus on Chr X include interleukin 1 receptor-associated kinase (Il1rak), and X-linked lymphocyte-regulated (Xlr)3a, Xlr3b, Xlr4 and Xlr5.
Akaike, H. 1974. A new look at the statistical model identification. IEEE Trans. Autom. Control AC-19:716-723.
Manly, K.F., R.H. Cudmore, Jr., and J.M. Meer. 2001. Map Manager QTX, cross-platform software for genetic mapping. Mamm Genome 12:930-932.