Regulatory T cells are expanded by Teriparatide treatment in humans and mediate intermittent PTH‐induced bone anabolism in mice

Abstract

Teriparatide is a bone anabolic treatment for osteoporosis, modeled in animals by intermittent PTH (iPTH) administration, but the cellular and molecular mechanisms of action of iPTH are largely unknown. Here, we show that Teriparatide and iPTH cause a ~two‐threefold increase in the number of regulatory T cells (Tregs) in humans and mice. Attesting in vivo relevance, blockade of the Treg increase in mice prevents the increase in bone formation and trabecular bone volume and structure induced by iPTH. Therefore, increasing the number of Tregs is a pivotal mechanism by which iPTH exerts its bone anabolic activity. Increasing Tregs pharmacologically may represent a novel bone anabolic therapy, while iPTH‐induced Treg increase may find applications in inflammatory conditions and transplant medicine.

Introduction

Primary hyperparathyroidism is a common cause of bone loss and fractures due to the continuous production of high levels of parathyroid hormone (PTH) by the parathyroid glands 1, 2. By contrast, when PTH is injected daily, a regimen known as intermittent PTH (iPTH) treatment, the hormone increases bone volume and strength due to a stimulation of bone formation tempered by a more moderate increase in resorption 3, 4. As a result, intermittent treatment with the 1–34 fragment of PTH is an FDA‐approved treatment modality for postmenopausal osteoporosis (PMO) 5.

PTH acts by binding to the PTH/PTHrP receptor PPR, which is expressed in all osteoblastic cells, including stromal cells (SCs), osteoblasts, and osteocytes 6, 7, 8, 9, 10. Moreover, PPR is expressed in conventional CD4⁺ and CD8⁺ T cells 11 and macrophages 12. iPTH stimulates bone formation by increasing osteoblast formation and life span. Activation of Wnt signaling in osteocytes and osteoblasts is one of the proposed mechanisms by which iPTH stimulates bone formation 4, 13, 14. Wnt signaling activation is achieved through multiple mechanisms including Wnt ligand‐independent activation of Wnt coreceptors 15, blunted osteocytic production of the Wnt inhibitor sclerostin 16, 17, 18, and decreased production by osteoblasts of the Wnt inhibitor Dkk1 19.

While osteocytes and their production of sclerostin are critical for the activity of iPTH, part of this hormone activity is sclerostin independent 20 and mediated by T cells 21, a cell lineage that potentiates the anabolic activity of iPTH in trabecular bone 11, 20, 22. Accordingly, iPTH fails to stimulate bone formation and increase bone mass in T‐cell null mice 11. By contrast, the effects of iPTH in cortical bone are completely T cell independent 11, 20, 22, likely due to the fact that T cells have no contacts with periosteal surfaces and have limited capacity to communicate with osteocytes. Among the T cells required for iPTH to exert its full anabolic activity are bone marrow (BM) CD8⁺ T cells 11. CD8⁺ T cells express higher levels of the receptor PPR than CD4⁺ T cells 11, 20, 22. Moreover, BM CD8⁺ T cells, but not CD4⁺ T cells, respond to iPTH by releasing Wnt10b 11, 20, 22, an osteogenic Wnt ligand that activates Wnt signaling in osteoblastic cells 23.

Regulatory T cells (Tregs) are a suppressive population of predominantly CD4⁺ T cells that play a critical role in maintaining immune tolerance and immune homeostasis. Tregs are comprised of thymus‐derived Tregs (tTregs, also known as nTregs) and peripherally derived Tregs (pTregs, also known as iTregs) 24. Tregs are defined by the expression of the transcription factor Foxp3 and the ability to block inflammatory diseases and maintain immune homeostasis and tolerance 25. Accordingly, defects in Treg numbers and/or activity have been implicated in several chronic inflammatory diseases. Moreover, Tregs have furthermore been found blunt bone resorption 26, 27, prevent ovariectomy‐induced bone loss 28, and regulate osteoclast formation 26, 29, 30.

In vitro, conventional CD4⁺ T cells differentiate into Tregs by TCR stimulation under the influence of TGFβ and IL‐2 31, 32, 33. Recently, IGF‐1 has been recognized as an additional inducer of Tregs 34, 35. Since iPTH increases TGFβ and IGF‐1 production in bone 36, 37, 38, it is likely that iPTH may induce and/or expand BM Tregs.

This study was designed to investigate the effects of iPTH on Treg formation and activity in humans and mice, and to determine whether Tregs play a role in the bone anabolic activity of iPTH in mice. We report that treatment with iPTH increases the number of Tregs in humans and mice. In rodents, an increase in the number of Tregs is required for iPTH to exert its bone anabolic activity.

Results

Teriparatide treatment increases the number of Tregs in human peripheral blood

The PTH fragment Teriparatide is the only approved bone anabolic treatment for osteoporosis but its intricate mechanism of action remains largely unknown. Among the pleiotropic effects of PTH is the capacity to increase the production of TGFβ1 and IGF‐1 by human osteoblasts 39, 40, 41, factors which induce Treg differentiation. To investigate whether Teriparatide regulates the number of Tregs in humans, 40 Italian women afflicted by PMO of similar age and years since menopause were enrolled in a 6‐month‐long prospective clinical trial. Twenty of the 40 women were treated with calcium and vitamin D (control treatment), while the remaining 20 women were treated with calcium, vitamin D, and human PTH 1‐34 (Teriparatide), a treatment modality referred to hereafter as Teriparatide treatment. The baseline demographic characteristics of the study population and the serum levels of calcium, PTH, and 25‐hydroxy vitamin D are shown in Table 1. Peripheral blood mononuclear cells (PBMCs) were obtained at baseline and 3 and 6 months of treatment. Analysis by flow cytometry revealed that Teriparatide treatment increased the absolute and relative number of Tregs (CD4⁺CD25⁺Foxp3⁺ cells) in human PBMC at 3 and 6 months of treatment, compared to baseline (Fig 1A and B). By contrast, treatment with calcium and vitamin D did not alter the number of Tregs during the 6 months of the study. As a result, both at 3 and 6 months the absolute and relative number of Tregs in PBMC was higher in women treated with Teriparatide than in those in the calcium and vitamin D control group.

Table 1.

Demographic and clinical data of patients with postmenopausal osteoporosis treated with calcium and vitamin D, or calcium and vitamin D and teriparatide

	Control	Teriparatide	P
n	20	20
Age	68.5 ± 1.8	69.7 ± 1.6	0.535
Years since menopause	18.4 ± 2.1	20.8 ± 1.6	0.232
Ca (mg/dl) [8.8–10.4 mg/dl]	9.6 ± 0.1	9.5 ± 0.1	0.851
Serum P (mg/dl) [2.5–4.48 mg/dl]	3.5 ± 0.2	3.4 ± 0.1	0.608
PTH (pg/ml) [10–65 pg/ml]	42.8 ± 9.3	46.8 ± 4.1	0.272
25OH vitamin D (ng/ml) [20–100 ng/ml]	30.2 ± 3.1	27.3 ± 2.7	0.997

Data are shown as mean ± SEM, and P values were calculated by unpaired t‐test. Values in squared parenthesis denote normal range.

Figure 1. Teriparatide treatment in humans increases the absolute and relative number of Tregs in peripheral blood and TGFβ expression by Tregs.

1. Relative frequency of Tregs in PBMC at 3 and 6 months of treatment. n = 20 patients per group.
2. Absolute frequency of Tregs in PBMC at 3 and 6 months of treatment. n = 20 patients per group.
3. mRNA levels of TGFβ1 in peripheral blood CD4⁺CD25⁺ T cells at 3 and 6 months of treatment. n = 9 patients per group.
4. Relative frequency of Tregs in cultures of peripheral blood CD4⁺CD25⁺ T cells stimulated with vehicle or PTH.
5. mRNA levels of TGFβ1 in peripheral blood CD4⁺CD25⁺ T cells stimulated with vehicle or PTH.

Data information: All data are expressed as mean ± SEM. All data were normally distributed according to the Shapiro–Wilk normality test and analyzed by two‐way ANOVA for repeated measures. *P < 0.05, **P < 0.01, and ***P < 0.001 compared to baseline. ^# P < 0.05 and ^## P < 0.01, compared to Ca + D group.

TGFβ1 not only is a critical inducer of Treg differentiation, but is also an important product of Tregs that contributes to suppress effector T cells in vivo 25, 42. Production of TGFβ1 by Tregs is therefore an indicator of Treg function. To determine if Teriparatide regulates the function of Tregs, we measured the level of TGFβ1 mRNA in sorted peripheral blood CD4⁺CD25⁺ T cells from the last 18 women enrolled in the trial. We selected this cell population because most CD4⁺CD25⁺ T cells are Foxp3⁺ Tregs, while measurements of TGFβ1 mRNA in Tregs sorted by Foxp3 staining is not feasible due to the loss of cell viability caused by intracellular staining. We found that the level of TGFβ1 mRNA in this Treg‐enriched population was higher in the Teriparatide‐treated group than in the control group (Fig 1C). Post hoc analysis showed that TGFβ1 mRNA levels were higher at 3 months than at baseline in the Teriparatide‐treated group. Moreover, at 6 months TGFβ1 mRNA levels were higher in the Teriparatide‐treated group than in the control group. Together, these findings demonstrate that Teriparatide treatment expands the number and the activity of circulating human Tregs.

To determine whether Teriparatide targets Tregs directly or indirectly, human peripheral blood CD4⁺CD25⁺ T cells were stimulated in vitro with anti‐CD3 Ab and IL‐2 for 6 days. Vehicle or Teriparatide were added every 2 days for 1 or 24 h. Analysis by flow cytometry revealed that Teriparatide did not increase the relative number of Tregs (CD4⁺CD25⁺Foxp3⁺ cells) in cultures of peripheral blood CD4⁺CD25⁺ T cells (Fig 1D), suggesting that Teriparatide regulates the number of Tregs via indirect mechanisms. We also found that in vitro stimulation with Teriparatide does not increase the expression of TGFβ mRNA in sorted peripheral blood CD4⁺CD25⁺ T cells (Fig 1E), confirming that Teriparatide does not directly targets Tregs.

iPTH treatment in mice expands the pool of BM Tregs by increasing Treg differentiation

As in human cells, in vitro PTH stimulation increases in production of TGFβ1 and IGF‐1 by murine osteoblasts 36, 37, 38. However, the effects of iPTH treatment on the production of these factors are largely unknown. To investigate this matter, 6‐week‐old mice were treated with vehicle or iPTH for 2 weeks. BM was then harvested and cultured for 24 h. ELISAs revealed that iPTH significantly increases the levels of TGFβ1 and IGF‐1 in the whole BM culture media (Appendix Fig S1A and B), suggesting that iPTH may regulate Treg differentiation.

To investigate the effect of iPTH treatment on the number of BM Tregs, 6‐week‐old mice were treated with vehicle or iPTH for 1, 2, or 4 weeks. BM was then harvested and stained for TCRβ, CD4, and Foxp3. iPTH treatment increased the relative and absolute number of BM Tregs (TCRβ⁺CD4⁺Foxp3⁺ cells) during the entire study period (Fig 2A and B). The increase in absolute number of BM Tregs was already significant at 1 week of iPTH treatment, peaked at 2 weeks, and remained significantly increased at 4 weeks of treatment. By contrast, iPTH did not increase the number of splenic, thymic, and intestinal Tregs (Appendix Fig S2A–E).

Figure 2. iPTH treatment increases the number of BM Tregs and Treg differentiation.

To determine the effect of treatment on the number of Tregs, BM TCRβ⁺CD4⁺Foxp3⁺ T cells were counted by flow cytometry following 1, 2, and 4 weeks of treatment with vehicle or iPTH. To assess Treg differentiation, conventional CD4⁺ T cells (TCRβ⁺CD4⁺eGFP⁻) from B6.Foxp3.eGFP reporter mice were transferred into TCRβ^−/− mice. Recipient mice were treated with vehicle or iPTH for 1, 2, or 4 weeks starting 2 weeks after the T‐cell transfer. The number of BM TCRβ⁺CD4⁺eGFP⁺ cells was then determined by flow cytometry.

1. Relative frequency of BM Tregs at 1, 2, and 4 weeks of treatment. n = 10–32 mice per group.
2. Absolute frequency of Tregs at 1, 2, and 4 weeks of treatment. n = 10–32 mice per group.
3. Relative frequency of eGFP⁺ Tregs at 1, 2, and 4 weeks of treatment n = 10 mice per group.
4. Absolute frequency of eGFP⁺ Tregs at 1, 2, and 4 weeks of treatment n = 10 mice per group.

Data information: Data are expressed as mean ± SEM. All data were normally distributed according to the Shapiro–Wilk normality test and analyzed by unpaired t‐tests. ***P < 0.001, and ****P < 0.0001 compared to the corresponding vehicle.

iPTH could expand the pool of BM Tregs via multiple mechanisms including increasing the differentiation of conventional CD4⁺ T cells into Tregs or the proliferation of Tregs within the BM. To gain mechanistic insights, we determined the effects of iPTH on Treg differentiation, which is defined as the induction of Foxp3 expression in CD4⁺Foxp3⁻ T cells 25. For this purpose, we made use of B6.Foxp3.eGFP reporter mice, a strain in which eGFP expression is co‐expressed with Foxp3 and restricted to CD4⁺ T cells. Conventional CD4⁺ T cells (CD4⁺eGFP⁻) were FACS‐sorted from the spleens of B6.Foxp3.eGFP reporter mice and transferred into TCRβ^−/− mice, a strain lacking αβ T cells. After 2 weeks, a length of time sufficient for the engraftment and expansion of donor T cells, recipient mice were treated with vehicle or iPTH for 1–4 weeks. We then determined the number of CD4⁺eGFP⁺ cells in the BM by flow cytometry. Treatment with iPTH increased the relative and the absolute number of CD4⁺eGFP⁺ cell in the BM at 1, 2, and 4 weeks of treatment (Fig 2C and D), demonstrating that iPTH increases the differentiation of BM Tregs. Additional studies that used BrdU incorporation to measure proliferation revealed that iPTH treatment for 1, 2, or 4 weeks does not increase BM Treg proliferation (Appendix Fig S2F). In an additional set of experiments, Tregs (CD4⁺eGFP⁺ cells) were FACS‐sorted from the spleens of B6.Foxp3.eGFP reporter mice and transferred into TCRβ^−/− mice. Recipient mice were treated with vehicle or iPTH for 2 weeks, starting the day of the Treg transfer. This design was selected to minimize the confounding effect of the partial loss of Foxp3 expression by CD4⁺eGFP⁺ cells, which may occur after Treg transfer into lymphopenic host mice 43. These studies revealed that iPTH does not affect the relative and the absolute number of CD4⁺eGFP⁺ cells residing in the BM and the spleen (Appendix Fig S2G–J). These findings, together with a lack of an effect of iPTH on Treg proliferation, indicate that iPTH does not alter the homing of Tregs to the spleen and the BM.

In addition to increasing the BM levels of TGFβ and IGF‐1 36, 37, 38, iPTH enhances the sensitivity of conventional CD4⁺ cells to TGFβ. This was disclosed by experiments in which unstimulated splenic CD4⁺CD25⁻ cells purified from iPTH‐treated mice were found to express lower levels of the negative regulator of TGFβ signaling SMAD7 as compared to CD4⁺CD25⁻ cells from vehicle‐treated mice (Fig 3A). To ascertain the functional relevance of this finding, splenic CD4⁺CD25⁻ cells from vehicle‐ or iPTH‐treated mice were stimulated in vitro with CD3/CD28 Ab, IL‐2, and TGFβ at 2.5 ng/ml for 72 h to induce their differentiation into Tregs 44. Measurements of phosphorylated SMAD2 and SMAD3 (pSMAD2 and pSMAD3) at the end of the culture period revealed higher concentrations of pSMAD2 and pSMAD3 (Fig 3B) in cells from iPTH‐treated mice as compared to those from control mice, suggesting that conventional CD4⁺ T cells from iPTH‐treated mice have a higher sensitivity to TGFβ. To confirm this hypothesis, splenic CD4⁺CD25⁻ cells were purified from vehicle‐ or iPTH‐treated mice and then cultured in vitro for 72 h with anti CD3/CD28 Ab, IL‐2, and increasing doses of TGFβ (0.1–5 ng/ml). At each dose of TGFβ, cultures of CD4⁺CD25⁻ T cells from iPTH‐treated mice yielded a greater percentage of Foxp3⁺ Tregs than those from vehicle‐treated mice (Fig 3C). Together, these findings indicate that CD4⁺ T cells from iPTH‐treated mice possess a greater sensitivity to TGFβ, which results in enhanced differentiation of CD4⁺ T cells into Tregs.

Figure 3. iPTH treatment increases the sensitivity of conventional CD4⁺ T cells to TGFβ.

1. Western blotting analysis of SMAD7 levels in unstimulated splenic conventional CD4⁺ cells. Data are from 1 representative experiment of a total of four experiments. R.I., relative intensity.
2. Western blotting analysis of pSMAD2 and pSMAD3 levels in splenic conventional CD4⁺ T cells stimulated with anti‐CD3/CD28 Ab, IL‐2, and recombinant TGFβ1 (rTGFβ1) at 2.5 ng/ml for 72 h to induce their differentiation into Tregs. Data are from one representative experiment of a total of four experiments. R.I., relative intensity.
3. Relative frequency (mean ± SEM) of Foxp3⁺ Tregs in cultures of conventional CD4⁺ T cells stimulated with anti‐CD3/CD28 Ab, IL‐2, and increasing doses of rTGFβ1. Splenic conventional CD4⁺ T cells were obtained after 1, 2, and 4 weeks of treatment with vehicle or iPTH. n = 10 mice per group. Data were analyzed by two‐way ANOVA and post hoc tests applying the Bonferroni correction for multiple comparisons. **P < 0.01 and ****P < 0.0001 compared to vehicle.

An increase in the number of Tregs is required for iPTH to induce bone anabolism in mice

A direct means to investigate the contribution of Tregs to the anabolic activity of iPTH is to assess the effects of iPTH in a model in which the increase in the frequency of Tregs is prevented. The surface marker CD25 is expressed at high levels by most CD4⁺ Foxp3⁺ Tregs 45. Accordingly, treatment with anti‐CD25 Abs capable of deleting CD25^hi is used to partially deplete Tregs in vivo 46, 47. We thus treated 6‐week‐old mice with vehicle or iPTH (days 1–28) plus four injections (days −2, 0, 5, and 7) of isotype control Ab or the anti‐CD25 Ab PC61 46, 47. We found CD25^hi to be expressed by CD4⁺ T cells and by a negligible fraction of CD8⁺ T cells (Appendix Fig S3A). Anti‐CD25 Ab decreased the frequency of CD25^hi CD4⁺ T cells but not that of CD25^lo CD4⁺ T cells. As previously reported 48, we also found that treatment with anti‐CD25 Ab decreased the number of CD25^hiFoxp3⁺CD4⁺ T cells, but not the number of CD25^loFoxP3⁻CD4⁺ T cells (Appendix Fig S3B). In addition, treatment with anti‐CD25 Ab did not decrease the percentage of BM of conventional CD4⁺ T cells (TCRβ⁺CD4⁺CD25⁺Foxp3⁻ cells) and that of CD8⁺ T cells (TCRβ⁺CD8⁺CD25⁺ cells) (Appendix Fig S3C–F). Together, these findings demonstrate that anti‐CD25 Ab specifically depletes Tregs.

At sacrifice control mice treated with anti‐CD25 Ab had ~37% fewer Tregs than mice treated with irrelevant (Irr.) Ab (Fig 4A and B). Moreover, treatment with anti‐CD25 Ab prevented the increase in BM Tregs induced by iPTH. The partial depletion of Tregs induced by anti‐CD25 Ab did not increase the production of inflammatory and lineage‐specific cytokines in the BM. In fact, in both the vehicle‐ and iPTH‐treated groups, BM cells from mice treated with anti‐CD25 Ab expressed similar levels of TNF, IL‐17A, IL‐6, IL‐4, and IL‐13 mRNAs to those from mice treated with Irr. Ab (Appendix Fig S4). Moreover, iPTH lowered the mRNA levels of IFNγ, in both the Irr. Ab and the anti‐CD25 Ab groups but did not affect the other cytokines. Since inflammatory cytokines blunt bone formation 49, these findings indicate treatment with anti‐CD25 Ab does not alter the bone anabolic activity of iPTH by inducing inflammation. Analysis by in vitro μCT of femurs harvested at sacrifice revealed that iPTH induced a significant increase in trabecular bone volume (BV/TV) in mice treated with Irr. Ab (Fig 4C and D), but not in those treated with anti‐CD25 Ab. Trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular space (Tb.Sp), which are indices of trabecular structure, were altered by iPTH in mice treated with Irr. Ab, but not in those treated with anti‐CD25 Ab (Appendix Fig S5A–C). By contrast, iPTH increases cortical volume (Ct.Vo) (Fig 4C and E) and cortical thickness (Ct.Th) (Appendix Fig S5D) in both groups of mice, confirming that T cells are not implicated in the mechanism by which iPTH increases cortical volume. Together, these findings demonstrate that the anabolic effects of iPTH in trabecular bone are dependent on increased numbers of Tregs.

Figure 4. Depletion of Tregs by treatment with anti‐CD25 Ab prevents the bone anabolic activity of iPTH .

• A, B
  Relative and absolute frequency of BM Tregs.
• C
  Images of representative three‐dimensional μCT reconstructions of examined femurs from each group.
• D
  Femoral trabecular bone volume (BV/TV) as measured by μCT scanning.
• E
  Femoral cortical bone volume (Ct.Vo) by μCT scanning
• F
  Images are representative sections displaying the calcein double‐fluorescence labeling. Original magnification 20×. Scale bar represents 300 μm.
• G
  Mineral apposition rate (MAR).
• H
  Bone formation rate per mm bone surface (BFR/BS).
• I
  The number of osteoblasts per mm bone surface (N.Ob/BS).
• J
  The percentage of bone surface covered by osteoblasts (Ob.S/BS).
• K
  The images show tartrate‐resistant acid phosphatase (TRAP)‐stained sections of the distal femur. Original magnification 40×. Scale bar represents 300 μm.
• L
  The number of osteoclasts per mm bone surface (N.Oc/BS).
• M
  The percentage of bone surface covered by osteoclasts (Oc.S/BS).
• N
  Serum levels of osteocalcin (OCN), a marker of bone formation.
• O
  Serum levels of type 1 cross‐linked C‐telopeptide (CTX), a marker of resorption.

Data information: n = 10–25 mice per group. Data are expressed as mean ± SEM. All data were normally distributed according to the Shapiro–Wilk normality test and analyzed by two‐way ANOVA and post hoc tests applying the Bonferroni correction for multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 compared to the indicated group.

Analysis of femoral cancellous bone by histomorphometry revealed that iPTH increased the dynamic indices of bone formation mineral apposition rate (MAR) and bone formation rate (BFR/BS) in Treg‐replete mice but not in those treated with anti‐CD25 Ab (Fig 4F–H). Moreover, iPTH increased two static indices of bone formation, the number of osteoblasts per bone surface (N.Ob/BS) (Fig 4I) and the percentage of surfaces covered by osteoblasts (Ob.S/BS) (Fig 4J) in Treg‐replete mice but not in those treated with anti‐CD25 Ab. The finding that treatment with anti‐CD25 Ab did not decrease bone formation in vehicle‐treated mice provides evidence that partial Treg depletion does not cause a nonspecific inhibitory effect of on bone formation. Two indices of bone resorption, the number of OCs per bone surface (N.Oc/BS) and the percentage of surfaces covered by OCs (Oc.S/BS), were not affected by iPTH (Fig 4K–M) in both control and Treg‐depleted groups. However, N.Oc/BS was higher in mice treated with iPTH and anti‐CD25 Ab, as compared to those treated with iPTH and Irr. Ab, suggesting that Treg depletion may stimulate bone resorption.

Measurements of serum levels of osteocalcin, a marker for bone formation, revealed that iPTH increased bone formation in mice treated with Irr. Ab but not in those treated with anti‐CD25 Ab (Fig 4N). Serum CTX, a marker for bone resorption, also increased significantly in response to iPTH in mice treated with Irr. Ab but not in those injected with anti‐CD25 Ab (Fig 4O). Moreover, mice treated with anti‐CD25 Ab had higher CTX levels than those treated with Irr. Ab, confirming that Treg depletion is associated with an increase in bone resorption. The differences between the CTX data and histomorphometric indices of bone resorption are explained by the fact that CTX reflects cortical and trabecular bone resorption, while the histomorphometric analysis was limited to the trabecular compartment.

To determine the role of Tregs in mediating the effects of iPTH on osteoblastogenesis, BM was harvested at sacrifice and cultured for 1 week to allow SCs to proliferate. SCs were then purified and counted. This analysis revealed that iPTH treatment increases the number of SCs in samples from mice treated with Irr. Ab while it had no effects in those treated with anti‐CD25 Ab (Fig 5A). To investigate the mechanism involved, BM was cultured for 1 week, and SCs were purified and used to determine their rate of proliferation and apoptosis. These experiments revealed that iPTH increases significantly the proliferation of SCs from mice treated with Irr. Ab, while it had no effect on the proliferation of SCs from mice treated with anti‐CD25 Ab (Fig 5B). iPTH decreased the rate of SC apoptosis in mice treated with Irr. Ab, while it had no effect on SC apoptosis in mice treated with anti‐CD25 Ab (Fig 5C). Analysis of the expression levels of osteoblastic genes in SCs revealed that iPTH treatment increased the expression of type 1 collagen (Col1), runt‐related transcription factor 2 (Runx2), osterix (Osx), bone sialoprotein (BSP), and osteocalcin (Ocn) mRNAs in SCs from mice treated with Irr. Ab, while it had no effect on SCs from mice treated with anti‐CD25 Ab (Fig 5D). These findings demonstrate that iPTH regulates osteoblast proliferation, differentiation, and life span through a Treg‐dependent mechanism.

Figure 5. Depletion of Tregs by treatment with anti‐CD25 Ab blocks the effects of iPTH on SC number, proliferation, apoptosis, and expression of osteoblast differentiation genes.

1. BM harvested at sacrifice was cultured for 1 week and SCs purified and counted.
2. SCs were purified from BM cultured for 1 week, seeded in equal number, and pulsed with [³H]‐thymidine for 18 h, to assess their proliferation. Data are expressed in CPM.
3. SCs were purified from BM cultured for 1 week and the rate of apoptosis quantified by determinations of caspase‐3 activity.
4. SCs were purified from BM, cultured for 1 week and the level of OB marker gene mRNAs, bone sialoprotein (BSP), type I collagen (Col1), osteocalcin (Ocn), osterix (Osx), and runt‐related transcription factor 2 (Runx2) analyzed by RT–PCR.

Data information: Data are expressed as mean ± SEM. n = 5 mice per group. All data were analyzed by two‐way ANOVA and post hoc tests applying the Bonferroni correction for multiple comparisons. **P < 0.01, ***P < 0.001, and ****P < 0.0001 compared to the indicated group or Irr. Ab vehicle.

To further investigate the relevance of Tregs for the anabolic activity of iPTH, experiments were conducted utilizing DEREG mice 50, a strain that expresses a fusion protein of the human diphtheria toxin (DT) receptor (hDTR) and eGFP under control of the Foxp3 promoter. Foxp3⁺ Tregs can be selectively depleted upon DT administration to DEREG mice, since WT mice do not express the hDTR receptor and are thus insensitive to DT. DT is known not to cause toxic effects in mice 51. While Treg ablation in DEREG mice causes scurfy‐like symptoms in newborn animals, older mice do not develop autoimmune diseases 50, 51 as DT treatment of older DEREG mice causes a partial Treg depletion 52 and the residual Treg population is sufficient to prevent disease in adult mice 52. We treated 6‐week‐old mice with DT (1 μg/mouse, i.p. two times per week for 4 weeks), a treatment modality titrated to block the increase in Tregs induced by iPTH. Mice were also treated with vehicle or iPTH for 4 weeks starting after the first two DT injections. Controls included DEREG mice not treated with DT and WT littermate (LM) mice treated with DT. Analysis of BM samples harvested at sacrifice revealed that iPTH had expanded Tregs in control mice but not in DEREG + DT mice (Fig 6A and B). As expected, analysis of vehicle‐treated groups showed that DEREG + DT mice had a lower absolute and relative numbers of Tregs as compared to control groups. Analysis by μCT of femurs harvested at sacrifice revealed that iPTH induced a significant increase in BV/TV in the two control groups but not in DT‐treated DEREG mice (Fig 6C and D). Analysis of distal femurs also revealed that Tb.Th, Tb.N, and Tb.Sp were differentially altered by IPTH in Treg‐replete and Treg‐depleted mice (Appendix Fig S5E–G). By contrast, iPTH increased Ct.Vo (Fig 6C and E) and cortical thickness (Ct.Th) (Appendix Fig S5H) in all groups of mice. Together, these findings confirmed that the anabolic effects of iPTH in trabecular bone are dependent on increased numbers of Tregs.

Figure 6. Depletion of Tregs by treatment of DEREG mice with diphtheria toxin (DT) prevents the bone anabolic activity of iPTH .

• A, B
  Relative and absolute frequency of BM Tregs.
• C
  Images of representative three‐dimensional μCT reconstructions of examined femurs from each group.
• D
  Femoral trabecular bone volume (BV/TV) as measured by μCT scanning.
• E
  Femoral cortical bone volume (Ct.Vo) by μCT scanning.
• F
  Images are representative sections displaying the calcein double‐fluorescence labeling. Original magnification 20×. Scale bar represents 300 μm.
• G
  Mineral apposition rate (MAR).
• H
  Bone formation rate (BFR).
• I
  The number of osteoblasts per mm bone surface (N.Ob/BS).
• J
  The percentage of bone surface covered by osteoblasts (Ob.S/BS).
• K
  The images show tartrate‐resistant acid phosphatase (TRAP)‐stained sections of the distal femur. Original magnification 40×. Scale bar represents 300 μm.
• L
  The number of osteoclasts per mm bone surface (N.Oc/BS).
• M
  The percentage of bone surface covered by osteoclasts (Oc.S/BS).
• N
  Serum levels of osteocalcin, a marker of bone formation.
• O
  Serum levels of type 1 cross‐linked C‐telopeptide (CTX), a marker of resorption.

Data information: n = 9–13 mice per group. Data are expressed as mean ± SEM. All data were normally distributed according to the Shapiro–Wilk normality test and analyzed by two‐way ANOVA and post hoc tests applying the Bonferroni correction for multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 compared to the indicated group.

Analysis of femoral cancellous bone by histomorphometry revealed that iPTH increased the dynamic indices of bone formation MAR and BFR/BS in the two control groups but not in DT‐treated DEREG mice (Fig 6F–H). N.Ob/BS and Ob.S/BS, which are static indices of bone formation, were also increased by iPTH in DEREG mice and littermate controls but not in DEREG + DT mice (Fig 6I and J). Two indices of bone resorption, N.Oc/BS and Oc.S/BS, were similar in all groups of mice (Fig 6K–M). Analysis of biochemical markers of bone turnover revealed that iPTH increased serum level of osteocalcin and CTX in control mice but not on Treg‐depleted mice (Fig 6N and O), confirming that the increase in the number of BM Treg is required for iPTH to increase bone turnover. In vehicle‐treated mice, serum CTX was higher in Treg‐depleted mice than in Treg‐replete controls, confirming that Treg depletion leads to a stimulation of bone resorption.

Discussion

iPTH treatment is an approved bone anabolic treatment for osteoporosis but its intricate mechanism of action remains largely unknown. We report that in vivo iPTH treatment increases the number of BM Tregs in mice. The increase in the BM Treg pool is essential for the bone anabolic activity of iPTH. The increase in the number of Tregs is due to enhanced differentiation of peripherally induced Tregs, a phenomenon driven by the capacity of iPTH to upregulate the levels of TGFβ and IGF‐1 in the BM and to increase the sensitivity of naïve CD4⁺ cells to TGFβ.

To translate our findings in humans, we measured the absolute and relative frequency of Tregs in the peripheral blood of osteoporotic women treated with Teriparatide, a form of iPTH treatment. We found that Teriparatide increases the number of Tregs at 3 and 6 months of treatment. Teriparatide also increases TGFβ1 mRNA levels, an indicator of Treg activity. By contrast, we found that in vitro treatment of Tregs with PTH does not affect the number of Tregs and TGFβ production, indicating that iPTH regulates human Tregs indirectly.

Tregs are a suppressive population of predominantly CD4⁺ T cells that play a critical role in maintaining immune tolerance and immune homeostasis. Accordingly, defects in Treg numbers and/or activity have been implicated in several chronic inflammatory diseases. Tregs are also known to regulate osteoclast formation 26, 29, 30, blunt bone resorption 26, 27, and prevent ovariectomy‐induced bone loss 28. However, CD4⁺Foxp3⁺ Tregs have not been previously reported to stimulate bone formation, and iPTH treatment was not known to regulate the number of BM Tregs.

Our findings demonstrate that one mechanism by which iPTH expands murine BM Tregs is enhanced differentiation of peripherally induced Tregs. We also found iPTH not to alter Treg proliferation and homing to the BM. Whether additional mechanisms, such as increased life span, contribute to the numeric increase in BM Tregs remains to be determined.

PPR is expressed by conventional CD4⁺ cells and CD8⁺ cells but not by Tregs. On the other hand, iPTH upregulates the BM levels of TGFβ and IGF‐1, factor capable of inducing Treg differentiation in vitro 31, 32, 33, 34, 35. Therefore, iPTH is likely to induce murine Treg differentiation indirectly.

In the mouse, iPTH treatment for 4 weeks did not increase the number of Tregs in the spleen, thymus, and intestinal wall, suggesting that a treatment period longer than 4 weeks might be required for iPTH to expand Tregs in peripheral blood and lymphoid organs, a hypothesis supported by our findings in humans at 3 and 6 months of treatment. It is also likely that the increase in Tregs induced by iPTH occurs first in the BM because of environmental cues. iPTH increases the osteoblastic production of TGFβ and the bone matrix is the largest reservoir of TGFβ in the body 36, 53. iPTH also increases the osteoblastic production of IGF‐1 36, 38. It is thus likely that in the first 4 weeks of treatment the increase in the number of Tregs induced by iPTH is confined to the BM because of the higher levels of TGFβ and IGF‐1 present in the BM as compared to the spleen.

We have used two experimental models to assess the relevance of Tregs for the anabolic activity of iPTH in the mouse, depletion of Tregs in WT mice by treatment with anti‐CD25 Ab, and depletion of Tregs in DEREG mice by treatment with DT. Both strategies prevented the increase in the number of Tregs induced by iPTH. In both cases, blockade of the increase in the number of Tregs prevented the increases in bone formation and trabecular bone volume induced by iPTH, demonstrating that an enlargement of the pool of BM Tregs is required for iPTH to increase trabecular bone mass. By contrast, blockade of Tregs did not blunt the capacity of iPTH to increase cortical bone volume. These findings are in agreement with previous reports from our laboratory demonstrating that T cells do not contribute to iPTH‐induced cortical bone anabolism 11, 20, 22. We hypothesize that iPTH induces cortical bone anabolism primarily by regulating the osteocytic production of sclerostin.

Recently, there has been an explosion in research investigating the potential to manipulate Tregs for clinical purposes 54, 55, 56, 57, 58. Among these studies are several phase I clinical trials to test whether boosting Treg numbers and/or function is a feasible, safe, and potentially effective way to treat diseases such as graft vs. host disease, type 1 diabetes and to prevent the rejection of transplanted organs 54, 55, 56, 57, 58. Osteoporosis is a common chronic disorder that represents a major source of disability in the elderly. Novel anabolic treatments are needed because the long‐term use of current antiresorptive agents is associated with significant adverse events and complications. An increase in the number of Treg is achievable by Treg transfer or treatment with mTOR inhibitors, self‐antigens, or cytokines 59, and inhibition of CD28 costimulation (e.g. CTLA‐4Ig). Based on our findings, pharmacological Tregs may represent a novel therapeutic modality for osteoporosis or for potentiating the anabolic activity of iPTH. Moreover, the use of iPTH to increase the number of Tregs may find applications in transplant medicine or as a treatment for inflammatory and autoimmune conditions.

Materials and Methods

Human study population

All human studies were approved by the Ethical Committee of the A.O.U. Città della Salute e della Scienza—A.O. Ordine Mauriziano—A.S.L. TO1, Turin, Italy, and informed consent was obtained from all participants. The study population was recruited from the patients of A.O.U. Città della Salute e della Scienza, Turin, Italy. The study population included 40 women affected by PMO. The demographic characteristics of the study population are shown in Table 1. The diagnosis of PMO was established based on WHO criteria 60: the presence of secondary osteoporosis was ruled out by medical history, physic examination and blood examinations quantifying calcium, phosphorus, 25OH vitamin D, and PTH. Twenty patients were treated with calcium carbonate (1,000 mg/day) and cholecalciferol 800 UI/day (control treatment). The remaining twenty patients were treated with human PTH 1‐34 (Teriparatide, Eli Lilli, Indianapolis) 20 μg/day subcutaneously, calcium carbonate (1,000 mg/day), and cholecalciferol 800 UI/day. Patients in the Teriparatide group had prior fragility fractures, whereas those in the control treatment group had not, according to guide lines for PMO treatment of the Italian Health ministry. Blood samples (40 ml) were collected in EDTA‐containing vacuum tubes at baseline and after 3 and 6 months of treatment.

Study design

No randomization procedure was used to assign humans and mice to experimental groups. All murine and human samples were analyzed in blinded fashion. The investigators analyzing the human samples were blind to the identity of the study participant, disease state, treatment status, and all other clinical variables. The investigators analyzing mouse samples were blind to the genotype and treatment group. Congenic WT mice or nontransgenic littermates were used as controls for transgenic mice.

Inclusion and exclusion criteria for the human study population

None of the subjects enrolled were affected by disease states affecting bone health other than PMO. Subjects affected by renal or hepatic insufficiency or chronic inflammatory conditions such as rheumatoid arthritis, psoriasis, and inflammatory bowel disease were excluded, as these diseases are known to influence bone status. Subjects treated with drugs active on bone turnover such as bisphosphonates, denosumab, teriparatide, hormone replacement therapy, selective estrogen receptor modulator, strontium ranelate, glucocorticoids, androgens, or GN‐RH agonists for any length of time during the 6 months prior to enrollment were excluded.

Flow cytometric analysis of human samples

Peripheral blood mononuclear cells were purified from 40 ml blood obtained from all patients at each time point using the Ficoll–Paque gradient method, as previously described 61. The Human Regulatory T‐cell Staining Kit (eBioscience Inc., San Diego, CA, USA) was used in accordance with manufacturer's instructions to stain PBMCs for Tregs. Briefly, the following labeled monoclonal antibodies and corresponding isotype controls were used: anti‐CD4 (FITC‐conjugated), anti‐CD25 (APC‐conjugated), and anti‐Foxp3 (PE‐conjugated). After surface staining for CD4 and CD25, cells were washed, fixed, and permeabilized (Fix‐Perm Buffer). Cells were then incubated with anti‐Foxp3 for intra‐nuclear staining. Flow cytometry was performed on an Accuri C6 flow cytometer (BD Biosciences).

Human T‐cell immunomagnetic separation

Human CD4⁺CD25⁺ T cells were isolated from whole blood at each time point using The Complete Kit for Human CD4⁺CD25⁺ Regulatory T Cells, which includes RosetteSep CD4⁺ T Cell Enrichment Cocktail and EasySep Human CD25 Positive Selection Kit (STEMCELL Technologies, Auburn, CA, USA) according to the manufacturer's instructions. The purity of CD4⁺CD25⁺ T cells was 93–95% as assessed by flow cytometry.

Real‐time RT–PCR and human primers

RT–PCR was used to evaluate the mRNA levels of TGFβ1 in CD4⁺CD25⁺ T cells. RNA was isolated using TRIzol reagent (Ambion, Huntingdon, UK), according to the manufacturer's protocol; 1 μg of RNA was reverse‐transcribed to single‐stranded cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied‐Biosystems). RT–PCR was performed with IQ SYBR Green Supermix (Bio‐Rad). The housekeeping control gene was β‐actin, and gene expression was quantified using the $2^{-\mathrm{\Delta }\mathrm{\Delta }{C}_{\mathrm{t}}}$ method. The primers used were as follows: 5′‐CTCTCCGACCTGCCACAGA‐3′ (forward) and 3′‐TCTCAGTATCCCACGGAAATAACC‐5′ (reverse) for TGFβ1; 5′‐CCTAAAAGCCACCCCACTTCT‐3′ (forward) and 3′‐CACCTCCCCTGTGTGGACTT‐5′ (reverse) for β‐actin.

In vitro PTH treatment of human CD4⁺CD25⁺ T cells

Purified CD4⁺CD25⁺ T cells were cultured for 6 days in RPMI medium containing 10% fetal bovine serum (FBS), 1% penicillin streptomycin (Gibco, Thermo Fisher Scientific, MA, USA), with 500 U/ml rIL‐2 (Tebu‐bio srl, Milano, Italy) and 2 μg/ml anti‐CD3 antibody (Biolegend, San Diego, CA, USA). Human PTH 1‐34 (50 ng/ml) or control vehicle was added to the cultures for 1 h or 24 h for three times during the 6‐day culture to mimic the anabolic effects of intermittent PTH 1‐34. Cells were then harvested and utilized for flow cytometry and RNA extraction.

Animals

All the animal procedures were approved by the Institutional Animal Care and Use Committee of Emory University. All in vivo experiments were carried out in female mice. In vitro experiments were conducted using primary cells from female mice or EL4 cells. Female C57BL/6 WT, TCRβ^−/−, DEREG, and Foxp3 eGFP knock‐in mice were purchased from Jackson Laboratories (Bar Harbor, ME). All mice were maintained under specific pathogen‐free conditions and fed sterilized food (5V5R chow) and autoclaved water ad libitum.

In vivo iPTH treatment

For the in vivo iPTH studies, 80 μg/kg/day of hPTH 1‐34 (Bachem California Inc., Torrance, CA, USA) or vehicle was injected daily subcutaneously into female mice for 1–4 weeks starting at the age of 6 weeks, as described 11, 20, 22, 62.

TGFβ1 and IGF‐1 ELISA

BM cells from long bones were cultured for 24 h. Supernatants were collected and assayed for TGFβ1 or IGF‐1 by ELISA kits (R&D Systems) following the manufacturer's instruction. When checking TGFβ1, the medium without the cells was run as the control of the baseline levels of TGFβ1. To isolate CD4⁺ T cells from BM, BM CD8⁺ T cells were discarded by EasySep Mouse CD8a Positive Selection Kit II (StemCell Technologies). CD4⁺ T cells were then positively isolated using APC‐anti‐TCRβ antibody and EasySep Mouse APC Positive Selection Kit (StemCell Technologies).

T‐cell purification and adoptive transfer

Splenic T cells from 6‐ to 8‐week‐old Foxp3 eGFP knock‐in mice were enriched by negative selection using EasySep Mouse T Cell Isolation Kit (StemCell Technologies). Next, conventional CD4⁺ T cells (CD4⁺eGFP⁻) were purified from enriched splenic T cells by FACS sorting and transferred into 4‐week‐old TCRβ^−/− recipient mice by IV injection (3 × 10⁶ cells per mouse). Recipient mice were treated with vehicle or iPTH for 1–4 weeks starting 2 weeks after the CD4⁺eGFP⁻ cell transfer. For Treg isolation, CD4⁺eGFP⁺ cells were sorted from enriched splenic T cells of Foxp3 eGFP knock‐in mice and injected into 6‐week‐old TCRβ^−/− recipient mice via tail vein (2 × 10⁶ cells per mouse). The purity of CD4⁺eGFP⁺ cells was 99% checked by flow cytometry. Vehicle or iPTH treatment was started the day of the Treg transfer for 2 weeks. At the end of the treatment period, the number of CD4⁺eGFP⁺ cells was determined by flow cytometry in spleen and BM cells.

Small intestine Lamina propria lymphocyte (SILP) isolation

Lamina propria lymphocyte isolation was performed as described 63. Briefly, the small intestine was removed and flushed of fecal contents, and intestinal segments containing Peyer's patches were excised. The intestine was opened longitudinally and cut into 5‐mm pieces. Tissues were transferred into a 50‐ml conical tube and shaken at 250 rpm for 20 min at 37°C in HBSS medium (Life Technologies, Grand Island, NY, USA) supplemented with 5% FBS (Mediatech Inc., Manassas, VA, USA) containing 2 mM EDTA. The tissue suspension was passed through a strainer, and the remaining intestinal tissue was washed and then minced, transferred into a fresh 50‐ml conical tube, and shaken for 20 min at 37°C in HBSS + 5% FBS containing type VIII collagenase (Sigma‐Aldrich, St. Louis, MO, USA) at 1.5 mg/ml and DNase I at 100 μg/ml (Roche). The tissue suspension was collected, passed through a strainer, and pelleted by centrifugation at 267 g for 5 min. The pellet was suspended in 5 ml HBSS and 5 ml 90% isotonic Percoll, and then transferred into a 15‐ml tube and mixed by tilting back and forth. The cell content was layered onto 2 ml of 70% isotonic Percoll. The gradient was centrifuged at 742 g for 20 min. Cells were collected from the interface area. When these cells were used for flow cytometry, the live cells were discriminated by LIVE/DEAD Fixable Yellow Dead Cell Stain Kit (ThermoFisher).

BrdU incorporation studies

Mice were injected IP with 40 mg/kg/day of BrdU solution for 4 days and sacrificed 24 h later. BrdU detection was performed by using the BrdU Flow Kit (BD Biosciences, San Diego, CA) and analyzing cells by flow cytometry. The percentage of BrdU⁺ Treg cells was quantified by gating CD4⁺Foxp3⁺ cells in the TCRβ⁺ cell population.

Anti‐CD25 Ab treatment

Six‐week‐old WT mice were injected daily with vehicle or PTH for 4 weeks (days 1–28). Mice were also injected with anti‐CD25 Ab (clone PC61, 500 μg/mouse/injection IP) (BioXCell, West Lebanon, NH, USA) on days −2, 0, 5, and 7 or isotype‐matched irrelevant Ab.

DT treatment

DT was purchased from Merck (catalog number 322326), and each lot was tested for toxicity in WT mice and titrated for potency in DEREG mice prior to use. DEREG and littermate control mice were administered 1 μg DT intraperitoneally on two consecutive days each week for total 4 weeks.

μCT measurements

μCT scanning and analysis of the distal femur were performed as reported previously 20, 64, 65 using a Scanco μCT‐40 scanner (Scanco Medical, Bassersdorf, Switzerland). Femoral trabecular and cortical bone regions were evaluated using isotropic 12‐μm voxels. For the femoral trabecular region, we analyzed 140 slices from the 50 slices under the distal growth plate. Femoral cortical bone was assessed using 80 continuous CT slides located at the femoral midshaft. X‐ray tube potential was 70 kVp, and integration time was 200 ms.

Quantitative bone histomorphometry

The measurements, terminology, and units used for histomorphometric analysis were those recommended by the Nomenclature Committee of the American Society of Bone and Mineral Research 66. Nonconsecutive longitudinal sections of the femur were prepared and analyzed as described previously 65. Mice were injected subcutaneously with calcein at day 7 and day 2 before sacrifice. Nonconsecutive longitudinal sections (5 μm thick) were cut from methyl methacrylate plastic‐embedded blocks along the frontal plane using a Leica RM2155 microtome and were stained with Goldner's trichrome stain for the static measurements. Additional sections were cut at 10 μm and left unstained for dynamic (fluorescent) measurements. Measurements were obtained in an area of cancellous bone that measured ≈ 2.5 mm² and contained only secondary spongiosa, which was located 0.5–2.5 mm proximal to the epiphyseal growth cartilage of the femurs. Measurements of single‐labeled and double‐labeled fluorescent surfaces and interlabel width were made in the same region of interest using unstained sections. Mineral apposition rate (MAR) and BFR were calculated by the software by applying the interlabel period. Histomorphometry was done using the Bioquant Image Analysis System (R&M Biometrics).

Markers of bone turnover

Serum CTX and osteocalcin were measured by rodent‐specific ELISA assays (Immunodiagnostic Systems, Scottsdale, AZ, USA).

Stromal cell purification

BM SCs were purified as previously described 11, 20, 22. In brief, BM cells from long bones were cultured for 7 days in α‐MEM medium containing 10% FBS, 100 mg/ml of penicillin, and 100 IU/ml of streptomycin, to allow the proliferation of SCs. After removing nonadherent cells, adherent macrophages were eliminated by positive selection using anti‐CD11c MACS Microbeads (Miltenyi Biotech, Auburn, CA, USA). The remaining adherent cells were defined as SCs as they express alkaline phosphatase (ALP), type I collagen, and Runx2, and have the capacity to form mineralization nodules when further cultured under mineralizing conditions.

SC thymidine incorporation assay

The proliferation of purified SCs was measured by [³H]‐thymidine incorporation assay. SCs were pulsed with [³H]‐thymidine (0.5 μCi/10,000 cells) for 18 h and were harvested with a Cell Harvester (Skatron, Inc., Sterling, VA, USA). [³H]‐thymidine incorporation was read by a LS 6000 IC Liquid Scintillation Counter (Beckman Coulter, Inc., Fullerton, CA, USA).

SC apoptosis assay

The activity of caspase‐3, the critical protease in the induction of apoptosis, was measured in SCs using CaspACE Assay System (Promega Corporation, Madison, WI, USA) according to the manufacturer's protocol.

In vitro Treg differentiation

Assessment of Treg differentiation in vitro was carried out as described 67. Splenic CD4⁺ T cells from WT mice treated with vehicle or iPTH were purified by negative selection using EasySep Mouse CD4⁺ T Cell Isolation Kit (StemCell Technologies). CD25⁺ cells were then removed using the EasySep Mouse CD25 Regulatory T Cell Positive Isolation Kit (StemCell Technologies). The remaining cells (CD25⁻ enriched CD4⁺ T cells) were cultured in plates coated with anti‐CD3 Ab (3 μg/ml) in the presence of anti‐CD28 Ab (3 μg/ml), IL‐2 (5 ng/ml), and TGFβ1 (0.1–5 ng/ml) for 3 days. Cells were then harvested and analyzed by flow cytometry to enumerate CD4⁺Foxp3⁺ cells.

Western blotting

Resting or cultured conventional CD4⁺ T cells were lysed in lysis buffer (20 mM Tris–HCl, pH 7.5, 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% Triton X‐100, 2.5 mM sodium pyrophosphate, 1 mM β‐glycerophosphate, 1 mM Na₃VO₄, and 1 μg/ml leupeptin, Cell Signaling Technology, Danvers, MA, USA). Halt Protease and Phosphatase Inhibitor Cocktail (ThermoFisher Scientific) were added into the reagents before using. Lysates were cleared by centrifugation, and the supernatants were boiled in SDS loading buffer. The same amount of proteins was separated on 10% Mini‐PROTEAN TGX Precast Gels (Bio‐Rad) and electroblotted to nitrocellulose membrane (ThermoFisher Scientific). Proteins were detected by anti‐Smad7 (catalog no. sc‐365846, Santa Cruz Biotechnology), anti‐phospho‐Smad2 (Ser465/467) (catalog no. 3108), or anti‐phospho‐Smad3 (Ser423/425) (catalog no. 9520) (Cell Signaling Technology, Danvers, MA) antibodies. Anti‐beta‐actin (catalog no. sc‐1616) antibody bought from Santa Cruz Biotechnology was used as the loading control. Western blot analysis was conducted by using Luminata Crescendo Western HRP substrate (EMD Millipore). Band intensities were quantified with Quantity One 1D Analysis Software (Bio‐Rad Laboratories) and expressed relative to beta‐actin.

Flow cytometry

Flow cytometry was performed on a LSR II system (BD Biosciences, Franklin Lakes, NJ, USA), and data were analyzed using FlowJo software (Tree Star, Inc., Ashland, OR). For intracellular Foxp3 staining, APC‐Foxp3 (clone FJK‐16s, eBioscience) antibody was added after cell fixation and permeabilization with BD Transcription Factor Buffer Set (BD Biosciences). The following anti‐mouse antibodies were used for cell surface staining: purified CD16/32, BV 421‐TCRβ (clone H57‐597), PerCP/Cy5.5‐CD4 (clone RM4‐5), PE‐CD25 (clone PC61), and BV 711‐CD8 (clone 53‐6.7) (Biolegend).

Real‐time RT–PCR and murine primers

Total RNA was isolated from total BM cells and SCs using TRIzol reagent (ThermoFisher Scientific). cDNA was synthesized with random hexamer primers (Roche) and AMV reverse transcriptase (Roche). mRNA levels of bone sialoprotein (BSP), collagen 1 (Col1), osteocalcin (Ocn), osterix (Osx), and runt‐related transcription factor 2 (Runx2) in SCs and that of TNF, IL‐6, IFNγ, IL‐4, IL‐13, and IL‐17A in total BM were quantified by real‐time PCR. Changes in relative gene expression between vehicle and iPTH groups were calculated using the $2^{-\mathrm{\Delta }\mathrm{\Delta }{C}_{\mathrm{T}}}$ method with normalization to 18S rRNA. The primers used are provided in Appendix Table S1.

Statistical analysis

Human Tregs and human TGFβ1 data were analyzed by ANOVA for repeated measures. Human demographic and clinical data were analyzed by unpaired two‐tailed t‐tests. When murine data were normally distributed according to the Shapiro–Wilk normality test, they were analyzed by unpaired two‐tailed t‐tests, one‐way or two‐way analysis of variance, as appropriate. This analysis included the main effects for animal strain and treatment plus the statistical interaction between animal strain and treatment. When the statistical interaction between animal strain and treatment group was not statistically significant (P > 0.05) nor suggestive of an important interaction (P > 0.10), P‐values for the main effects tests were reported. When the statistical interaction was statistically significant (P < 0.05) or suggestive of an important interaction, then t‐tests were used to compare the differences between the treatment means for each animal strain, applying the Bonferroni correction for multiple comparisons. Data that were not normally distributed (as tested by Shapiro–Wilk normality test) were analyzed by Kruskal–Wallis nonparametric tests.

Author contributions

MY, RDP, MNW, and RP designed the animal studies. PDA, FS, and IB designed and performed the human studies and analyzed the human data. MY, CV, AMT, J‐YL, EH, and JA performed the research and analyzed the animal data. RP wrote the manuscript.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Review Process File

EMBO Reports (2018) 19: 156–171