CD4+CD25+Foxp3+ regulatory T cells (Tregs) play a central role in cancer tolerance. However, mechanisms leading to their accumulation in cancer remain unknown. Although the thymus is the main site of Treg development, thymic contribution to Treg expansion in cancer has not been directly examined. Herein, we used two murine models of multiple myeloma (MM), 5T2 MM and 5T33 MM, to examine Treg accumulation in peripheral lymphoid organs, including spleen, lymph nodes, bone marrow, and blood, and to explore thymic Treg development during malignancy. We found that peripheral ratios of suppressive-functional Tregs increased in both models of MM-inflicted mice. We found that thymic ratios of Treg development in MM increased, in strong association with thymus atrophy and altered developmental processes in the thymus. The CD4+CD8+ double-positive population, normally the largest thymocyte subset, is significantly decreased, whereas the CD4CD8 double-negative population is increased. Administration of thymocytes from MM-inflicted mice compared with control thymocytes resulted in increased progression of the disease, and this effect was shown to be mediated by Tregs in the thymus of MM-inflicted mice. Our data suggest that increased ratios of Treg development in the thymus may contribute to disease progression in MM-inflicted mice.

Interest in the involvement of CD4+CD25+Foxp3+ regulatory T cells (Tregs)3 in cancer tolerance has been continuously growing in recent years. Elevated Treg frequency has been reported in various mouse and human malignancies (1, 2, 3, 4, 5), and their functional role in reducing antitumor response has been demonstrated in mice (6, 7, 8, 9). In humans, Treg contribution to tumor tolerance was strongly suggested by the significant correlation between Treg levels and poor survival of ovarian cancer patients (2). Treg levels have been also associated with poor survival and tumor relapse in other malignancies, such as breast cancer and non-small cell lung cancer (10, 11).

It is essential to reveal the mechanisms leading to Treg expansion for the development of strategies to eliminate them and improve the results of cancer immunotherapy (12). Several mechanisms describing Treg induction or recruitment to the tumor site have been described. It is generally thought that Tregs are induced at the tumor site and further affect tumor microenvironment and draining lymph nodes (LNs) (12, 13). Indeed, it was recently shown that Tregs were induced at the tumor site, as a result of IL-10 and TGF-β secretion by the tumor cells (14, 15, 16). Additionally, “naturally occurring” Tregs (nTregs) were shown to be specifically recruited to the tumor by chemotaxis, which was mediated by the release of CCL22 and CCL17 by the tumor cells (2, 17). Notably, the thymus is recognized as the main site of Treg development (18, 19, 20, 21). Treg development in the thymus has been discussed as a possible mechanism contributing to Treg accumulation in malignancy (1), and thymus output was indirectly tested in human multiple myeloma (MM) patients (22). However, Treg development in the thymus during malignancy has not been directly explored.

In contrast to data from solid tumors and other hematological tumors (1), several studies in B cell malignancies have reported an association of lower Treg levels with poor prognosis (23, 24). This phenomenon was suggested to reflect the suppressive function of Tregs on B cell proliferation and activity (24). However, other studies in B cell malignancies reported a correlation between higher Treg levels and disease morbidity (22, 25). Specifically, two recent studies in MM patients reported contradictory findings concerning Treg levels and activity (22, 26). In one, reduced Treg levels were reported as well as Treg dysfunction (26). In the other, increased Treg levels in the blood and bone marrow (BM) of MM patients were observed, and Treg suppressive function was retained (22). Therefore, contradictory data exist concerning Treg levels and activity in B cell malignancies as a whole, and specifically in MM.

In the present study we used two mouse models of MM, 5T2 MM (5T2) and 5T33 MM (5T33) (27, 28), to examine Treg levels in the periphery, as well as developmental processes that may occur in the thymus to increase Treg ratios in MM. We found increased Treg proportions in the spleen, LNs, BM, and peripheral blood (PB) of MM-inflicted mice. Additionally, thymus atrophy was associated with increased Treg-to-effector T cell (Teff) proportions in the thymus of MM-inflicted mice.

C57BL KaRij mice were purchased from Harlan CPB and further bred at the Weizmann Institute breeding center. All experimental procedures were approved by the Weizmann Institute’s Animal Care Committee.

5T2 MM (IgG2aκ) and 5T33 MM (IgG2bκ) tumor cells were propagated by injection in young syngeneic mice via the lateral tail vein and further maintained by injecting BM from diseased mice into appropriate recipients. We also established a 5T33 MM cell line into tissue culture. BM cells from tibiae and femora of sick mice were cocultured with normal splenocytes in RPMI 1640 supplemented with 10% FCS, penicillin (100 IU/ml), streptomycin (100 μg/ml), nonessential amino acids (1 ml/100 ml), l -glutamine (2 mM), and sodium pyruvate (5 mM). Within 2 wk of culture, an independent cell line emerged.

5T33 MM cells (3 × 104) propagated in culture, or 3 × 105 BM cells from 5T2 MM-inflicted C57BL KaRij mice, were i.v. injected into C57BL KaRij normal females at the age of 8–10 wk. Injected mice were euthanized when hind limb paralysis or other signs of morbidity occurred. All healthy mice used as controls were females age-matched to MM-inflicted mice.

Cells from the BM, spleen, LNs, and PB were stained with the following Abs: anti-CD4 (GK1.5), anti-CD8 (53-6.7), anti-CD25 (PC61.5), anti-CD69 (H1.2F3), anti- glucocorticoid-induced TNFR (GITR) (DTA-1), and anti-TCRβ (H57-597) and their matched isotype controls, all purchased from eBioscience. PerCPcy5.5-conjugated anti-CD8 Ab (53-6.7) was purchased from BD Pharmingen. Intracellular Foxp3 staining was conducted using the FITC anti-mouse/rat anti-Foxp3 staining kit (eBioscience) following the manufacturer’s instructions. This kit contains a FITC-conjugated anti-mouse/rat Foxp3 Ab (FJK-16s). For cell cycle analysis, cells fixed and permeabilized using the Foxp3 staining kit (eBioscience) were stained with 20 μl of 7-aminoactinomycin D (BD 51-2359KC) for 10 min on ice and analyzed. For the analysis of the percentage of cycling cells by BrdU incorporation, 1 mg BrdU was i.p. injected to 5T2-inflicted mice and to normal controls. After 1 day, organs were taken out and cells stained with the BrdU Flow kit (BD Pharmingen) following the manufacturer’s instructions. For intracellular CTLA-4 staining, membranal staining with anti-CD4-FITC was followed by fixation and permeabilized using the Leukoperm kit (Serotec). Cells were fixed in 100 μl of fixative (reagent A), washed, and resuspended in 100 μl of permeabilization reagent (reagent B) in the presence of anti-CTLA-4-PE Ab (clone 1B8, SouthernBiotech) or PE-conjugated isotype control (SouthernBiotech). Stained cells were analyzed using FACSort (BD Biosciences) and with CellQuest 3.1 software (BD Pharmingen). PBMCs were obtained using Ficoll-Paque Plus (GE Healthcare) density centrifugation. CD4, CD25, and Foxp3 staining was used to evaluate Treg ratios among CD4+ T cells, and the CD4/CD8 ratio was calculated as the percentage of CD4+ T cells divided by the percentage of CD8+ T cells as described before (29).

Thymus cell suspensions were obtained using a stainless steel mesh. Cells were suspended in PBS, counted, and total thymocyte number was calculated as cell number per milliliter amplified by the volume of the solution. Teff and Treg numbers per thymus were calculated as the percentage of Teff and Treg in the thymus as found in FACS analysis, amplified by the total thymocyte number.

Suppression assay was performed as reported earlier (30, 31) and as modified by Kipnis and colleagues (32). Treg and Teff cells were purified from the spleens of healthy controls (n = 8–9 spleens) or 5T2-inflicted mice (n = 5–7 spleens). CD4+ cells were enriched by magnetic bead separation using IMag anti-mouse CD4 particles (BD Biosciences cat. no. 551539). CD4+ cells were incubated with sterile anti-CD4-APC and anti-CD25-PE Abs (eBioscience) and CD4+CD25high (Tregs) and CD4+CD25 (Teffs) were sorted using FACSVantage cell sorter (BD Biosciences). Purity was >98%. Purified Teffs (106 cells/ml) were maintained for 72 h in “resting medium”, that is, RPMI 1640 supplemented with l -glutamine (2 mM), sodium pyruvate (5 mM), penicillin (100 IU/ml), streptomycin (100 μg/ml), nonessential amino acids (1 ml/100 ml), 5% FCS, and 10% T cell growth factor (medium collected from splenocytes cultured at the concentration of 2 × 106 cells/ml in the presence of concavalin A (1.25 ng/ml)) for 48 h. Purified Tregs (5 × 105 cells/ml) were activated for 72 h in 24-well plates for 3 days in “activation medium”, that is, RPMI 1640 supplemented with l -glutamine (2 mM), 2-ME (5 × 10−5 mM), sodium pyruvate (5 mM), penicillin (100 IU/ml), streptomycin (100 μg/ml), nonessential amino acids (1 ml/100 ml), and 2.5% FCS in the presence of mouse recombinant IL-2 (5 ng/ml), soluble anti-CD3 Ab (1 μg/ml, 145-2C11, eBioscience), and irradiated (2500 rad) splenocytes (1.5 × 106 cells/ml). After 72 h, Teffs were activated in the presence of different Treg numbers to test for suppression activity. Teffs (50 × 103 cells/well) were cocultured in quadruplicates with decreasing numbers of activated Tregs for 72 h in 96-well flat-bottom plates in the presence of irradiated splenocytes (106/ml) and anti-CD3 Ab (1 μg/ml). [3H]thymidine (1 μCi) was added for the last 16 h of culture. After the cells were harvested, their [3H]thymidine content was analyzed by the use of a gamma counter.

5T2 cells (3 × 105 cells) were i.v. injected to 8-wk-old C57BL KaRij females, followed by the injection of thymocytes, once a week for 3 wk, starting at 42 days after tumor injection. Thymocytes from 5T2-inflicted mice showing hind limb paralysis were purified and mixed together. Equivalent numbers of donors and recipients were used, so thymocytes from one thymus in average (∼30 × 106) were injected to one recipient. Similar numbers of control thymocytes were injected to each mouse. A third group of 5T2-injected mice were not injected with thymocytes. Recipient mice were killed when hind leg paralysis occurred, and tumor growth in the bones was examined by H&E staining. To purify the different thymus populations of single-positive (SP) CD4+CD25+ Tregs, SP CD4+CD25 Teffs, and double-negative (DN), double-positive (DP), and CD8+ cells, thymocytes from 16 5T2-inflicted paralyzed females and from 18 age-matched control female mice were fractionated by the BD IMag separation system using anti-mouse CD8a particles (BD Biosciences cat. no. 551516). Positively selected cells were DP and CD8+ cells. Negatively selected cells were DN and CD4+ cells. The latter were then sorted into CD4+CD25+ cells, CD4+CD25 cells, and DN cells using a FACSAria cell sorter (BD Biosciences). Whole thymocytes (∼30 × 106, taken from the thymocyte mix before separation) and Tregs (150,000) from control or 5T2-inflicted mice were injected to 42-day-old 5T2-injected females, and the procedure was repeated 1 wk later. A mix of DP cells, CD8+ cells, DN cells, and SP CD4+CD25 cells was prepared according to the original numbers found in the thymus of 5T2-inflicted mice, which therefore contained all thymocytes excluding Tregs (6 × 106 DN cells, 4.2 × 106 CD4+CD25 Teffs, 16 × 106 DP + CD8+ cells). A mix containing the same cell numbers from each thymic population was prepared from normal thymocytes as a completely equivalent control. Hind limb paralysis and tumor growth in the bones were examined as described above.

A two-tailed unpaired Student’s t test was used for the analysis of the differences between MM-inflicted mice and controls presented in Figs. 1–4 and between the different experimental groups presented in Fig. 6 C. Univariant regression analyses assessing the correlations between each of the two variables were conducted using Matlab 7.0 software or SAS 9.1 software. A logarithmic stepwise multiple regression analysis was conducted to assess the predictive values of the determining variables (thymus size, DP, DN, SP CD4+, and SP CD8+ populations, and TCR mean fluorescence intensity (MFI)) on the dependent variables (TCR MFI, CD25 expression, and Foxp3 expression at the SP and DP stages) using SAS 9.1 software.

FIGURE 1.

Increased Treg ratios in the periphery of MM-inflicted mice. A, Representative FACS samples of CD25 vs Foxp3 expression by CD4+ splenocytes from a control, a 5T2-inflicted mouse, and a 5T33-inflicted mouse. Percentages of CD25highFoxp3+ cells are indicated. B, Percentage (mean ± SEM) of CD25highFoxp3+CD4+ splenocytes from 5T2- and 5T33-inflicted mice. ∗, p < 0.05 and ∗∗, p < 0.005 vs controls. Results are representative of two and three experiments in 5T2- and 5T33-inflicted mice, respectively. C, Percentage (mean ± SEM) of GITR+Foxp3+, CTLA-4+Foxp3+, and CD69+Foxp3+ cells among CD4+ splenocytes from 5T2- and 5T33-inflicted mice. ∗, p < 0.005; ∗∗, p < 10−6; ∗∗∗, p < 10−13 vs controls. Results are representative of two and three experiments in 5T2- and 5T33-inflicted mice, respectively. D, Representative FACS samples of GITR, CTLA-4, and CD69 vs Foxp3 expression by CD4+ splenocytes from a control and a 5T2 mouse. Percentages of GITR+Foxp3+, CTLA-4+Foxp3+, and CD69+Foxp3+ cells are indicated. E, In vitro proliferation of normal Teffs in the presence of the indicated ratios of CD4+CD25high (Treg) splenocytes (mean of quadruplicates ± SEM). [3H]thymidine incorporation was tested after 3 days. Results are representative of two independent experiments. F, Percentage (mean ± SEM) of CD4+ cells from surrounding LNs (SLN) and distal LNs (DLN) expressing CD25highFoxp3+ from 5T2-inflicted mice. ∗, p = 0.0005 and ∗∗, p < 0.0005 vs control for SLN and DLN. Results are representative of two experiments in 5T2-inflicted mice and three experiments in 5T33-inflicted mice. G, Percentage (mean ± SEM) of CD4+ cells from the spleen, BM, and PB of healthy mice and mice injected with 5T2 cells for the indicated times. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p ≤ 0.0005 vs control. One experiment representative of two with similar results.

FIGURE 1.

Increased Treg ratios in the periphery of MM-inflicted mice. A, Representative FACS samples of CD25 vs Foxp3 expression by CD4+ splenocytes from a control, a 5T2-inflicted mouse, and a 5T33-inflicted mouse. Percentages of CD25highFoxp3+ cells are indicated. B, Percentage (mean ± SEM) of CD25highFoxp3+CD4+ splenocytes from 5T2- and 5T33-inflicted mice. ∗, p < 0.05 and ∗∗, p < 0.005 vs controls. Results are representative of two and three experiments in 5T2- and 5T33-inflicted mice, respectively. C, Percentage (mean ± SEM) of GITR+Foxp3+, CTLA-4+Foxp3+, and CD69+Foxp3+ cells among CD4+ splenocytes from 5T2- and 5T33-inflicted mice. ∗, p < 0.005; ∗∗, p < 10−6; ∗∗∗, p < 10−13 vs controls. Results are representative of two and three experiments in 5T2- and 5T33-inflicted mice, respectively. D, Representative FACS samples of GITR, CTLA-4, and CD69 vs Foxp3 expression by CD4+ splenocytes from a control and a 5T2 mouse. Percentages of GITR+Foxp3+, CTLA-4+Foxp3+, and CD69+Foxp3+ cells are indicated. E, In vitro proliferation of normal Teffs in the presence of the indicated ratios of CD4+CD25high (Treg) splenocytes (mean of quadruplicates ± SEM). [3H]thymidine incorporation was tested after 3 days. Results are representative of two independent experiments. F, Percentage (mean ± SEM) of CD4+ cells from surrounding LNs (SLN) and distal LNs (DLN) expressing CD25highFoxp3+ from 5T2-inflicted mice. ∗, p = 0.0005 and ∗∗, p < 0.0005 vs control for SLN and DLN. Results are representative of two experiments in 5T2-inflicted mice and three experiments in 5T33-inflicted mice. G, Percentage (mean ± SEM) of CD4+ cells from the spleen, BM, and PB of healthy mice and mice injected with 5T2 cells for the indicated times. ∗, p < 0.05; ∗∗, p < 0.005; ∗∗∗, p ≤ 0.0005 vs control. One experiment representative of two with similar results.

Close modal
FIGURE 2.

Changes in thymic structure and cell composition in MM-inflicted mice. A, Decreased thymus weight in 5T2 (p < 1 × 10−10 vs control) and 5T33-inflicted mice (p < 1 × 10−8 vs control). B, Representative H&E staining of a control thymus and atrophied thymuses from 5T2- and 5T33-inflicted mice. Scale bar represents 200 μm. C, Decreased thymocyte numbers (mean ± SEM) in 5T2-inflicted mice. ∗, p = 0.0002. D, Thymus size (mean ± SEM) in mice injected with 5T2 cells for the indicated times. ∗, p < 10−5; ∗∗, p < 10−9; ∗∗∗, p < 10−10 vs control. E, Representative FACS samples demonstrating thymic subset division in control and in 5T2- and 5T33-inflicted mice. F, Altered percentage (mean ± SEM) of DN, DP, CD4+ SP, and CD8+ SP subpopulations in 5T2-inflicted mice. ∗, p = 0.01 and ∗∗, p < 0.005 vs control. Sum of three experiments in 5T2-inflicted mice. Similar results were obtained in two experiments in 5T33-inflicted mice.

FIGURE 2.

Changes in thymic structure and cell composition in MM-inflicted mice. A, Decreased thymus weight in 5T2 (p < 1 × 10−10 vs control) and 5T33-inflicted mice (p < 1 × 10−8 vs control). B, Representative H&E staining of a control thymus and atrophied thymuses from 5T2- and 5T33-inflicted mice. Scale bar represents 200 μm. C, Decreased thymocyte numbers (mean ± SEM) in 5T2-inflicted mice. ∗, p = 0.0002. D, Thymus size (mean ± SEM) in mice injected with 5T2 cells for the indicated times. ∗, p < 10−5; ∗∗, p < 10−9; ∗∗∗, p < 10−10 vs control. E, Representative FACS samples demonstrating thymic subset division in control and in 5T2- and 5T33-inflicted mice. F, Altered percentage (mean ± SEM) of DN, DP, CD4+ SP, and CD8+ SP subpopulations in 5T2-inflicted mice. ∗, p = 0.01 and ∗∗, p < 0.005 vs control. Sum of three experiments in 5T2-inflicted mice. Similar results were obtained in two experiments in 5T33-inflicted mice.

Close modal
FIGURE 3.

Increased ratios of Treg development in the thymus of MM-inflicted mice. A, Percentage (mean ± SEM) of CD25+ CD4+ SP and DP cells in the thymus of 5T2-inflicted mice. ∗, p = 0.001 and ∗∗, p = 10−5 vs control. Sum of three experiments in 5T2-inflicted mice. Similar results were obtained in two experiments in 5T33-inflicted mice. B, A high percentage of CD25+ SP thymocytes (gated as shown in the left panel) express Foxp3 in control and 5T2-inflicted mice. C, left, Representative FACS samples of CD25highFoxp3+ expression by DP thymocytes from a control and a 5T2-inflicted mouse. R5, CD25+Foxp3+ cells; R6, CD25+Foxp3 cells. Percentages of cells in R5 and R6 are indicated. Right, The percentage (mean ± SEM) of DP CD25+ cells expressing Foxp3 in normal and 5T2-inflicted mice. ∗, p < 0.05. SEM is not seen because it is smaller than 0.1. Sum of three experiments in 5T2-inflicted mice. D, The percentage (mean ± SEM) of CD62Lhigh Tregs in the periphery (spleen + LN) and thymus of control and 5T2-inflicted mice. ∗, p < 0.05. E, Treg numbers in controls and 5T2-inflicted mice (mean ± SEM) are similar (p = 0.80). F, Teff numbers (mean ± SEM) in 5T2-inflicted mice are reduced. ∗, p = 0.01. G, Numbers of Tregs and Teffs (mean ± SEM) in control thymuses and in slightly atrophied (Slightly A.), moderately atrophied (Moderately A.), and severely atrophied (Severely A.) thymuses from 5T2-inflicted mice show a decrease in Treg number only in severely atrophied thymuses. p values are indicated only when differences are significant. ∗, p < 0.05.

FIGURE 3.

Increased ratios of Treg development in the thymus of MM-inflicted mice. A, Percentage (mean ± SEM) of CD25+ CD4+ SP and DP cells in the thymus of 5T2-inflicted mice. ∗, p = 0.001 and ∗∗, p = 10−5 vs control. Sum of three experiments in 5T2-inflicted mice. Similar results were obtained in two experiments in 5T33-inflicted mice. B, A high percentage of CD25+ SP thymocytes (gated as shown in the left panel) express Foxp3 in control and 5T2-inflicted mice. C, left, Representative FACS samples of CD25highFoxp3+ expression by DP thymocytes from a control and a 5T2-inflicted mouse. R5, CD25+Foxp3+ cells; R6, CD25+Foxp3 cells. Percentages of cells in R5 and R6 are indicated. Right, The percentage (mean ± SEM) of DP CD25+ cells expressing Foxp3 in normal and 5T2-inflicted mice. ∗, p < 0.05. SEM is not seen because it is smaller than 0.1. Sum of three experiments in 5T2-inflicted mice. D, The percentage (mean ± SEM) of CD62Lhigh Tregs in the periphery (spleen + LN) and thymus of control and 5T2-inflicted mice. ∗, p < 0.05. E, Treg numbers in controls and 5T2-inflicted mice (mean ± SEM) are similar (p = 0.80). F, Teff numbers (mean ± SEM) in 5T2-inflicted mice are reduced. ∗, p = 0.01. G, Numbers of Tregs and Teffs (mean ± SEM) in control thymuses and in slightly atrophied (Slightly A.), moderately atrophied (Moderately A.), and severely atrophied (Severely A.) thymuses from 5T2-inflicted mice show a decrease in Treg number only in severely atrophied thymuses. p values are indicated only when differences are significant. ∗, p < 0.05.

Close modal
FIGURE 4.

Increased activation of early thymocytes in MM-inflicted mice. A, Increased percentage (mean ± SEM) of CD69+ DP cells and in TCR cell-surface levels (TCR MFI) on DP cells in the thymus of 5T2-inflicted mice. ∗, p < 0.005 vs control. Sum of three experiments in 5T2-inflicted mice. Similar results were obtained in two experiments in 5T33-inflicted mice. B, Representative FACS samples showing TCR expression levels of DP cells (left) and SP cells (right) from a control and a 5T2-inflicted mouse. A DP population appears in 5T2-inflicted mice with TCR cell-surface levels similar to SP cells. C, Representative FACS analysis showing no increase in the percentage of thymocytes expressing CD138 in the thymus of 5T2-inflicted mice. D, Correlations (R2) between the thymic parameters (thymus weight (Weight), DN, DP, CD4+ SP (CD4), and CD8+ SP (CD8) subpopulations) and the parameters CD25 expression (CD25), Foxp3 expression (Foxp3), and TCR MFI at the DP stage (TCR MFI).

FIGURE 4.

Increased activation of early thymocytes in MM-inflicted mice. A, Increased percentage (mean ± SEM) of CD69+ DP cells and in TCR cell-surface levels (TCR MFI) on DP cells in the thymus of 5T2-inflicted mice. ∗, p < 0.005 vs control. Sum of three experiments in 5T2-inflicted mice. Similar results were obtained in two experiments in 5T33-inflicted mice. B, Representative FACS samples showing TCR expression levels of DP cells (left) and SP cells (right) from a control and a 5T2-inflicted mouse. A DP population appears in 5T2-inflicted mice with TCR cell-surface levels similar to SP cells. C, Representative FACS analysis showing no increase in the percentage of thymocytes expressing CD138 in the thymus of 5T2-inflicted mice. D, Correlations (R2) between the thymic parameters (thymus weight (Weight), DN, DP, CD4+ SP (CD4), and CD8+ SP (CD8) subpopulations) and the parameters CD25 expression (CD25), Foxp3 expression (Foxp3), and TCR MFI at the DP stage (TCR MFI).

Close modal
FIGURE 6.

Representative H&E stained section of lumbar vertebrae from a mouse injected with 5T2 cells and thymocytes from paralyzed 5T2-inflicted mice showing severe bone fractions and massive tumor invasion into the surrounding skeletal muscle. A, A small magnification (×10) image of the area showing the severely destructed spine (SP), and a large tumor mass surrounding the vertebrae. Arrowhead points are at residual skeletal muscle tissue in the tumor mass. B, A larger magnification (×40) image of a single vertebra indicated in A, showing myeloma cells (MM) that replaced the BM and invaded the soft tissue through the destructed bone. Arrowheads point to severe bone fractions. Scale bar represents 100 μm in A and B. C, Percentage of paralyzed mice in “whole thymus” (control, n = 6; 5T2, n = 5), Treg (control, n = 4; 5T2, n = 4) and “Treg-depleted” (control, n = 3; 5T2: n = 3) injected mice at the indicated time points after thymocyte administration (first time point of day 42).

FIGURE 6.

Representative H&E stained section of lumbar vertebrae from a mouse injected with 5T2 cells and thymocytes from paralyzed 5T2-inflicted mice showing severe bone fractions and massive tumor invasion into the surrounding skeletal muscle. A, A small magnification (×10) image of the area showing the severely destructed spine (SP), and a large tumor mass surrounding the vertebrae. Arrowhead points are at residual skeletal muscle tissue in the tumor mass. B, A larger magnification (×40) image of a single vertebra indicated in A, showing myeloma cells (MM) that replaced the BM and invaded the soft tissue through the destructed bone. Arrowheads point to severe bone fractions. Scale bar represents 100 μm in A and B. C, Percentage of paralyzed mice in “whole thymus” (control, n = 6; 5T2, n = 5), Treg (control, n = 4; 5T2, n = 4) and “Treg-depleted” (control, n = 3; 5T2: n = 3) injected mice at the indicated time points after thymocyte administration (first time point of day 42).

Close modal

In light of the contradictory data concerning Treg behavior in MM (22, 26) and other B cell malignancies (23, 24, 25), we first sought to characterize Treg frequency and function in MM-inflicted mice. Expression of the Treg markers CD25 and Foxp3 by CD4+ splenocytes was tested using flow cytometry (FACS). The percentage of CD4+ splenocytes expressing a CD25highFoxp3+ phenotype significantly increased in 5T2- (p < 0.05) and in 5T33-inflicted mice (p < 0.005), as shown in Fig. 1, A and B. These results were confirmed using the Treg markers GITR (p < 10−8 and p < 10−7 for 5T2- and 5T33-inflicted mice, respectively), CTLA-4 (p < 0.005 and p < 10−6, respectively), and CD69 (p < 10−7 and p < 10−13, respectively) as shown in Fig. 1,C. Since GITR, CTLA-4, and CD69 are expressed by activated CD4+ cells as well, coexpression with Foxp3 was tested. As shown in Fig. 1,D, the percentage of CD4+ splenocytes presenting Foxp3+GITR+, Foxp3+CTLA-4+, and Foxp3+CD69+ expression significantly increased in 5T2 mice (24.0 ± 1.5% vs 14.9 ± 0.5%, 13.4 ± 0.9% vs 9.1 ± 0.1%, and 15.9 ± 1.6% vs 7.0 ± 0.1%, respectively; p < 0.005), indicating increased Treg ratios among CD4+ cells. To test whether Tregs accumulating in the spleen of MM-inflicted mice were suppressive-functional, we compared the in vitro suppressive function of CD4+CD25high splenocytes from normal and 5T2-inflicted mice. The vast majority of CD4+CD25high splenocytes coexpressed Foxp3 in healthy and MM-inflicted mice (94.5 ± 0.7% and 93.3 ± 2.6%, respectively). As shown in Fig. 1 E, CD4+CD25high (Treg) cells from 5T2-inflicted mice suppressed the proliferation of CD4+CD25 responder cells (Teff) as efficiently as control Tregs, and they were anergic to in vitro proliferation, indicating that the cells were regulatory functional.

LN cells were purified from individual mice in parallel to splenocytes, and CD25 and Foxp3+ expression was examined. As shown in Fig. 1 F, Treg ratios significantly increased in the LNs of 5T2-inflicted mice. Notably, Treg ratios similarly increased in LNs surrounding the main sites of tumor infiltration (surrounding LNs, or SLN; inguinal, caudal, and lumbar nodes) and in LNs distal to the main tumor sites (distal LNs, or DLN; superficial cervical nodes, axillary nodes, and brachial nodes). Similar data were collected in 5T33-inflicted mice (data not shown).

In human patients, MM cells locate mainly to the BM, and increased Treg levels are reported in the PB and BM of MM patients (22). The 5T2 tumor resembles the human disease in its main localization to the BM (28, 33). Hind limb paralysis occurs as an early symptom of the disease in 5T2-inflicted mice, due to spinal cord compression. Further progression of the disease symptoms involves signs of paraplegia, bone lesions, and morbidity. To follow the course of Treg accumulation during disease progression, CD25 vs Foxp3 expression by CD4+ cells was tested in the spleen, BM, and PB of 5T2-inflicted mice at different time points (28, 42, 66, 90, and 104 days) following 5T2 tumor injection. In the first two time points mice were asymptomatic. Paralysis appeared around day 66 and became severe with latency. As shown in Fig. 1,G, Treg frequency among CD4+ splenocytes significantly increased at an early stage (28 days; p < 10−5) and did not increase with latency. In contrast, Treg frequency in the BM remained normal at 66 days after tumor injection in both paralyzed and nonparalyzed mice (Fig. 1,G and data not shown). However, at 90 and 104 days after tumor injection, a significant increase in BM Treg levels was observed (Fig. 1 G; p < 0.005). In PB, a mild but significant increase was found before paralysis onset (42 days; p < 0.005) and remained constant during disease progression. To conclude, Treg frequencies increased in the spleen and PB early after 5T2 tumor injection and did not further increase with disease latency and/or progression. A significant increase in BM Treg levels was observed only at later stages of the disease.

Since Tregs normally develop in the thymus (18, 19, 20, 21), we wanted to examine whether thymic processes were involved in the increased Treg frequency in the periphery of MM-inflicted mice (Fig. 1). To this end, we characterized thymus structure and composition following disease. Atrophy of the thymus in MM-inflicted mice was evident and was reflected by a significant reduction in thymus weight (p < 10−10 and p < 10−8 in 5T2 and 5T33, respectively; Fig. 2,A), as well as by the distortion of normal distinction between cortical and medullar areas (Fig. 2,B) in mice inflicted with both tumors. Thymus cellularity decreased by ∼5.5 on average, as shown in Fig. 2,C. The decrease in thymocyte numbers was significantly correlated with thymus atrophy, as expected (R2 = 0.56, p < 0.01). To follow the process of thymus atrophy during disease development, we tested thymus size in 5T2-injected mice starting at 42 days after tumor injection, before onset of paralysis. Hind limb paralysis occurs at 60–70 days after tumor injection as an early complication of the disease in 5T2-inflicted mice, caused by spinal cord compression. Further progressive severity of the disease symptoms involves signs of paraplegia, bone lesions, and morbidity. As shown in Fig. 2,D, no thymus atrophy was found in asymptomatic mice 42 days after tumor injection (p = 0.3). Thymus atrophy was observed 60–68 days after tumor injection (p < 10−5) when paralysis appeared in part of the mice (Fig. 2,D and data not shown). Importantly, thymus atrophy occurred only in paralyzed mice (20 of 25 paralyzed vs 0 of 5 nonparalyzed mice). Thymus atrophy became more prominent with disease latency, as shown in Fig. 2 D (90 and 104 days; p < 10−9 and p < 10−10, respectively). Notably, severity of paralysis increased, and bone lesions appeared along with disease latency. To conclude, thymus atrophy in 5T2-injected mice occurred around the time of paralysis onset and was associated with disease latency and/or severity.

To examine developmental processes in the thymus of MM-inflicted mice, the proportions of the main thymocyte subpopulations were analyzed by FACS. As shown in Fig. 2, E and F, the proportion of the CD4+CD8+ DP population, normally the largest thymocyte population, significantly decreased in 5T2-inflicted mice (p < 0.005), while the proportions of the CD4CD8 DN population significantly increased (p = 0.01). The proportions of the most mature populations of CD4+ and CD8+ SP cells significantly increased in the thymus of 5T2-inflicted mice (p < 0.005), suggesting that a change of kinetics rather than a developmental block at the DN stage was responsible for changes in subpopulation proportions. Changes in the proportions of thymus subpopulations were significantly correlated with thymus atrophy (p < 0.05). Similar data were collected from 5T33-inflicted mice (data not shown). These data demonstrate that thymuses in MM-inflicted mice were atrophied and that this atrophy was associated with changes in developmental processes in the thymus.

Although the thymus is normally the main site of Treg development (18, 19, 20, 21), Treg development in the thymus during malignancy has not been directly examined. Following our findings of significant changes in thymus characteristics in MM-inflicted mice (Fig. 2), we examined whether Treg development was altered in the thymus of MM-inflicted mice. We found that the frequency of mature (CD4+SP) thymocytes expressing CD25 significantly increased in 5T2-inflicted mice (p = 10−5; Fig. 3,A) by ∼2-fold. As shown in Fig. 3,B, most of the CD25+CD4+ SP cells coexpressed Foxp3. There was no significant difference in the percentage of Foxp3+ cells among CD4+ SP CD25+ cells in healthy and 5T2-inflicted mice (77.5 ± 5.5% and 72.2 ± 4.0%, respectively). Interestingly, CD25 expression was increased already at the DP stage (p = 0.001; Fig. 3,A). Although most CD25+ DP cells did not express Foxp3 (Fig. 3,C), the frequency of CD25+Foxp3+ cells significantly increased at this stage as demonstrated in Fig. 3 C (p < 0.05; R5: CD25+Foxp3+ cells), and this increase was accompanied by a decrease in the ratios of CD25+Foxp3 DP cells (R6). These results are in accordance with previous data suggesting the commitment to the Treg lineage as early as the DP stage (34, 35, 36).

Increased CD25+Foxp3+ expression in the DP stage (Fig. 3, A and C) implies that increased Treg ratios among mature thymocytes result from changes in developmental processes in the thymus of MM-inflicted mice. However, to exclude the possibility that increased Treg ratios reflect Treg recirculation from the periphery, we compared the naive phenotype of Tregs in the thymus and in the periphery of 5T2-inflicted mice. Naive T cells are marked by the bright expression of CD62L in the mouse. It is expected that Tregs originating in the thymus would present a naive phenotype, while Tregs in the periphery might have been activated and lost their naive phenotype. As shown in Fig. 3,D, the percentage of peripheral (spleen and LN mix) CD62Lhigh Tregs significantly decreased in 5T2-inflicted mice compared with controls (p < 0.05), indicating the loss of naive phenotype. However, Tregs in the thymus of the same mice retained a naive phenotype, and a statistically insignificant increase was observed compared with controls (p = 0.4). Teffs in the periphery and thymus of 5T2-inflicted mice showed similar trends (data not shown). These results present a distinct and naive phenotype of Tregs in the thymus of 5T2-inflicted mice and suggest that Treg recirculation from the periphery is not the cause of Treg increased ratios in the thymus. Treg memory phenotype was also tested in the thymus and periphery using the memory marker CD44. The percentage of Tregs expressing CD44 was not changed in the periphery (spleen and LN mix) or in the thymus of 5T2-inflicted mice (47.3 ± 1.6% vs 47.3 ± 1.1% and 34.5 ± 1.7% vs 35.8 ± 0.9%, respectively). Analysis of CD44 and CD62L coexpression showed a similar decreased in CD62Lhigh-expressing cells as found in the total Treg population (shown in Fig. 3 D), suggesting a shift from a CD44highCD62high (central memory) to a CD44highCD62low (effector memory) phenotype among peripheral Tregs.

Thymus atrophy in MM-inflicted mice was associated with a dramatic reduction in cellularity (Fig. 2,C). Therefore, we were interested in testing Treg numbers developing in the atrophied thymus to assess whether Tregs produced in the thymus could still be physiologically significant. Notably, Treg absolute numbers in the thymus of 5T2-inflicted mice were not altered compared with control mice, as shown in Fig. 3,E. In contrast, the number of Teffs (CD25) was reduced by ∼2.5-fold (p = 0.01, Fig. 3,F). To follow the amount of Treg development during progression of thymus atrophy, we divided thymuses from 5T2-inflicted mice into three groups. A group of slightly atrophied thymuses included thymuses that weighed more than two SDs below the normal average (>25 mg). Moderately atrophied thymuses weighed more than three SDs below the normal average (18 mg ≤ weight ≤ 25 mg), and severely atrophied thymuses weighed <18 mg. As shown in Fig. 3,G, Treg numbers were even increased in slightly and moderately atrophied thymuses, and they decreased only in the severely atrophied thymuses (Fig. 3,G). In contrast, Teff numbers constantly decreased and the decrease was significant already in the moderately atrophied thymuses (p = 0.01 and p = 0.002 in moderately and severely atrophied thymuses, respectively). These data show that until severe thymus atrophy occurs, Treg numbers do not fall below normal numbers. Even in severely atrophied thymuses, Treg numbers are reduced by ∼3-fold, compared with a more dramatic ∼9.5-fold reduction of Teff numbers (Fig. 3 G).

The effects of thymus atrophy on peripheral Teff numbers and function in MM are largely unknown (37). Teff numbers were tested in the PB of 5T2-inflicted mice, since T cell numbers in the spleen and BM may be largely affected by T cell migration to these tumor sites. Progression of MM is known to be associated with depletion of the CD4+ population, measured by the decreased CD4+/CD8+ ratio in the blood of human patients (29, 37). Using the same measure to test Teff levels in the blood, we found a significant correlation between thymus size and the CD4+/CD8+ ratio in the PB of 5T2-inflicted mice (p < 0.05). PB CD4+ consisted of ≥95% Teffs (Fig. 1 G). These results imply that low Teff frequency in the PB of 5T2-inflicted mice may be associated with thymus atrophy.

To test whether Teff function was affected by thymus atrophy, Teff cell cycle, proliferation, and activation were tested in the spleen and LNs of mice with different levels of thymus atrophy. No differences in cell cycle were found in 5T2-inflicted mice compared with controls or among the diseased mice (data not shown). Using BrdU incorporation, we found a significant increase in the percentage of proliferating Teffs in the spleen and LNs of 5T2 mice compared with controls (spleen, 5.6 ± 0.3% vs 2.7 ± 0.8%; LN, 2.6 ± 0.2% vs 1.1 ± 0.2%; p < 0.005). Similarly, Teffs were highly activated in the spleen and LNs of 5T2-inflicted mice as indicated by bright CD69 expression (spleen, 30.9 ± 2.5 vs 8.2 ± 0.4; LN, 22.6 ± 2.6 vs 8.5 ± 0.1; p < 0.05). However, we did not find a significant correlation between thymus size and Teff proliferation or activation (p > 0.08). Additionally, Teff activation did not increase in the PB, suggesting that Teff activation in the spleen and LN of 5T2-inflicted mice resulted from peripheral rather than thymic processes. To conclude, Teff depletion, but not activation or proliferation in the periphery, is significantly correlated with thymus atrophy. However, since thymus atrophy is associated with disease severity/progression (Fig. 2 D), we cannot conclude whether Teff depletion results from thymus atrophy or from other processes associated with the disease.

Several reports have suggested that Treg development is promoted by high avidity interactions of the TCR with the thymic stroma (35, 38, 39, 40). We therefore examined whether TCR engagement was increased in the thymus of MM-inflicted mice and could contribute to Treg induction in the thymus. TCR activation at the DP stage leads to positive selection, resulting in the expression of the thymic selection marker CD69 and further maturation, reflected by a continuous increase in TCR surface levels (41, 42). Therefore, we examined the percentage of CD69 expression by thymocytes at different developmental stages. Additionally, relative TCR cell-surface levels were estimated by TCR MFI. As shown in Fig. 4,A, both the frequency of CD69+ cells and TCR MFI significantly increased in DP thymocytes from 5T2-inflicted mice (p < 0.005). No significant difference was observed in mature thymocytes (p ≥ 0.095). Similar results were found in 5T33-inflicted mice (data not shown). As presented in Fig. 4,B, the increase in TCR MFI reflected a minor increase in TCR levels on the major DP thymocyte population, as well as the appearance of a DP subpopulation with TCR levels comparable to SP cells. These data indicate that TCR engagement at the DP stage is elevated in the thymus of MM-inflicted mice. Since tumor cells localize to the BM and spleen, it seems possible that some tumor cells could localize to thymic tissue, possibly causing increased thymocyte activation and thymus atrophy. However, we could not trace any MM cells by histological examination of the thymus (high magnitude examination of histological sections as shown in Fig. 2,B) or by FACS analysis, using an Ab specific to the plasma cell marker CD138 (Fig. 4 C).

Regression analysis showed that TCR MFI at the DP stage was significantly correlated with thymus atrophy in thymuses from 5T33 (p < 0.005; Table I). Interestingly, regression analysis revealed a highly significant correlation between DN and DP proportions and TCR MFI in both the 5T2 and 5T33 tumors (Table I) that was stronger than the association of thymus atrophy with this dependent parameter (R2 values from 5T33-inflicted mice are presented in Fig. 4,D). Although correlations in the 5T2 model were relatively low compared with the 5T33 model, possibly as a result of several samples showing extreme TCR MFI values, trends were clearly similar in the two tumor models (Table I). In accordance with results from univariant regression analysis (Table I and Fig. 4,D), a stepwise multiple regression analysis showed that the proportion of the DP population was the dominant variable predictive of TCR MFI both in the 5T2 and 5T33 tumor models (R2 = 0.83 and R2 = 0.93, respectively; p < 0.0001). In contrast, a stepwise multiple regression analysis performed to test the effect of early thymocyte activation on Treg development did not indicate a predictive value of TCR MFI for Treg development, as described below. Univariant regression analysis indicated a significant correlation of DP TCR MFI with CD25 and Foxp3 expression by SP thymocytes in 5T33-inflicted mice (Table I), but not in 5T2-inflicted mice (Table I). CD69 expression was not correlated with CD25/Foxp3 expression, nor with thymus atrophy and subpopulation proportions in the involuted thymus (R2 ≤ 0.17). Taken together, these results suggest that TCR surface levels at the DP stage are affected by thymus atrophy and more so by DP and DN subpopulation proportions in the atrophied thymus. Additionally, CD69 and TCR expression at this stage seem to be controlled by different mechanisms.

Table I.

Correlation between thymic characters and expression of CD25, Foxp3, and TCR MFI in thymocyte populations from 5T2- and 5T33-inflicted micea

CD25 CD4+ SPFoxp3 CD4+ SPTCR MFI DP
Thymus weight    
 5T2 <0.05 <0.05 NS 
 5T33 <0.01 <0.005 <0.005 
DN proportions    
 5T2 <10−6 <10−5 <0.005 
 5T33 <10−4 <10−4 <10−11 
DP proportions    
 5T2 <0.0005 <0.005 <0.005 
 5T33 <10−6 <0.0005 <10−16 
CD4 proportions    
 5T2 NS <0.01 NS 
 5T33 <0.05 NS <0.05 
CD8 proportions    
 5T2 <0.01 <0.05 NS 
 5T33 NS NS NS 
TCR MFI DP    
 5T2 NS NS p = 1 
 5T33 <10−4 <10−7 p = 1 
CD25 CD4+ SPFoxp3 CD4+ SPTCR MFI DP
Thymus weight    
 5T2 <0.05 <0.05 NS 
 5T33 <0.01 <0.005 <0.005 
DN proportions    
 5T2 <10−6 <10−5 <0.005 
 5T33 <10−4 <10−4 <10−11 
DP proportions    
 5T2 <0.0005 <0.005 <0.005 
 5T33 <10−6 <0.0005 <10−16 
CD4 proportions    
 5T2 NS <0.01 NS 
 5T33 <0.05 NS <0.05 
CD8 proportions    
 5T2 <0.01 <0.05 NS 
 5T33 NS NS NS 
TCR MFI DP    
 5T2 NS NS p = 1 
 5T33 <10−4 <10−7 p = 1 
a

p values of >0.05 are not significant.

A univariant regression analysis showed that the frequency of mature Treg development, measured by CD25 and Foxp3 expression, significantly correlated with thymus atrophy in both tumor models. As described above for TCR MFI, DN and DP proportions showed the highest correlation with Treg differentiation (Table I and Fig. 4,D). A stepwise multiple regression analysis showed that DP proportions independently predicted CD25 expression by CD4+ SP cells (R2 = 0.62, p < 0.0001), and DN proportions independently predicted Foxp3 expression (R2 = 0.54, p < 0.0001) in the 5T33 tumor. Analysis of 5T2 data revealed similar trends, where DN proportions and thymus atrophy were the dominant factors affecting CD25 and Foxp3 expression (R2 = 0.62 and R2 = 0.46, respectively; p < 0.0001). Correlations of thymus atrophy and subset proportions with CD25 at the DP stage were highly similar to correlation with mature Treg frequency (data not shown). Importantly, the proportion of the CD4+ SP subset was not markedly correlated with CD25+ and Foxp3+ frequency (Table I), in contrast to the possibility that the increase in Treg proportions resulted from recirculating Tregs from the periphery. Collectively, our data suggest that thymus atrophy and altered DN and DP populations are dominant variables affecting Treg developmental frequency as early as the DP stage. These results are in accordance with recent data suggesting the dominant effect of DP cells in the thymus on Treg fate adoption by early thymocytes in trans (36).

Increased ratios of Treg-to-Teff development in MM-inflicted mice could result from the induction of Treg differentiation. Alternatively, Treg increased ratios could result from increased resistance to apoptosis in the atrophied thymus, or from decreased survival of the SP CD25 (Teff) population. To discriminate between these possibilities, Treg (SP CD25+) and Teff (CD25) cells from control and atrophied thymuses of 5T2-inflicted mice were gated as described in Fig. 5,A (top) and apoptotic levels were tested by FACS. As shown in Fig. 5 A, apoptotic rates were negligible in both SP CD25+ and CD25 cells, in controls and in atrophied thymuses. Similarly, no apoptotic cells could be detected among CD25+ and CD25 cells at the DP stage (data not shown). These data did not result from the loss of apoptotic cells during the staining procedure, since a high percentage of apoptotic cells could be detected when gating on DN cells in both control and 5T2-inflicted mice. Therefore, differences in Treg and Teff survival were not responsible for increased Treg ratios in the atrophied thymuses.

FIGURE 5.

A, Representative cell cycle analysis of 7-aminoactinomycin D (7-ADD) CD4+ SP CD25+ (SP CD25+) cells, CD4+ SP CD25 (SP CD25) cells, and DN cells from a control and an atrophied thymus from a 5T2-inflicted mouse. Control (n = 4) and 5T2 (n = 7). Similar results were obtained in two independent experiments. B, Representative FACS analysis of the percentage of cycling (R3 indicates BrdU+) cells among CD4+ SP CD25+ cells (SP CD25+) and DP CD25+ cells from a control and a 5T2-inflicted mouse. Control (n = 4) and 5T2 (n = 5); one of two experiments with similar results.

FIGURE 5.

A, Representative cell cycle analysis of 7-aminoactinomycin D (7-ADD) CD4+ SP CD25+ (SP CD25+) cells, CD4+ SP CD25 (SP CD25) cells, and DN cells from a control and an atrophied thymus from a 5T2-inflicted mouse. Control (n = 4) and 5T2 (n = 7). Similar results were obtained in two independent experiments. B, Representative FACS analysis of the percentage of cycling (R3 indicates BrdU+) cells among CD4+ SP CD25+ cells (SP CD25+) and DP CD25+ cells from a control and a 5T2-inflicted mouse. Control (n = 4) and 5T2 (n = 5); one of two experiments with similar results.

Close modal

To evaluate whether increased Treg proliferation could be the cause of increased Treg ratios in the atrophied thymuses, thymocyte proliferation in normal and atrophied thymuses from 5T2-inflicted mice was tested 24 h after BrdU injection. As shown in Fig. 5,B, a similar percentage of cycling cells was found among SP CD25+ thymocytes in normal and 5T2-inflicted mice (2.9 ± 0.9 and 3.3 ± 0.6, respectively). A higher proliferation rate was found at the DP stage, but no difference was found between normal and 5T2-driven thymuses (21.4 ± 5.0 and 23.6 ± 3.7, respectively), as presented in Fig. 5 B. No cycling cells were detected among CD25 cells (data not shown). Taken toogether, these data suggest that increased Treg-to-Teff ratios in the thymus of 5T2-inflicted mice do not result from altered thymocyte survival or increased Treg proliferation. It is therefore suggested that differentiation processes that do not involve proliferation, such as the control of gene expression, might be involved.

Patients with MM commonly develop bone disease including bone pain, osteolytic lesions, pathologic fractures, and hypercalcemia. Bone destruction in MM results from asynchronous bone turnover. Normal osteoclasts are induced by osteoclast-activating factors produced by myeloma cells or other cells in the microenvironment, while this process is not accompanied by increased bone formation by osteobalsts (43, 44, 45).

The 5T2 myeloma model closely mimics the human disease, specifically by the development of bone lesions as a primary symptom of the disease (28, 33). The 5T2 cells localize primarily to the BM, replacing the normal BM cells and causing bone lesions. The mice develop hind limb paralysis as a result of spinal cord compression (28, 33).

To test whether Treg thymocytes in the thymus of MM-inflicted mice possessed regulatory activity for the tumor cells, we used an adoptive transfer assay as described by Deng et al. (46) to determine whether thymocytes from 5T2-inflicted mice could specifically support in vivo tumor progression. To this end, 8-wk-old mice were injected with 5T2 cells. After 42 days, mice received three doses (30 × 106 cells injected i.v. once a week) of thymocytes from paralyzed 5T2-inflicted mice, or from control healthy mice. A third group of 5T2-injcected mice was not further injected with thymocytes. Mice in the three groups developed hind limb paralysis at the same time, between 63 and 71 days after tumor cells injection, and were humanely killed and tested. Although there was no difference in the time of hind limb paralysis onset, a significant difference in the severity of disease symptoms was apparent between the three experimental groups: in the group receiving thymocytes from MM-inflicted mice, we noted severe bone destruction and massive tumor infiltration into the surrounding muscle in 8 of 10 mice, compared with 2 of 8 mice that had been injected with control thymocytes (Table II; sum of two independent experiments with similar results). In almost all cases, massive tumor growth was detected around the spine, as shown in Fig. 6,A. As shown in Fig. 6, A and B, tumor cells that originated in the bone caused severe fractures of the lumbar vertebra and invaded the skeletal muscle. This phenomenon was not observed in mice that had not received thymocyte injections, in contrast to 2 of 8 mice receiving normal thymocytes (Table II). This result probably reflects a regulatory activity of normal thymocytes. However, tumor-specific regulatory activity was apparently more abundant in thymuses from diseased mice compared with controls (Table II). Collectively, these results demonstrate that thymocyte administration from 5T2-inlicted mice led to increased severity of bone destruction by tumor cells in vivo.

Table II.

Effects of thymocyte administration on tumor progression in 5T2-injected mice

Treatment of Mice% Hind Limb Paralysis (63–71-day latency)% Bone Destruction and Invasion into Skeletal Muscle
3 × 105 MM i.v. 9/9 (100%) 0/9 
3 × 105 MM i.v. → 3 × 107 normal thymocytes 8/8 (100%) 2/8 (25%) 
3 × 105 MM i.v. → 3 × 107 thymocytes from paralyzed 5T2 10/10 (100%) 8/10 (80%) 
Treatment of Mice% Hind Limb Paralysis (63–71-day latency)% Bone Destruction and Invasion into Skeletal Muscle
3 × 105 MM i.v. 9/9 (100%) 0/9 
3 × 105 MM i.v. → 3 × 107 normal thymocytes 8/8 (100%) 2/8 (25%) 
3 × 105 MM i.v. → 3 × 107 thymocytes from paralyzed 5T2 10/10 (100%) 8/10 (80%) 

To test if thymocyte transfer was mediated by Treg thymocytes, Tregs from the thymus of 5T2-inflicted mice and controls were injected separately (150,000 per mouse as found in the 5T2-derived thymus), in addition to whole thymocytes as done in the former experiments. A mix containing all other 5T2-derived thymocytes excluding Tregs (“Treg depleted”) was injected to test whether any of these cells could mediate the effect caused by whole thymocyte administration. This mix contained purified DP, CD8+, SP, DN, and SP CD4+CD25 Teffs in the same numbers found in the 5T2-derived thymus, as described in Materials and Methods. A mix containing normal thymocyte populations in numbers similar to 5T2 was injected as a completely equivalent control.

As shown in Fig. 6,C, we found a large variance among recipients in the paralysis onset time, and the administration of whole thymocytes from 5T2-inflicted mice resulted in an earlier paralysis onset compared with controls. In contrast to the former experiments, severe bone lesions were found in the limbs and spine of all paralyzed mice in the two groups. Therefore, the effect of thymocyte administration from 5T2-inflicted mice was pronounced by the early onset of disease symptoms rather than the occurrence of severe bone fractions. Notably, the administration of 5T2-derived Tregs alone had a similar and stronger effect, compared with whole thymocytes from the same mice (p > 0.5), and this effect was significant compared with mice receiving control thymocytes (p < 0.01). These data indicated that Tregs alone could account for the tumor-progressive effect of 5T2-derived thymocytes. Importantly, the transfer 5T2-derived Treg-depleted thymocytes did not lead to early paralysis onset (p < 0.05 vs Tregs). Treg thymocytes from normal mice had an effect comparable to 5T2-derived Tregs, in accordance with results presented in Table II and discussed above. Taken together, these results indicate that Tregs, but not other thymocyte populations, mediate the tumor-progressive effect of thymocytes from 5T2-inflicted mice.

Although the thymus is the main site of Treg development, mechanisms leading to Treg expansion have been suggested to occur in the periphery (12, 13). Using two murine MM models, 5T2 and 5T33, we report that suppressive-functional Tregs accumulate in the spleens, LNs, BM, and PB of MM-inflicted mice. For the first time, we directly examined thymic Treg development in malignancy, and we report that the proportions of Tregs generated in the thymus of MM-inflicted mice were significantly increased. Similar data were collected from the two tumor models, which differ in localization and progression kinetics in the mouse (28). Although thymus cellularity was reduced, Treg numbers were not severely decreased in MM-inflicted mice, while Teff numbers were dramatically affected. Adoptive transfer of Tregs, but not other thymocytes from the thymus of 5T2-inflicted mice, led to increased progression of disease symptoms, strongly suggesting that Tregs in the thymus of MM-inflicted mice have an in vivo contribution to tolerance against MM cells.

Although Treg numbers did not increase in the thymus, Treg-to-Teff ratios significantly increased in the thymus of MM-inflicted mice due to dramatic Teff depletion. A similar phenomenon occurring in the thymus of pre-Tα-deficient mice leads to increased Treg-to-Teff ratios in the thymus as well as in the periphery (36). Importantly, the balance between Tregs and effectors has been stressed as critical for the decision between immune response and suppression (36, 47). Therefore, our data showing increased Treg-to-Teff proportions among mature thymocytes suggest an effect on immune balance in the periphery of MM-inflicted mice.

A previous study demonstrated increased levels of Tregs with a naive phenotype in MM patients. Peripheral expansion of naive nTregs was suggested as a main mechanism responsible for nTregs increase in MM, since peripheral proliferation of naive nTregs was implied in MM patients as well as in healthy individuals (22). Additionally, recent data point out that Treg induction occurs at the tumor site (12, 13, 14, 15, 16), and that Tregs are specifically recruited to the tumor site by CCL22 and CCL17 secreted by the tumor (2, 17). Thus, our results suggest a thymic contribution to increased Treg ratios among CD4+ cells as found in our models and in human MM patients (20), in addition to peripheral mechanisms reported to contribute to Treg accumulation at the tumor site. The relative contribution of thymic and peripheral mechanisms, as well as the relationships between different peripheral mechanisms such as Treg recruitment to the tumor site and Treg induction, should be further tested.

Thymus atrophy was demonstrated in 5T2- and 5T33-inflicted mice. Thymus atrophy was significantly associated with elevated ratios of Treg-expressing cells at the DP and SP stages and with TCR MFI at the DP stage in both tumors. However, DN and DP proportions were more strongly associated with Treg ratios and with DP TCR MFI in both tumors, and they were the most dominant variables predictive of Treg development ratios and of TCR MFI. It therefore seems that DN and DP proportions strongly affect thymocytes, starting at an early developmental stage. Recent data from studies in pre-Tα−/− mice have strongly suggested the dominant control of DP cells on gene expression and fate adoption by early thymocytes in trans (36, 48). Differentiation toward a Treg phenotype is negatively controlled by DP cells, leading to increased Treg differentiation in the pre-Tα−/− thymus, containing a severely reduced DP population, and to increased Treg peripheral levels (36). Based on these data, our data imply that an altered DP-to-DN proportion may dominantly affect Treg ratios in the atrophied thymuses, and may mediate the effects of thymus atrophy on Treg development in MM. Due to the strong association between DN and DP frequencies, we cannot establish whether both populations affect Treg development. Our data also suggest that Treg induction, rather than resistance to apoptosis or increased proliferation, is responsible to Treg increased development in the atrophied thymus.

Notably, thymus atrophy was reported in human and mouse malignancies, and a change of subpopulation proportions similar to our data was found in a mouse mammary tumor model (49). Additionally, increased concentrations of vascular endothelial growth factor (VEGF), a factor secreted by various tumors including MM, were shown to cause thymus atrophy with similar thymic subset changes (50). We found a strong association between Treg developmental ratios and thymus atrophy as well as altered proportions of the DN and DP subpopulations in both tumor models. It may therefore be of importance to test whether increased Treg proportions might be found in other cancer types demonstrating thymus atrophy.

It is unclear how proportions of SP cells were generated in the thymus of MM-inflicted mice, while their DP progenitors were dramatically reduced. Notably, decreased DP proportions accompanied by increased SP proportions were reported in atrophied thymuses in a mammary tumor model as well as in VEGF-injected mice (49, 50, 51). However, no mechanisms were offered to explain this phenomenon. In the present study, we did not find differences in proliferation levels or apoptotic rates of 5T2 thymocyes in the DN, DP, or SP developmental stages (Fig. 5 and accompanying text). We therefore conclude that neither decreased apoptosis of SP cells nor increased SP proliferation could explain increased SP ratios relative to their DP progenitors. Since DP cells in the thymus of MM-inflicted mice present a premature phenotype, as reflected in increased CD69 and TCR expression (Fig. 4), we think it possible that a change in differentiation kinetics in the thymus of MM-inflicted mice accelerates the DP-to-SP transition, leading to relatively increased SP-to-DP proportions.

Recent data have indicated that considerable amounts of Tregs recirculate to the thymus (52, 53), and therefore it could be claimed that increased Treg levels in the thymus of MM-inflected mice might reflect Treg recirculation from the periphery. However, while recirculating cells are mature SP cells (52, 53), increased prevalence of CD4+Foxp3+ cells in the thymus of MM-inflicted mice is demonstrated as early as the DP stage. Additionally, one would expect a major correlation between the proportion of CD4+ SP cells and Treg ratios, in the case that Treg increased levels resulted from mature Treg recirculation. However, only a weak correlation was found, while a marked correlation was found between Treg ratios and the proportions of other thymic subsets (DN and DP). Most importantly, Tregs in the thymus of 5T2-inflicted mice retained a naive phenotype, while Tregs in the periphery showed a dramatic decrease of naive expression (Fig. 3 D). We therefore conclude that the data presented herein reflect true change in Treg development in the thymus of MM-inflicted mice.

Conflicting results concerning Treg levels and function were reported from studies in B cell malignancies (23, 24, 25) and specifically in MM patients (10, 26). Our data showed increased ratios of functional Tregs that correlated with disease progression, similar to findings in MM patients and patients with other B cell malignancies (22, 25). Notably, it seems that differences in research strategies may be responsible for the contradictory data in the area of B cell malignancies. Since the balance between Tregs and Teffs is critical for antitumor response, and since the prevalence of other cells is altered in the blood and tumor tissues in disease, we and others (22, 25) consider Treg ratios among CD4+ cells to be a relevant measure for Treg prevalence in MM and other B cell malignancies.

New approaches in the area of cancer immunotherapy are aimed at neutralizing Tregs in the periphery before or in combination with other cancer immunotherapies (12). Based on the results presented herein, it is of concern that eliminating Tregs in the periphery while thymic Treg development is imbalanced would not lead to a stable Treg-to-Teff balance in the long term. Importantly, VEGF neutralization reversed tumor-associated thymus atrophy in mice (50). Our data suggest that reducing thymic Treg development in malignancy, through tumor load reduction or the neutralization of specific factors such as VEGF, may improve the results of Treg depletion strategies and immunotherapy in both the short and long term.

We thank J. Radl for his kind help in providing information concerning the 5T MM tumor and in enabling us to use the 5T MM tumor, and to J. A. Levy for helpful suggestions for improvement of the manuscript. We also thank Varda Segal for her help in growing the 5T33 tumor cell line.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by a grant (to D.Z.) from the Israel Cancer Research Fund and by a grant from the Wolfson Family Charitable Trust, London, on tumor cell diversity.

3

Abbreviations used in this paper: BM, bone marrow; DN, double negative; DP, double positive; GITR, glucocorticoid-induced TNF receptor; LN, lymph nodes; MFI, mean fluorescence intensity; MM, multiple myeloma; nTreg, naturally occurring Treg; PB, peripheral blood; SP, single positive; Teff, effector T cell; Treg, regulatory T cell; VEGF, vascular endothelial growth factor.

1
Beyer, M., J. L. Schultze.
2006
. Regulatory T cells in cancer.
Blood
108
:
804
-811.
2
Curiel, T. J., S. Wei, H. Dong, X. Alvarez, P. Cheng, P. Mottram, R. Krzysiek, K. L. Knutson, B. Daniel, M. C. Zimmermann, et al
2003
. Blockade of B7-H1 improves myeloid dendritic cell-mediated antitumor immunity.
Nat. Med.
9
:
562
-567.
3
Liyanage, U. K., T. T. Moore, H. G. Joo, Y. Tanaka, V. Herrmann, G. Doherty, J. A. Drebin, S. M. Strasberg, T. J. Eberlein, P. S. Goedegebuure, D. C. Linehan.
2002
. Prevalence of regulatory T cells is increased in peripheral blood and tumor microenvironment of patients with pancreas or breast adenocarcinoma.
J. Immunol.
169
:
2756
-2761.
4
Marshall, N. A., L. E. Christie, L. R. Munro, D. J. Culligan, P. W. Johnston, R. N. Barker, M. A. Vickers.
2004
. Immunosuppressive regulatory T cells are abundant in the reactive lymphocytes of Hodgkin lymphoma.
Blood
103
:
1755
-1762.
5
Ormandy, L. A., T. Hillemann, H. Wedemeyer, M. P. Manns, T. F. Greten, F. Korangy.
2005
. Increased populations of regulatory T cells in peripheral blood of patients with hepatocellular carcinoma.
Cancer Res.
65
:
2457
-2464.
6
Onizuka, S., I. Tawara, J. Shimizu, S. Sakaguchi, T. Fujita, E. Nakayama.
1999
. Tumor rejection by in vivo administration of anti-CD25 (interleukin-2 receptor α) monoclonal antibody.
Cancer Res.
59
:
3128
-3133.
7
Shimizu, J., S. Yamazaki, S. Sakaguchi.
1999
. Induction of tumor immunity by removing CD25+CD4+ T cells: a common basis between tumor immunity and autoimmunity.
J. Immunol.
163
:
5211
-5218.
8
Sutmuller, R. P., L. M. van Duivenvoorde, A. van Elsas, T. N. Schumacher, M. E. Wildenberg, J. P. Allison, R. E. Toes, R. Offringa, C. J. Melief.
2001
. Synergism of cytotoxic T lymphocyte-associated antigen 4 blockade and depletion of CD25+ regulatory T cells in antitumor therapy reveals alternative pathways for suppression of autoreactive cytotoxic T lymphocyte responses.
J. Exp. Med.
194
:
823
-832.
9
Turk, M. J., J. A. Guevara-Patino, G. A. Rizzuto, M. E. Engelhorn, S. Sakaguchi, A. N. Houghton.
2004
. Concomitant tumor immunity to a poorly immunogenic melanoma is prevented by regulatory T cells.
J. Exp. Med.
200
:
771
-782.
10
Bates, G. J., S. B. Fox, C. Han, R. D. Leek, J. F. Garcia, A. L. Harris, A. H. Banham.
2006
. Quantification of regulatory T cells enables the identification of high-risk breast cancer patients and those at risk of late relapse.
J. Clin. Oncol.
24
:
5373
-5380.
11
Petersen, R. P., M. J. Campa, J. Sperlazza, D. Conlon, M. B. Joshi, D. H. Harpole, Jr, E. F. Patz, Jr.
2006
. Tumor infiltrating Foxp3+ regulatory T-cells are associated with recurrence in pathologic stage I NSCLC patients.
Cancer
107
:
2866
-2872.
12
Zou, W..
2005
. Immunosuppressive networks in the tumour environment and their therapeutic relevance.
Nat. Rev. Cancer
5
:
263
-274.
13
Kim, R., M. Emi, K. Tanabe.
2006
. Cancer immunosuppression and autoimmune disease: beyond immunosuppressive networks for tumour immunity.
Immunology
119
:
254
-264.
14
Jarnicki, A. G., J. Lysaght, S. Todryk, K. H. Mills.
2006
. Suppression of antitumor immunity by IL-10 and TGF-β-producing T cells infiltrating the growing tumor: influence of tumor environment on the induction of CD4+ and CD8+ regulatory T cells.
J. Immunol.
177
:
896
-904.
15
Larmonier, N., M. Marron, Y. Zeng, J. Cantrell, A. Romanoski, M. Sepassi, S. Thompson, X. Chen, S. Andreansky, E. Katsanis.
2007
. Tumor-derived CD4+CD25+ regulatory T cell suppression of dendritic cell function involves TGF-β and IL-10.
Cancer Immunol. Immunother.
56
:
48
-59.
16
Liyanage, U. K., P. S. Goedegebuure, T. T. Moore, C. T. Viehl, T. A. Moo-Young, J. W. Larson, D. M. Frey, J. P. Ehlers, T. J. Eberlein, D. C. Linehan.
2006
. Increased prevalence of regulatory T cells (Treg) is induced by pancreas adenocarcinoma.
J. Immunother.
29
:
416
-424.
17
Mizukami, Y., K. Kono, Y. Kawaguchi, H. Akaike, K. Kamimura, H. Sugai, H. Fujii.
2008
. CCL17 and CCL22 chemokines within tumor microenvironment are related to accumulation of Foxp3+ regulatory T cells in gastric cancer.
Int. J. Cancer
122
:
2286
-2293.
18
Itoh, M., T. Takahashi, N. Sakaguchi, Y. Kuniyasu, J. Shimizu, F. Otsuka, S. Sakaguchi.
1999
. Thymus and autoimmunity: production of CD25+CD4+ naturally anergic and suppressive T cells as a key function of the thymus in maintaining immunologic self-tolerance.
J. Immunol.
162
:
5317
-5326.
19
Kim, J. M., J. P. Rasmussen, A. Y. Rudensky.
2007
. Regulatory T cells prevent catastrophic autoimmunity throughout the lifespan of mice.
Nat. Immunol.
8
:
191
-197.
20
Sakaguchi, S..
2005
. Naturally arising Foxp3-expressing CD25+CD4+ regulatory T cells in immunological tolerance to self and non-self.
Nat. Immunol.
6
:
345
-352.
21
Shevach, E. M..
2000
. Regulatory T cells in autoimmmunity.
Annu. Rev. Immunol.
18
:
423
-449.
22
Beyer, M., M. Kochanek, T. Giese, E. Endl, M. R. Weihrauch, P. A. Knolle, S. Classen, J. L. Schultze.
2006
. In vivo peripheral expansion of naive CD4+CD25high FoxP3+ regulatory T cells in patients with multiple myeloma.
Blood
107
:
3940
-3949.
23
Alvaro, T., M. Lejeune, M. T. Salvado, R. Bosch, J. F. Garcia, J. Jaen, A. H. Banham, G. Roncador, C. Montalban, M. A. Piris.
2005
. Outcome in Hodgkin’s lymphoma can be predicted from the presence of accompanying cytotoxic and regulatory T cells.
Clin. Cancer Res.
11
:
1467
-1473.
24
Carreras, J., A. Lopez-Guillermo, B. C. Fox, L. Colomo, A. Martinez, G. Roncador, E. Montserrat, E. Campo, A. H. Banham.
2006
. High numbers of tumor-infiltrating FOXP3-positive regulatory T cells are associated with improved overall survival in follicular lymphoma.
Blood
108
:
2957
-2964.
25
Yang, Z. Z., A. J. Novak, M. J. Stenson, T. E. Witzig, S. M. Ansell.
2006
. Intratumoral CD4+CD25+ regulatory T-cell-mediated suppression of infiltrating CD4+ T cells in B-cell non-Hodgkin lymphoma.
Blood
107
:
3639
-3646.
26
Prabhala, R. H., P. Neri, J. E. Bae, P. Tassone, M. A. Shammas, C. K. Allam, J. F. Daley, D. Chauhan, E. Blanchard, H. S. Thatte, et al
2006
. Dysfunctional T regulatory cells in multiple myeloma.
Blood
107
:
301
-304.
27
Radl, J., J. W. Croese, C. Zurcher, M. H. Van den Enden-Vieveen, A. M. de Leeuw.
1988
. Animal model of human disease: multiple myeloma.
Am. J. Pathol.
132
:
593
-597.
28
Vanderkerken, K., H. De Raeve, E. Goes, S. Van Meirvenne, J. Radl, I. Van Riet, K. Thielemans, B. Van Camp.
1997
. Organ involvement and phenotypic adhesion profile of 5T2 and 5T33 myeloma cells in the C57BL/KaLwRij mouse.
Br. J. Cancer
76
:
451
-460.
29
Koike, M., I. Sekigawa, M. Okada, M. Matsumoto, N. Iida, H. Hashimoto, K. Oshimi.
2002
. Relationship between CD4+/CD8+ T cell ratio and T cell activation in multiple myeloma: reference to IL-16.
Leuk. Res.
26
:
705
-711.
30
Piccirillo, C. A., E. M. Shevach.
2001
. Cutting edge: Control of CD8+ T cell activation by CD4+CD25+ immunoregulatory cells.
J. Immunol.
167
:
1137
-1140.
31
Thornton, A. M., E. M. Shevach.
2000
. Suppressor effector function of CD4+CD25+ immunoregulatory T cells is antigen nonspecific.
J. Immunol.
164
:
183
-190.
32
Kipnis, J., M. Cardon, H. Avidan, G. M. Lewitus, S. Mordechay, A. Rolls, Y. Shani, M. Schwartz.
2004
. Dopamine, through the extracellular signal-regulated kinase pathway, downregulates CD4+CD25+ regulatory T-cell activity: implications for neurodegeneration.
J. Neurosci.
24
:
6133
-6143.
33
Dingli, D., S. J. Russell.
2007
. Mouse models and the RANKL/OPG axis in myeloma bone disease.
Leukemia
21
:
2090
-2093.
34
Bayer, A. L., A. Yu, T. R. Malek.
2007
. Function of the IL-2R for thymic and peripheral CD4+CD25+ Foxp3+ T regulatory cells.
J. Immunol.
178
:
4062
-4071.
35
Cabarrocas, J., C. Cassan, F. Magnusson, E. Piaggio, L. Mars, J. Derbinski, B. Kyewski, D. A. Gross, B. L. Salomon, K. Khazaie, et al
2006
. Foxp3+ CD25+ regulatory T cells specific for a neo-self-antigen develop at the double-positive thymic stage.
Proc. Natl. Acad. Sci. USA
103
:
8453
-8458.
36
Pennington, D. J., B. Silva-Santos, T. Silberzahn, M. Escorcio-Correia, M. J. Woodward, S. J. Roberts, A. L. Smith, P. J. Dyson, A. C. Hayday.
2006
. Early events in the thymus affect the balance of effector and regulatory T cells.
Nature
444
:
1073
-1077.
37
Raitakari, M., R. D. Brown, J. Gibson, D. E. Joshua.
2003
. T cells in myeloma.
Hematol. Oncol.
21
:
33
-42.
38
Apostolou, I., A. Sarukhan, L. Klein, H. von Boehmer.
2002
. Origin of regulatory T cells with known specificity for antigen.
Nat. Immunol.
3
:
756
-763.
39
Jordan, M. S., A. Boesteanu, A. J. Reed, A. L. Petrone, A. E. Holenbeck, M. A. Lerman, A. Naji, A. J. Caton.
2001
. Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide.
Nat. Immunol.
2
:
301
-306.
40
Lerman, M. A., J. Larkin, III, C. Cozzo, M. S. Jordan, A. J. Caton.
2004
. CD4+ CD25+ regulatory T cell repertoire formation in response to varying expression of a neo-self-antigen.
J. Immunol.
173
:
236
-244.
41
Brandle, D., S. Muller, C. Muller, H. Hengartner, H. Pircher.
1994
. Regulation of RAG-1 and CD69 expression in the thymus during positive and negative selection.
Eur. J. Immunol.
24
:
145
-151.
42
Suzuki, H., J. A. Punt, L. G. Granger, A. Singer.
1995
. Asymmetric signaling requirements for thymocyte commitment to the CD4+ versus CD8+ T cell lineages: a new perspective on thymic commitment and selection.
Immunity
2
:
413
-425.
43
Callander, N. S., G. D. Roodman.
2001
. Myeloma bone disease.
Semin. Hematol.
38
:
276
-285.
44
Terpos, E., O. Sezer, P. Croucher, M. A. Dimopoulos.
2007
. Myeloma bone disease and proteasome inhibition therapies.
Blood
110
:
1098
-1104.
45
Yeh, H. S., J. R. Berenson.
2006
. Myeloma bone disease and treatment options.
Eur. J. Cancer
42
:
1554
-1563.
46
Deng, S., D. J. Moore, X. Huang, M. Mohiuddin, M. K. t. Lee, E. Velidedeoglu, M. M. Lian, M. Chiaccio, S. Sonawane, A. Orlin, et al
2006
. Antibody-induced transplantation tolerance that is dependent on thymus-derived regulatory T cells.
J. Immunol.
176
:
2799
-2807.
47
Belkaid, Y., B. T. Rouse.
2005
. Natural regulatory T cells in infectious disease.
Nat. Immunol.
6
:
353
-360.
48
Silva-Santos, B., D. J. Pennington, A. C. Hayday.
2005
. Lymphotoxin-mediated regulation of γδ cell differentiation by αβ T cell progenitors.
Science
307
:
925
-928.
49
Lopez, D. M., V. Charyulu, B. Adkins.
2002
. Influence of breast cancer on thymic function in mice.
J. Mammary Gland Biol. Neoplasia
7
:
191
-199.
50
Ohm, J. E., D. I. Gabrilovich, G. D. Sempowski, E. Kisseleva, K. S. Parman, S. Nadaf, D. P. Carbone.
2003
. VEGF inhibits T-cell development and may contribute to tumor-induced immune suppression.
Blood
101
:
4878
-4886.
51
Adkins, B., V. Charyulu, Q. L. Sun, D. Lobo, D. M. Lopez.
2000
. Early block in maturation is associated with thymic involution in mammary tumor-bearing mice.
J. Immunol.
164
:
5635
-5640.
52
Bosco, N., F. Agenes, A. G. Rolink, R. Ceredig.
2006
. Peripheral T cell lymphopenia and concomitant enrichment in naturally arising regulatory T cells: the case of the pre-Tα gene-deleted mouse.
J. Immunol.
177
:
5014
-5023.
53
Zhan, Y., D. Bourges, J. A. Dromey, L. C. Harrison, A. M. Lew.
2007
. The origin of thymic CD4+CD25+ regulatory T cells and their co-stimulatory requirements are determined after elimination of recirculating peripheral CD4+ cells.
Int. Immunol.
19
:
455
-463.