Mechanisms limiting the long-term anabolic effects of teriparatide (PTH 1-34) on bone.

PTH, a peptide hormone produced by chief cells of the parathyroid gland, acts on bones, intestines and kidneys to regulate blood calcium levels, essential for muscle contraction, nerve impulse transmission, blood clotting, and skeletal mineralization. 1 PTH stimulates bone remodeling (ie, increases both bone resorption and formation). However, it’s net effect on bone depends on the exposure kinetics: continuous PTH predominantly stimulates osteoclastic bone resorption leading to net bone loss, whereas pulsatile, intermittent PTH increases osteoblast number, bone formation, and net bone mass but also increases bone resorption. 2 , 3 The period during which PTH-induced bone formation exceeds bone resorption is known as the “anabolic window.” The first 34 N-terminal amino acids of PTH comprise the drug teriparatide. In 2002, teriparatide became the first osteoanabolic medication to receive approval from the Food and Drug Administration (FDA) for the treatment of osteoporosis. Teriparatide effectively reduces the risk of new vertebral fractures and increases BMD of the spine and hip in postmenopausal women and men with osteoporosis. 4 , 5 Teriparatide is also approved for the treatment of glucocorticoid-induced osteoporosis. 6 Current guidelines recommend teriparatide use for 2 yr. 7 , 8 Although longer use of teriparatide can be considered if fracture risk remains high, the durability of its clinical benefits with extended therapy remains unclear, given the known waning anabolic response. The anabolic effects of teriparatide are most potent in the first 6-12 mo of administration. Afterwards, the rate of increase in BMD plateaus, as observed in multiple clinical trials. 4 , 9 , 10 For example, in the Fracture Prevention Trial, 1637 postmenopausal women with osteoporosis were randomized to receive placebo or teriparatide (20 or 40 μg/d). 4 In teriparatide-treated patients, LS-BMD increased rapidly for the first 6 mo, followed by slower gain from 6 to 18 mo. 11 The European Study of Forsteo (EUROFORS) was an randomized controlled trial (RCT) designed to determine the effects of prior treatment on teriparatide efficacy and the effects of various sequential treatment regimens following 1-yr of teriparatide therapy. ΔLS-BMD in 865 patients receiving teriparatide (20 μg/d) was greatest for the initial 6 mo of treatment compared to the following 6-12 mo. 9 In 503 patients who continued to receive teriparatide, the ΔLS-BMD continued to decline from 12 to 24 mo of treatment regardless of whether patients were previously treatment-naïve or treated with antiresorptives. 12 In 245 patients treated with antiresorptive drugs prior to teriparatide, the same decline in ΔLS-BMD was demonstrated irrespective of the type of antiresorptive therapy. 13 The TOWER trial investigated the effect of once weekly teriparatide (56.5 μg/wk) in 578 Japanese patients with osteoporosis. Here, LS-BMD increased most rapidly for the first 6 mo of treatment. Subsequently, ΔLS-BMD decreased from 6 to 12 mo followed by an additional decline from 12 to 18 mo. 10 This characteristic pattern of early rapid BMD improvement followed by a plateau has been reproduced across diverse patient populations, pre-treatment regimens, teriparatide doses, and dosing intervals ( Table 1 ). Clinical studies showing decreased ΔLS-BMD over time. Number of participants included in study arms treated with teriparatide. Abbreviations: ΔLS-BMD, change in LS-BMD; RCT, randomized controlled trial; s.c., subcutaneous; TPTD, teriparatide. Similar waning effects are observed in the impact of teriparatide on serum and urine bone turnover markers ( Table 2 ). Within the first months of administration, teriparatide rapidly increases bone formation markers, such as bone-specific alkaline phosphatase (ALP), osteocalcin, type I collagen C-terminal propeptide, and type I collagen N-terminal propeptide (P1NP). Bone resorption markers such as crosslinked C-telopeptide of type I collagen, crosslinked N-telopeptide of type I collagen (NTX), and tartrate-resistance acid phosphatase type 5b (TRAP) increase with a delay. 34 , 35 During long-term treatment with teriparatide, both bone formation and resorption markers peak after 6-12 mo followed by a decline. 36 Clinical studies showing progression of bone turnover markers over time. Abbreviations: Bone ALP, bone alkaline phosphatase; CTX, cross-linked C-terminal telopeptide of type 1 collagen; DPD, deoxypyridinolone; NTX, cross-linked N-terminal telopeptide of type 1 collagen; OHP, hydroxyproline; PICP, serum procollagen I carboxy-terminal; PINP, serum procollagen I N-terminal; PYD, pyridinoline; s.c., subcutaneous; TPTD, teriparatide. The waning effects of intermittent PTH treatment over time have also been reported in pre-clinical in vivo studies. In ovariectomized rats, PTH rapidly increased spine and whole skeleton BMD during the first 6 wk of treatment, after which the rate of BMD gain slowed. Serum markers of bone formation (ALP and osteocalcin) followed a similar trend. 37 In another study, female ovariectomized or Sham-operated mice were treated with PTH for 8 wk. The percentage increase in trabecular volumetric BMD from pretreatment values was greater at 4 wk than 8 wk in both ovariectomized and Sham-operated mice. 38 Another study evaluated the effect of PTH on ectopic ossicle formation in 4- to 6-wk-old nude mice after injection of BM stromal cells isolated from young C57BL/6 mice. Mice treated with PTH for 3 wk developed ossicles with increased radiopacity compared to controls, while mice treated for 7 wk had similar radiopacity irrespective of treatment. 39 This demonstrates that increased bone formation at 3 wk was not sustained. In this review, we will discuss different hypotheses for the waning effects of PTH(1-34) on bone, as evidenced by the reversal in serum bone turnover markers and stagnating ΔLS-BMD, including osteoprogenitor depletion, changes in bone remodeling dynamics, counter-regulatory molecules (Wnt inhibitors), the influence of mechanical stimulation, and downstream signaling adaptations.

An important anabolic mechanism of teriparatide is the activation of the Wnt signaling pathway that promotes osteogenic differentiation. PTH affects the Wnt signaling pathway via at least 3 mechanisms. 92 First, PTH has been shown to decrease the expression of Wnt signaling inhibitors. PTH receptor (PTH1R) is a G-protein coupled receptor that activates the G s protein and protein kinase A (PKA). 93 PKA phosphorylates downstream targets (eg, salt-induced kinase 2), causing histone deacetylases 4 and 5 to translocate to the nucleus and inhibit MEF2C-driven SOST expression, coding for the Wnt antagonist sclerostin. 94 PTH can also decrease expression of the Wnt antagonist Dickkopf protein-1 (DKK-1). 95 Second, PTH-PTH1R can also form a complex with low-density lipoprotein receptor-related protein 6 (LRP6) causing activation of canonical Wnt signaling. 96 Lastly, it has been shown that PKA (upon activation by PTH/PTH1R) can phosphorylate and stabilize β-catenin. 95 These 3 mechanisms (summarized in Figure 1 ) could be altered in long-term treatment with PTH. The following section focuses on the first mechanism as it has been suggested that a gradual increase in the expression of these Wnt antagonists during prolonged teriparatide treatment may limit activation of Wnt signaling and thereby contribute to the waning anabolic response to teriparatide. The interactions of PTH signaling and the Wnt signaling pathway. (1) PTH(1-34) interacts with its G-protein coupled receptor (PTH1R) and activates Gs leading to the activation of protein kinase A (PKA). PKA phosphorylates downstream targets (indicated by dashed lines) causing decreased expression of genes coding for the Wnt-antagonists sclerostin (SOST) and Dickkopf-protein 1 (DKK-1). In the absence of Wnt-antagonists, Wnt associates with frizzled receptors (Frizzled) and low-density lipoprotein receptor-related protein (LRP5/6). As a result, the β-catenin destruction complex (consisting of Axin, adenomatous polyposis coli (APC), and glycogen synthase kinase 3 β (GSK-3 β)) is disabled and β-catenin translocates to the nucleus and activates transcription of osteoanabolic genes. (2) PTH binds to PTH1R and activates PKA. PKA phosphorylates LRP6 and PTH-PTH1R forms a complex with phosphorylated LPR6, which recruits Axin and leads to inhibition of the β-catenin destruction complex and stabilization of β-catenin and its translocation to nucleus. (3) PKA phosphorylates β-catenin, leading to its stabilization and translocation to the nucleus. It should be noted that this figure only highlights the interactions between PTH and Wnt signaling pathways and does not depict the full range of mechanisms involved in bone anabolic effects of PTH. Figure is created using https://BioRender.com . Intriguingly, the Wnt antagonists DKK-1 and sclerostin have been shown to compensate for each other. DKK-1 inhibition increases expression of the Sost gene, coding for sclerostin, in mice. Moreover, DKK-1 antibody has potent bone anabolic effects in Sost −/− mice, while only having negligible effects in Sost +/+ mice. 97 Patients with deficient sclerostin production leading to sclerosteosis or Van Buchem disease, both have increased levels of DKK-1, 98 and mice treated with sclerostin antibodies have increased DKK-1 protein levels. 99 , 100 As PTH(1-34) affects the expression of these antagonists, a compensatory upregulation of Wnt antagonists over time could attenuate the anabolic effects of the drug. Several clinical studies have measured DKK-1 levels during long-term teriparatide treatment, with somewhat variable results. Two prospective cohort studies observed that DKK-1 levels were increased as early as 6 mo of treatment and remained elevated after 18 mo compared to baseline. 101 , 102 In contrast, another study showed that serum levels of DKK-1 first increased after 6 mo of treatment, at the same time as bone turnover markers started to stagnate. 103 A third study found a slight decrease of DKK-1 after 6 mo, followed by a significant increase after 12 mo. 104 Two of these studies also measured serum sclerostin levels and found these to be constant. 103 , 104 Lastly, a study investigated the expression of serum microRNAs after 3 and 12 mo of treatment with teriparatide and discovered decreased expression of miR-33-3p, which negatively regulates DKK-1 expression. 105 These observations raise the question of whether increased circulating DKK-1 contributes to attenuation of teriparatide efficacy over time. However, preclinical studies do not support this interpretation. Intermittent PTH has consistently been shown to suppress DKK-1 expression in vitro and in vivo, and in mice with osteoblast-targeted overexpression of DKK-1, intermittent PTH still produced a bone anabolic response comparable to that in wild-type mice. 106 Thus, although clinical studies suggest that circulating DKK-1 may rise during long-term teriparatide treatment, current preclinical data indicate that this increase is unlikely to be a major driver of the waning anabolic response. Several studies have monitored the serum levels of sclerostin in patients treated with teriparatide ( Table 4 ), reporting variable results including increased, decreased and constant serum sclerostin levels in response to teriparatide treatment. In addition, a recent retrospective clinical study found that patients switched from teriparatide to romosozumab experienced further increases in LS, FN, and TH BMD, with greater gains than those switched in the reverse sequence. 113 However, because the study did not include a comparator arm that continued teriparatide alone, it does not directly prove that sclerostin inhibition extends the anabolic window of teriparatide. Clinical studies investigating the effects of teriparatide on sclerostin serum levels. Number of participants included in study arms treated with teriparatide. Abbreviations: ↓, decrease; →, constant level; ↑, increase; RCT, randomized controlled trial; s.c., subcutaneous; TPTD, teriparatide. In preclinical studies, both increased and decreased levels of sclerostin can impact the anabolic effects of PTH(1-34) on bone. In transgenic mice, overexpressing Sost (SOST Tg) treated with intermittent PTH(1-34) for 2 mo, the increase in BMD in the LS, tibia, and distal femur was smaller in SOST Tg mice compared to WT mice. Moreover, the anabolic response to PTH(1-34) on trabecular bone volume fraction and trabecular thickness were attenuated in SOST Tg mice compared to WT mice. Dynamic bone histomorphometry revealed decreased bone formation rates in PTH(1-34)-treated SOST-Tg mice compared to PTH(1-34)-treated WT mice. 114 PTH(1-34)-treated global Sost KO mice have an attenuated increase in BMD compared to PTH(1-34)-treated WT mice. While one study identified reduced bone formation rates as an underlying mechanism, 114 another study found an increase in cortical porosity at constant cortical bone formation parameters. 115 Importantly, PTH(1-34)-induced increase in trabecular bone volume fraction was not impaired in Sost KO mice compared to WT controls. 115 Lastly, 2 studies on ovariectomized rats found that 12 wk of combined sclerostin antibody and PTH(1-34) treatment produced greater improvements in micro-CT-derived structural parameters, mechanical strength, and healing of a cylindrical bone defect than either monotherapy alone. 116 , 117 Overall, the preclinical literature supports a biologically plausible role for sclerostin in modulating the anabolic response to intermittent PTH(1-34), and some clinical observations are compatible with this idea. However, it should be considered that the additive or synergistic skeletal effects of anti-sclerostin antibody combined with intermittent PTH(1-34) do not, by themselves, prove that sclerostin contributes to narrowing of teriparatide’s anabolic window. In addition, the human data are inconsistent, and definitive proof that rising sclerostin contributes to narrowing of the anabolic window during long-term teriparatide treatment is lacking. Notably, no clinical study has directly tested whether antagonizing sclerostin during ongoing teriparatide treatment can prevent or delay waning of the anabolic response. Overall, current evidence does not support DKK-1 as a major driver of the waning anabolic effect of teriparatide, although circulating DKK-1 may increase during long-term treatment. By contrast, both preclinical and some clinical findings suggest that sclerostin may contribute to modulation of the anabolic response to PTH(1-34). In preclinical models, altered sclerostin levels affect the skeletal response to intermittent PTH, and combined PTH plus sclerostin antibody treatment produces greater anabolic effects than either monotherapy alone. Clinical evidence remains insufficient to establish causality, however, and no study has directly tested whether inhibiting sclerostin during continued teriparatide treatment can extend the anabolic window. Such a question would best be addressed in a head-to-head trial comparing teriparatide alone with teriparatide plus romosozumab after the initial treatment period. In addition, further mechanistic studies are needed because different skeletal cell populations may respond differently to PTH(1-34) and Wnt modulation, and compensatory interactions among Wnt pathway components may influence the overall outcome.

Another mechanism that may contribute to the waning anabolic effect of teriparatide is the altered coupling of bone formation and resorption during the remodeling process. Although teriparatide initially increases bone formation more than resorption, this imbalance is temporary. In clinical studies, bone formation markers rise early, whereas bone resorption markers increase later with continued treatment, narrowing the anabolic window. 34 Evidence from combination therapy trials demonstrates that co-administration of denosumab (antibody targeting RANKL) with teriparatide can enhance or sustain BMD gains by mitigating resorption, underscoring the critical role of remodeling balance in determining the durability of the anabolic response. 83 However, not all antiresorptives interact with PTH the same way, as for example, alendronate has been shown to limit teriparatide’s anabolic effect. 30 In addition, the capacity to initiate new remodeling units may plateau over time, 84 and osteoclast-driven feedback may further modulate the coupling process. During bone resorption, matrix-derived factors such as TGF-β are released 85 and play a role in recruiting osteoblast progenitors to the bone surface, thereby coupling bone formation to resorption. 85 However, TGF-β also inhibits osteoblast differentiation in vitro 86 , 87 and in vivo 88 and excessive TGF-β compromises skeletal homeostasis. 88 In addition, elevations in serum calcium upon bone resorption may stimulate calcitonin secretion, which may indirectly restrain bone formation by suppressing osteoclast-derived pro-osteogenic coupling signals, such as sphingosine 1-phosphate. 89 Together, these feedback mechanisms may limit further anabolic responses during prolonged teriparatide treatment. Furthermore, prolonged PTH-induced remodeling can also increase cortical bone porosity, offsetting gains in trabecular compartments and limiting net improvements in BMD at certain sites. 90 , 91 In summary, the balance of bone remodeling appears crucial: teriparatide’s early advantage (formation >> resorption) diminishes as resorption accelerates and catches up. Strategies that rebalance remodeling in favor of formation can prolong BMD gains, highlighting that the coupling between formation and resorption is a key determinant of the durability of teriparatide’s anabolic response.

This review aimed to explore different hypotheses proposed to explain the waning effects of teriparatide on bone. Several preclinical and clinical studies have attempted to overcome these waning effects, but the causes underlying the phenomenon remain partly understood. We discussed how osteoprogenitor supply, bone remodeling dynamics, counter-regulatory molecules (like Wnt inhibitors), mechanical feedback, and receptor signaling adaptations might contribute to the tapering of teriparatide’s efficacy. For each of these factors, we found only circumstantial or partial evidence of their role, rather than definitive proof. It is likely that the waning effect is multifactorial, with several of these mechanisms acting in concert. Recently, an observational study suggested that patients previously treated with teriparatide have attenuated responses to the sclerostin antibody romosozumab. 145 A preclinical study reported that osteoprogenitor frequency limited the effects of sclerostin antibodies, 146 and another study showed dampened bone formation in response to repeat dosing with sclerostin antibody in mice concurrent with upregulation of Wnt antagonists. 147 In addition, inhibition of DKK-1 showed a synergistic effect on bone formation induced by anti-sclerostin antibody. 100 These parallel findings suggest that a shared adaptive mechanism (possibly depletion of a target cell population or induction of feedback inhibitors) may underlie the waning effectiveness of diverse anabolic agents. Identifying these common limiting factors is therefore of high significance for improving osteoporosis treatment. In conclusion, continued research is needed to determine how to prevent or counteract the body’s adaptive responses to PTH, whether by combination therapies (eg, concurrent antiresorptives or Wnt signaling modulators), altered dosing schedules, or adjuvant mechanical interventions, to maintain bone formation at its peak. Unraveling the exact mechanisms behind the waning anabolic effect of PTH will enable the development of informed strategies to sustain bone gains and improve long-term outcomes for patients on osteoanabolic therapy.

Multiple tightly regulated mechanisms allow PTH(1-34) to increase the number of active osteoblasts. 40 In skeletally mature bone, new osteoblasts are supplied by heterogeneous skeletal stem/progenitor cells (SSPCs) rather than by a single uniform progenitor pool that can be identified using a universal marker. 41 , 42 It has been hypothesized that the waning anabolic effect of teriparatide may partly reflect a progressive limitation in the recruitment or osteogenic differentiation of these progenitor populations over time. Preclinical studies have shown that increased numbers of SSPCs and osteoprogenitors can enhance the anabolic effects of PTH(1-34). Nuclear matrix protein 4 (Nmp4) also known as Cas-interacting zinc finger protein (CIZ) is a ubiquitously expressed transcription factor that shuttles between the cytoplasm and nucleus and has been implicated in restraining the anabolic response of bone to PTH. 43 Global Nmp4 KO (Nmp4 (−/−) ) mice exhibit no major baseline skeletal abnormalities but show significantly greater trabecular bone gain in response to intermittent PTH(1-34) than WT mice. 44 , 45 Notably, PTH-treated Nmp4 (−/−) mice had a 4-fold increase in CD45 − /CD146 + /CD105 + /nestin + BM stromal cells after 3 wk of PTH(1-34) treatment, and their BM produced 4 times more ALP + colonies (CFU-Fs) in vitro than PTH-treated WT mice. These findings suggest that Nmp4/CIZ restricts the anabolic effects of PTH(1-34) by suppressing the size of the osteoprogenitor pool from which osteoblasts are recruited. 46 Furthermore, conditional KO of Nmp4 in mesenchymal progenitors (using Prx1-Cre) but not in mature osteoblasts (Bglap-Cre) or osteocytes (Dmp1-Cre) enhanced the bone formation response to PTH(1-34), underscoring that the pool of early progenitors is essential for teriparatide’s full efficacy. 47 Given that increased numbers of osteoprogenitors enhance the anabolic response to PTH(1-34) treatment, it is reasonable to hypothesize that gradual depletion of osteoprogenitors could contribute to limiting the anabolic response to PTH(1-34) over time. Whether intermittent PTH(1-34) expands or depletes the osteoprogenitor pool depends on its influence on progenitor proliferation vs differentiation. If PTH primarily drives osteoprogenitors to differentiate into osteoblasts without replenishing them via proliferation, the progenitor pool could decline over time. Studies on this topic have yielded mixed results, as detailed in Table 3 . The observed variability likely reflects differences in differentiation state, treatment pattern, and cell population. Among these, the studied osteoprogenitor population seems to be an especially important factor. Lineage tracing studies in mice have for example indicated increased proliferation of Gli+-metaphyseal mesenchymal progenitors 52 but decreased proliferation of Sox9+-progenitors in response to short-term intermittent PTH(1-34). 59 A limitation of these studies is that both Gli+-progenitors 63 and Nes+-progenitors 50 play a role in bone development and are known to decline with age. Besides, young, skeletally immature mice with a higher basal bone turnover rate were used, making it difficult to translate the results to an aged, human population. In this regard, although Sox9+ progenitors have been shown to contribute to osteoblast lineage cells in young adult mice, their specific role in aged, remodeling-dominant bone remains insufficiently defined. Additional studies are needed to define how prolonged PTH(1-34) administration affects osteoprogenitor populations relevant to bone remodeling in the mature and aging skeleton. Importantly, this should also be examined in patients receiving long-term teriparatide treatment to determine whether comparable changes occur in the clinical setting. The effects of intermittent PTH on osteoprogenitor proliferation and cell numbers. If the gradual depletion of osteoprogenitors contributes to the waning effects of teriparatide over time, cyclic regimens and treatment holidays might sustain the anabolic response. Cyclical teriparatide regimens and “drug holidays” have been explored as strategies to mitigate the plateau, although direct evidence of progenitor recovery is lacking. Several such regimens have been tested. An initial clinical study in osteoporotic women on alendronate compared continuous daily teriparatide vs a cyclic regimen (3 mo on teriparatide, 3 mo off, repeated five times). In the continuously treated group, LS-BMD did not increase further from 12 to 15 mo, while the cyclically treated group had not yet reached a treatment plateau. With only 60% of the daily dose, the cyclic approach achieved 88.5% of the increase in BMD that the continuous regimen provided. 16 Twenty-seven women, who continued to be at high risk of fracture after the 15 initial months of treatment, continued in a follow-up study, consisting of 12 mo of treatment with alendronate only, followed by 12 mo of retreatment with daily teriparatide and alendronate. Women experienced similar ΔLS-BMD in the retreatment period as in the first treatment course. 64 Of note, concurrent treatment with alendronate has since been shown to limit the effectiveness of teriparatide, 30 limiting the generalizability of the data to patients treated with teriparatide only. As this clinical study was underpowered for detecting changes in fracture risk, the group developed a mouse model. Mice were treated continuously for 7 wk or cyclically, alternating between PTH(1-34) injections and vehicle every week. Mice treated with the cyclic regimen achieved higher femur and vertebral BMD, osteocalcin concentration, and femur strength per unit PTH administered than mice treated with the continuous regimen. 65 Microarchitectural parameters of the lumbar vertebrae improved equally in both groups, whereas the microarchitecture of the femur was affected to a larger extent by the continuous regimen. 66 Overall, the mouse model indicated that cyclic PTH(1-34) regimens are reasonable alternatives to continuous regimens. Another study randomized postmenopausal women to constant (30 μg/day for 18 mo) vs escalating doses of teriparatide (20 μg/day for 6 mo, 30 μg/day for 6 mo, 40 μg/day for 6 mo). The escalating regimen prevented waning of bone formation markers during months 12-18, while achieving similar increases in BMD. 67 It is possible that the escalating dose prevented a rapid depletion of osteoprogenitors. On the contrary, other studies did not observe these positive effects of cyclic regimens. Women and men with osteoporosis were treated with teriparatide for 24 mo followed by 12 mo of treatment holiday and then treated with teriparatide again for another 12 mo. The response to teriparatide as measured by spine BMD and increase in osteocalcin, P1NP, and NTX remained attenuated in the second treatment period despite of the long treatment holiday. 68 Similarly, another randomized trial compared the effects of daily vs cyclic (3 mo on, 3 mo off) teriparatide in treatment naïve women and women previously treated with alendronate. In treatment-naïve women, the daily regimen achieved a 2-fold greater gain in BMD than the cyclic group, while alendronate treated women had an equal increase in BMD in both regimens. 69 Women with a T-score <2 after 2 yr of treatment with teriparatide, were invited to participate in a follow-up study. After 4 yr of cyclic treatment with teriparatide, a total of 24 mo on teriparatide over 8 cycles, there was no significant increase in BMD in women receiving a cyclic regimen compared to women receiving daily teriparatide for 2 yr followed by alendronate for 2 yr. 70 While it could be argued that the chosen population consisted of low responders to teriparatide, the study did not find that cyclic treatment with teriparatide has advantages over continuous regimens. Overall, some cyclic or dose-escalating teriparatide regimens can maintain bone formation and BMD gains longer, while other studies have not been able to replicate these benefits. Further research is needed to elucidate whether the success of some cyclic regimens is truly due to allowing osteoprogenitor recovery, or whether other factors (eg, re-sensitization of PTH signaling or re-balancing formation vs resorption) underlie the improved outcomes. Aging is associated with decreased numbers and proliferation potential of various osteoprogenitor populations. 71–73 Moreover, aging reduces transcriptomic diversity of SSPCs and affects interactions of the skeletal and hematopoietic lineages in the bone marrow niche, increasing fragility. 74 Therefore, older patients with a lower reserve of osteoprogenitors, may have a more limited response to teriparatide, especially if prolonged treatment further depletes these cells. A study treating aged mice with PTH(1-34) for 4 wk showed that aged mice experienced greater increase in spinal BMD and a larger increase in osteoblast number than young adult mice, indicating that age did not attenuate the anabolic response. 75 In vitro PTH(1-34) treatment increased proliferation of MSCs isolated from younger human donors, while it had no effect on MSCs isolated from older donors, hinting that progenitors from elderly patients may be less responsive or fewer in number. 76 Lastly, a meta-analysis of 15 clinical trials reported that the increase in spine BMD after teriparatide treatment was blunted by increasing age. 77 However, this analysis could be confounded by age-related factors such as spinal osteoarthritis (which can artifactually affect BMD measurements). Overall, whether age-related osteoprogenitor decline limits teriparatide’s long-term efficacy remains uncertain. Additional studies are needed to clarify if older osteoporosis patients exhibit a faster waning of response due to diminished progenitor supply or altered PTH signaling. In conclusion, there is currently no direct evidence that prolonged PTH(1-34) treatment truly depletes osteoprogenitors in vivo, nor that actively restoring the osteoprogenitor pool reverses the waning anabolic response. Future research into this important question must therefore address several key challenges. First, the pronounced heterogeneity within the SSPC landscape poses a challenge. Distinct human adult SSPC populations have been identified in specialized niches of long bones, including CD146+ cells in the bone marrow stroma and CD164+, CD73+, and PDPN+ cells in the periosteum and upper metaphysis. 78 , 79 In addition, vertebral skeletal stem cells expressing XIC1 and PAX1 among others have been described. 80 These populations of stem cells are further influenced by aging. CD200 has been proposed as a marker distinguishing fetal from adult stem cells. 74 Further studies should therefore examine the effects of PTH(1-34) on these SSPC subsets in aged populations. Lineage tracing in studies in animal models can provide important mechanistic insides but must be long-term and focus on aged or osteoporotic animal models. Second, possible osteoprogenitor exhaustion through prolonged PTH(1-34) treatment might involve other aspects of osteoprogenitor physiology than proliferation such as altered fate regulation and senescence. 81 , 82 The effect of long-term PTH(1-34) treatment on these factors should be considered in future studies.

The interaction between mechanical loading of the skeleton and PTH’s effects is another factor that may limit long-term bone anabolism. The mechanostat hypothesis, first described by Frost in 1987, proposed that bone mass is regulated by homeostatic feedback loops responding to mechanical strain. When the set-point for bone mass changes, there is an initial rapid adaptation of bone mass toward the new set-point, followed by stabilization at a higher level once the new steady-state is reached. In the case of teriparatide, treatment increases bone mass without a corresponding increase in mechanical usage (patients’ physical activity typically remains the same). As bone mass rises while mechanical load stays constant, mechanical strain per unit bone mass gradually decreases. This could trigger a feedback response that slows further bone gain—essentially, bone “senses” that it is now carrying the habitual loads more easily (lower strain) and thus reduces the stimulus for additional formation. It has been hypothesized that this mismatch between increased bone mass and unchanged mechanical strain may contribute to the drug’s waning effects. 118 If so, one would predict that low mechanical strain (disuse or immobilization) would limit teriparatide’s anabolic action, whereas high mechanical strain (exercise or loading) would augment it. In experimental settings, mechanical strain on bone comes from two sources: the strain generated by normal daily activities and additional externally applied loads (exercise regimens or device-based loading). At a given level of daily activity, a smaller bone mass experiences higher strain than a larger bone mass (since the same force is distributed over less area). Therefore, when teriparatide increases bone mass without a change in activity, the relative strain per bone decreases. If PTH’s efficacy is indeed strain-sensitive, then a scenario of low bone mass combined with external loading should yield the greatest response to PTH, while a high bone mass model with no mechanical loading (eg, immobilization) should yield the poorest response ( Table 5 ). In support of this idea, it is shown that genetically high-bone-mass mice (sFRP1 or Sost KOs) show blunted responses to PTH, consistent with the notion that their bones, already strong and lightly strained under normal activity, get less benefit from PTH. 114 , 119 The following section will investigate the effects of external loading on the effects of PTH(1-34). The expected responsiveness to PTH(1-34) from combined mechanical strain. Several studies have examined if skeletal disuse attenuates the effects of PTH(1-34). One study subjected growing rats to 8 d of skeletal unloading with or without concurrent treatment with PTH(1-34). Rats experienced significant bone loss in response to unloading. While PTH(1-34) could not completely restore periosteal bone formation and overall tibial mass, skeletal unloading did not reduce the responsiveness of cancellous bone to PTH(1-34). 120 Therefore, animals with low bone mass and low mechanical load still experienced a response to PTH(1-34). Another study immobilized mice by tail suspension for 8 or 15 d. Treatment with PTH(1-34) still had positive effects on bone volume fraction in the proximal tibia on day 8, while there were no effects on day 15. Moreover, PTH(1-34) treatment still increased expression of osteocalcin, osterix, and PTH1R expression on day 15. 121 Lastly, a study investigated the effects of 2 wk of hindlimb unloading on adult male rats. PTH(1-34) treatment prevented bone loss in cancellous but not in cortical bone. 122 PTH(1-34) treatment has also proven effective in treatment of immobilized patients. In patients with spinal cord injury, 12 mo of teriparatide therapy resulted in significant increases in spine aBMD compared with baseline. 123 Another study compared the effects of teriparatide in patients with low levels of walking state to patients with high-level walking state. Here, the authors found similar increases in LS-BMD in both subgroups. 124 Overall, these studies indicate that immobilization or unloading possibly reduces but does not abolish the osteoanabolic effects of PTH(1-34). Consistent with the concept that mechanical input modulates the anabolic response to PTH, a remobilization study showed that in rats subjected to 19 wk of hindlimb immobilization, followed by continued immobilization or remobilization with concurrent PTH(1-34) treatment, remobilized limbs exhibited higher periosteal bone formation rates than immobilized limbs, 125 indicating synergy between mechanical stimulation and PTH(1-34). Several studies indicate that adding mechanical loading enhances PTH’s effects, particularly in cortical bone. For example, one study evaluated the effects of uniaxial compressive loading of the right tibia on PTH(1-34) treatment in ovariectomized mice. While there were no additive effects of the two treatments in cancellous bone, cortical thickness was higher with combination treatment than with PTH(1-34) treatment alone. The increased benefits were most prominent in those areas of the tibia subjected to higher strains under compressive loading. 126 In 13-wk-old mice, combined mechanical stimulation and intermittent PTH(1-34) treatment for 6 wk yielded synergistic effects on tibial cortical bone volume. 127 Moreover, 2 studies reported increased effects of combined compression of tail vertebrae and PTH(1-34) on bone formation rate in rats after a single injection, 128 or after 2 and 4 wk of treatment. 129 This result was confirmed in a recent study, showing additive effects of tail compression and PTH(1-34) on predicted strength and static morphometric parameters in the sixth caudal vertebrae of female mice. 130 In rats subjected to overload of the right hindlimb and concurrent immobilization of the left hind limb, PTH(1-38) increased fracture load only in overloaded but not immobilized hindlimbs. 131 In rats subjected to external loading by a 4-point bending device and PTH(1-34) for 3 wk, combined treatment resulted in synergistic effects on cortical bone formation surface, mineral apposition rate, and bone formation rate. 132 These studies confirm the beneficial effects of high external loading in combination with PTH(1-34) on skeletal parameters. Positive effects of mechanical stimulation have also been observed in clinical studies. One study treated 40 patients with teriparatide for 6 wk prior to hip replacement surgery and investigated the impact of the treatment on the 2 areas of the FN that undergo tension vs compression under normal daily activity. The endocortical bone formation rate was greater on the tensile region of the FN than on the compressed region of the FN, indicating that mechanical stimulation affects the anabolic response of bone to teriparatide. 133 In addition, an RCT, comprising 35 postmenopausal women, reported increased effects of teriparatide on LS-BMD when combined with whole body vibration exercise for 12 mo. 18 On the contrary, a study investigating the combined effect of mechanical loading and PTH(1-34) in aged 19-mo-old female mice found that PTH(1-34) abrogated the positive effects of loading on trabecular bone volume and trabecular thickness. In cortical bone, mechanical loading and PTH(1-34) had additive or independent effects on cortical thickness and porosity, 134 suggesting that mechanical loading does not affect the bone anabolic response to PTH(1-34) in aged mice. In conclusion, several studies have shown either additive or synergistic effects of concurrent external loading and PTH(1-34) treatment, while low mechanical strain seems to limit the effectiveness of PTH(1-34). These results support the hypothesis that lower mechanical strain could limit the long-term effects of PTH(1-34). However, aging may limit the sensitivity of bone to mechanical loading. Therefore, it remains unclear to what extent mechanical loading plays a role in the waning effects of teriparatide in older osteoporotic populations. From a practical standpoint, incorporating mechanical loading (eg, exercise) during teriparatide therapy may help to sustain its benefits, particularly in younger patients, although age-related declines in mechanosensitivity could dampen this advantage.

The downstream signaling pathways of PTH are intricate and could be altered during long-term treatment with teriparatide. PTH binds to PTH1R, a G-protein coupled receptor that activates G s and G q proteins. Once phosphorylated, PTH1R can recruit β-arrestin 2, leading to either dissociation from the G-protein and subsequent internalization of the β-arrestin-PTH1R complex or internalization of a β-arrestin-PTH1R-G-protein complex. The internalized complex can be either degraded or recycled. 93 Additionally, the internalized β-arrestin-PTH1R-G-protein complex can continue to signal, prolonging the cAMP response. 135 Several components of this downstream signaling pathway could contribute to the waning effects of teriparatide. First, PTH1R could be affected by receptor downregulation or reduced receptor affinity in osteoprogenitors upon extended treatment with PTH. Pretreatment of primary calvarial cells with PTH(1-34) for 24-48 h downregulated PTH1R expression in vitro. 136 Furthermore, case studies have reported acquired PTH resistance due to autoantibodies to PTH1R. 137 However, this phenomenon has not been observed during teriparatide treatment. Second, both insufficient and excessive desensitization could limit the bone-anabolic effects of teriparatide. Insufficient desensitization may prolong signaling, resulting in increased activation of bone resorption relative to formation, while excessive desensitization could reduce receptor availability. Repeat dosing of PTH(1-34) caused refractory cAMP responses to subsequent PTH(1-34) challenge in several studies. 138–140 This desensitization involved internalization of the receptor, 139 activation of kinases, such as PKA and PKC, 138 and β-arrestin recruitment. 140 The Na+/H+ exchange regulatory factor 1 (NHERF1) inhibited desensitization by preventing interaction with β-arrestin. 140 Notably, PTH(1-34)-treated β-arrestin-2 (−/−) mice had higher periosteal bone formation rates compared to PTH(1-34)-treated β-arrestin-2 (+/+) mice, suggesting that limiting desensitization through β-arrestin inhibition may enhance the osteoanabolic effects of teriparatide. 141 Conversely, abaloparatide, a PTH-related protein analogue, induced larger increases in cortical bone thickness compared to teriparatide in vivo, despite causing greater β-arrestin recruitment in vitro. 142 This finding implies that β-arrestin recruitment may be positively associated with cortical bone gain. However, 2 other studies showed that abaloparatide causes less 143 or similar 144 β-arrestin recruitment compared to teriparatide in PTH1R-expressing HEK293 cells, indicating that the precise role of β-arrestin recruitment remains incompletely understood. In addition, the impact of long-term teriparatide treatment on β-arrestin recruitment has not been characterized further. Overall, the downstream signaling pathways activated by PTH warrant further investigation, at the preclinical and clinical level, in relation to the waning of its anabolic effects.