Mechanisms underlying the waning of osteoanabolic therapy effects in osteoporosis

Abstract

Osteoanabolic therapies such as parathyroid hormone (PTH) analogs (teriparatide [TPT], abaloparatide [ABL]) and the sclerostin inhibitor romosozumab rapidly stimulate new bone deposition but their bone-building effects are not indefinite. Clinical trials including bone turnover markers and bone histomorphometry studies show that the initial gains in bone formation and density wane with continued therapy—typically within 12-18 mo for PTH analogs and even sooner for romosozumab. This invited review explores the biological mechanisms that may explain this waning response, including receptor desensitization at the PTH1 receptor level, depletion or saturation of bone-forming surfaces, exhaustion or downregulation of osteoblast progenitors, upregulation of endogenous Wnt pathway antagonists (eg, sclerostin, DKK1), and mechanostatic feedback from increased bone mass. The contrasting kinetics of TPT, ABL, and romosozumab are compared, highlighting how their modes of action (remodeling- vs modeling-based formation) influence the duration of anabolic activity. Finally, potential strategies to extend or rekindle anabolic effects are discussed—including sequential and combination therapies and adjunctive mechanical loading—with a view toward maximizing bone gains and reducing fracture risk. Understanding why anabolic drug efficacy attenuates provides insight into bone biology and guides the design of regimens to achieve more sustained increases in bone mass and strength.

Introduction

Osteoanabolic therapies, namely parathyroid hormone (PTH) receptor agonists—teriparatide (TPT) and abaloparatide (ABL)-, and sclerostin inhibitors, -romosozumab-, have significantly improved osteoporosis management by directly stimulating new bone formation and are now largely advocated as first-line therapy in patients at very high fracture risk.¹ Indeed, through their direct and/or indirect action on osteoblasts, these drugs induce large and rapid gains in BMD—although at spine more than the hip, which could be at least partly explained by the larger bone surface area in cancellous than cortical bone-, improve bone microstructure, and eventually reduce fracture incidence as compared to oral bisphosphonates and/or placebo.^2–6 Although continuous BMD gains are observed for the duration of therapy, ie, 12 mo with Romosozumab, 18 mo with ABL, and up to 24 mo with TPT,⁷ clinical observations and trial data indicate that the early robust anabolic effects wane earlier. For example, increases in bone turnover markers (BTMs: P1NP and CTx) and BMD gains in response to TPT are most pronounced in the first year and attenuate thereafter, whereas ABL induces a less robust increase in BTMs, with P1NP already declining by 50% between 6 and 12 mo of administration, whereas CTX barely raises above baseline (<20% at peak).⁴ Romosozumab exhibits an even earlier waning of bone formation: after an initial surge in P1NP within the first 2-3 mo, its anabolic activity subsides completely by 9 to 12 mo despite ongoing treatment.² Importantly, with sclerostin inhibition there is a dissociation of the bone anabolic stimulus and bone resorption, as CTX levels are durably suppressed by about 25%-40%. These patterns suggest that lower levels of remodeling-based bone formation as well as intrinsic regulatory mechanisms in bone curtail the anabolic response over time. Understanding why the bone-building response to therapy diminishes is crucial for optimizing treatment sequences and developing strategies to achieve a more sustained benefit on skeletal health. In this review, we synthesize findings from clinical trials, biomarker studies, and bone biology research to discuss the possible mechanisms underlying the waning of osteoanabolic effects. We also discuss potential approaches to extend the anabolic window or amplify bone gains, thereby moving closer to the goal of building sufficient bone mass to durably reduce fracture risk.

Histomorphometry confirms waning osteoanabolic effects

In keeping with the pattern of BTMs described above, bone biopsies from the TPT-Alendronate comparator trial indicate that the stimulation by TPT of both the activation frequency and bone formation rate declines by 50% between 6 and 18 mo of exposure.^8,9 In the Shotz Study comparing TPT and Zoledronate, the percentage of bone remodeling and, to a lesser extent, modeling surfaces induced by TPT in every compartment (cancellous, endocortical, and periosteal) at 6 mo was markedly attenuated, if not entirely gone, by 24 mo.¹⁰

Moreover, iliac crest bone biopsies taken at 12-18 mo in the ACTIVE trial failed to show a significant increase in most bone-forming indices by either TPT or ABL compared to placebo (besides a modest increase in the mineral apposition rate with TPT), further demonstrating the waning of their anabolic effects before the end of therapy.¹¹ Hence the modest increase in BMD after 12 m. of therapy with these PTHRc agonists (ie, 1.5%-2% in m. 12-18 compared to 5.5%-6.5% in m. 0-6) can probably be explained by some residual bone-forming effects of the drugs but mostly by secondary mineralization of the newly formed bone.

Consistent with the biochemical markers, histomorphometry from the romosozumab trial biopsy sub-study showed an early burst of bone formation that diminished with time.^12,13 At 2 mo, romosozumab-treated patients had significant increases in dynamic formation indices at both trabecular and endocortical surfaces, with evidence that a large proportion of this new bone formation was modeling-based, occurring on previously quiescent surfaces without prior resorption.¹³ However, after 12 mo of romosozumab, bone formation indices were not only reduced from their early peak, but were actually lower than those in placebo-treated patients.¹² At that time, bone resorption remained below baseline—an outcome indicative of self-limited anabolic activity leading into a predominantly antiresorptive mode.¹²

Differential mechanisms of PTHRc agonists vs Romosozumab on bone formation

The 3 anabolic therapies differ in the magnitude and duration of bone formation they induce, which reflects their differential mechanisms of action. As explained above, TPT acts predominantly on remodeling, and to a lesser extent on modeling, surfaces.¹⁰ It stimulates new bone formation eventually coupled to increased bone resorption, ie, by increasing activation frequency of remodeling units.¹⁰ This leads to a progressive creation of new bone multicellular units (BMUs), especially on trabecular bone, and endocortical surfaces over time, whereas within cortical bone initiation of a BMU creates a resorption cavity (osteonal canal), which refilling is slow and often incomplete, leading to cortical porosity. Abaloparatide, acting on the same PTH1 receptor but inducing a modified cAMP signaling vs TPT (see below), exerts a more transient effect on resorption.^14,15 Head-to-head data from ACTIVE showed that ABL caused a smaller rise in CTX than TPT for a given increase in PINP, suggesting relatively more bone formation vs resorption (“anabolic window”).^4,16 Clinically, ABL achieved significantly greater BMD gains at the hip than TPT from 6 to 18 mo,⁴ which may reflect these differences. In contrast, romosozumab triggers modeling-based bone formation to a much larger extent initially.¹³ This results in rapid deposition of bone on periosteal, endocortical, and trabecular surfaces without needing prior osteoclast activity. The FRAME biopsy analysis confirmed that ~90% of the new bone formation at 2 mo with romosozumab was modeling-based (with smooth cement lines) rather than remodeling-based (scalloped cement lines).¹³ However, this uncoupled formation is short-lived. Once those surfaces have been activated and packed with new bone, continuing romosozumab cannot further increase formation unless new surfaces are made available. This may also explain why romosozumab effects are markedly attenuated after antiresorptive therapy.¹⁷

Taken together, these observations first suggest that the duration of bone-forming effects induced by osteoanabolics is proportional to the stimulation of bone turnover: with lesser (ABL) or even negative effects (romosozumab) on bone resorption, a relative saturation of the remodeling space by newly formed bone may occur earlier, while an extensive activation of quiescent bone surfaces (romosozumab > PTHRc agonists) may also rapidly exhaust the potential of modeling-based bone formation. In addition, the surface-to-volume ratio changes as bone mass increases—thicker trabeculae and cortices mean relatively less surface area per unit volume on which osteoblasts can act. For instance, doubling the romosozumab dose in a rodent model after the initial anabolic effects had waned, failed to extend bone-forming surfaces and/or to further improve bone micro-architecture,¹⁸ implying that simply providing more drug cannot overcome the potential limitation of available bone surface.

Molecular and cellular mechanisms of waning anabolic response

Multiple non-mutually exclusive mechanisms have been proposed by which bone tissue adapts or “desensitizes” to continued osteoanabolic stimulation. These operate at the receptor signaling level, the cellular level, and the tissue/system level (Figure 1).

Figure 1.

Schematic representation of the main tissular, cellular, and molecular processes that limit the sustained anabolic response to prolonged treatment with TPT, ABL, and ROMO. (1) Reduced availability of activatable bone surfaces for bone remodeling during TPT/ABL therapy and modeling surfaces under ROMO progressively limits the capacity for new bone accrual. (2) Decreased mechanical strain, a consequence of increased bone mass (mechanostat feedback), attenuates osteocyte mechanotransduction and contributes to the decline in bone formation during all 3 therapies. (3) Upregulation of endogenous Wnt pathway inhibitors, including SOST and DKK1, occurs both through osteocyte-intrinsic negative feedback and extrinsically as a result of reduced mechanical strain; this mechanism is most prominent with ROMO but also contributes under TPT and ABL. (4) PTH1R desensitization and downregulated downstream signaling emerge with chronic exposure to TPT and ABL, blunting the intermittent activation that normally drives their anabolic effect. (5) Reduced osteoblast number and functional capacity (“vigor”), particularly under prolonged ROMO therapy, further limits sustained bone formation. (6) Exhaustion of the osteoprogenitor pool, especially relevant for ROMO, constrains the availability of responsive precursor cells required to maintain an anabolic response over time. Abbreviations: ABL, abaloparatide; ROMO, romosozumab; TPT, teriparatide; PTH1R, PTH1 receptor.

PTH1 receptor desensitization and signaling adjustments

Continuous or repeated activation of G protein-coupled receptors like the PTH1 receptor (PTH1R) can lead to receptor down-regulation or signal attenuation within target cells. In vitro studies demonstrate that PTH1R signaling is subject to regulation by C-terminal phosphorylation and β-arrestins, which can uncouple the receptor from G-protein signaling and target it for internalization.^19–21 In fact, PTH1R can adopt distinct conformations when bound to different ligands, and these conformations influence whether the receptor is recycled or degraded after internalization.²² A possibility is that TPT and ABL, though both agonists of PTH1R, may not engage identical signaling cascades or receptor trafficking pathways. Abaloparatide has been reported to cause more transient cAMP signaling and less sustained receptor occupancy than TPT, theoretically resulting in less receptor down-regulation per dose.¹⁵ Nonetheless, with daily stimulation over many months, PTH1Rs on osteoblasts and osteocytes could undergo some degree of down-regulation (reduced cell-surface expression) and desensitization (diminished second-messenger response). This would blunt cellular responses to PTH analogs over time, contributing to the reduced BTM levels seen during continuous therapy. Evidence for PTH1R desensitization includes preclinical findings that continuous infusion of PTH (as opposed to daily injections) leads to an attenuated bone formation response, as well as cellular data showing PTH1R coupling to adenylyl cyclase is dampened after prolonged exposure due to β-arrestin recruitment.¹⁹ Designing agonists that bias signaling away from G_s/cAMP toward alternate pathways (β-arrestin-biased ligands) has been explored as a means to preserve bone formation with less desensitization.²⁰

In summary, on the receptor signaling level, tachyphylaxis from repetitive PTH1R stimulation could be one factor limiting sustained response to TPT and ABL.

Upregulation of endogenous Wnt inhibitors (negative feedback)

Perhaps the clearest molecular feedback mechanism is seen with romosozumab and the Wnt signaling pathway. Sclerostin neutralization leads to strong activation of Wnt/β-catenin signaling in osteoblasts, driving bone formation. The skeleton responds by increasing production of several Wnt antagonists. Holdsworth et al. and others demonstrated in mice that after one dose of sclerostin antibody (Scl-Ab), there was a significant increase in the mRNA expression of multiple secreted Wnt inhibitors—including sclerostin itself (Sost gene), DKK1, DKK2, Sostdc1, and sFRPs (secreted frizzled-related proteins).^18,23,24 After multiple doses, these antagonists were even more upregulated, and an additional inhibitor, sFRP5, was induced.²³ Recent clinical observations also suggest that romosozumab treatment is associated with a significant rise in serum DKK1 levels, which correlates with a decrease in bone formation markers over time.²⁵ This suite of Wnt pathway blockers would act to dampen further Wnt signaling and bone formation, establishing a negative feedback loop.

Parathyroid hormone analogs might also trigger counter-regulatory inhibitors of the Wnt pathway, though the specifics differ. Notably, short-term intermittent PTH decreases sclerostin production,²⁶ which is part of its anabolic mechanisms. However, some studies have found elevated circulating DKK1 or sclerostin levels with TPT,^27,28 suggesting a complex temporal pattern.

Overall, molecular counter-regulation by Wnt signaling antagonists—exemplified strongly with romosozumab therapy—likely plays a key role in the waning of osteoanabolic effects.

Osteoblast and osteoblast precursor dynamics

Another potential mechanism of waning is the exhaustion or quiescence of the osteoblast lineage with prolonged stimulation. Anabolic drugs initially stimulate mature osteoblasts and bone lining cells to ramp up matrix synthesis and tissue mineralization, thereby also accelerating the transition of these cells toward osteocytes. Moreover, osteoanabolics recruit osteoprogenitors from the marrow or bone lining into osteoblastic differentiation. However, continuous exposure to PTH or PTHrP analogs may cause pre-osteoblasts to differentiate more rapidly than they can self-renew. In the case of sclerostin inhibition, some data suggest that osteoprogenitor proliferation is curtailed after an initial wave of differentiation. Ominsky et al. compared Scl-Ab and intermittent PTH in young rats and found that sclerostin-Ab caused an initial spike in osteoblast numbers that could not be further increased with ongoing treatment, whereas PTH maintained osteoblast recruitment to a greater extent.²⁹ After 6 wk, sclerostin-Ab-treated rats had fewer osteoblast progenitors than PTH-treated rats, implying a feedback mechanism limiting the supply of new osteoblasts under sclerostin inhibition.²⁹ This aligns with the clinical biopsy finding that romosozumab’s bone formation is mostly a 1-t activation of lining cells (ie, it uses up the readily available mature cells).¹² If mesenchymal stem cells or pre-osteoblasts become scarce or less responsive, bone formation will slow. There is also the possibility of osteoblast senescence or “vigor” loss after extended high activity. Osteoblasts have a finite lifespan; an initial surge of formation might push a cohort of osteoblasts through their life cycle faster, and without a fresh supply, formation declines. Notably, when anabolic therapy is halted, the bone-formation capacity can often be rekindled after a period, suggesting the cells are not permanently exhausted and/or the pool of osteoblast precursors has regenerated. For example, in a mouse study, after an initial 6-wk course of sclerostin-Ab that caused bone formation to wane, a period off the drug allowed bone formation capacity to “reset” and respond again to a later course of treatment.²³ Similarly, in patients, a second course of romosozumab after a break can restimulate bone formation (P1NP) and BMD gain significantly.³⁰

Mechanostatic feedback (the “mechanostat” hypothesis)

The skeleton’s response to mechanical loading provides another possible mechanism to explain waning osteoanabolic drug effects. According to Frost’s mechanostat theory, bone mass increases to accommodate mechanical forces and will stop increasing once an optimal strain environment is reached.³¹ Anabolic therapy increases bone mass and strength, thereby reducing the mechanical strain for a given load. Osteocytes, which are mechanosensors, may then respond by emitting signals to slow bone formation (one such signal is sclerostin). This could act as a natural brake on further gains—essentially a homeostatic ceiling for bone mass. Hence, Gasser wrote in 2003 about TPT that “bone gain is not infinite and indeed is limited by a feedback mechanism (mechanical sensor).”³²

Preclinical suggestions for mechanostatic regulation come from studies where mechanical loading was superimposed on anabolic treatment. In mice on Scl-Ab, adding axial loading of the tibia prevented the usual increase in Sost (sclerostin) expression and sustained higher cortical bone formation than drug alone.^18,33 In addition, the mechanical stimulus restored periostin expression (a co-activator of Wnt signaling),³⁴ which was suppressed by Scl-Ab.¹⁸ Regarding PTH effects, Roberts et al. demonstrated that concomitant mechanical stimulation (bending) of the femur in rats exerted synergistic effects on the total endocortical bone apposition surfaces.³⁵

These findings suggest that the mechanostat can override pharmacologic stimulation: when mechanical usage becomes insufficient relative to the higher bone mass, even an anabolic agent cannot sustain bone formation indefinitely. However, higher mechanical loading may enhance and prolong the drug’s effect.

In summary, the attenuation of osteoanabolic drug effects can be attributed to a combination of receptor-level desensitization, activation of endogenous brakes (Wnt inhibitors, etc.), cellular source limitations, and biomechanical feedback once bone strength improves. These factors are interrelated and together ensure that bone formation induced by therapy does not run unchecked indefinitely.

Strategies to sustain or enhance anabolic efficacy

Given the above limitations, a key question is how to extend the anabolic window or amplify bone gains in practice. Several strategies have emerged, informed by both clinical studies and the mechanistic insights from research.

Sequential therapy (anabolic followed by antiresorptive)

While following anabolic therapy with an antiresorptive agent does not prolong the bone-forming phase per se, it consolidates the new bone by allowing its secondary mineralization and preventing its remodeling away. Indeed, studies have shown that without such follow-up, the BMD increments from osteoanabolics regress post-therapy, at least until bone turnover returns to baseline.^7,36,37 Thus, sequencing exploits the strengths of each class—anabolism first to build bone, then antiresorption to preserve it.

Combined anabolic + antiresorptive therapy

Combining an anabolic and antiresorptive simultaneously has been tested to see if one can obtain the high bone formation of an anabolic while curbing any resorptive increases, theoretically yielding a larger anabolic window and net bone gain. Early attempts with PTH + bisphosphonate were disappointing—alendronate seemed to blunt TPT’s effects, particularly on the trabecular spine, as they have “opposite” effects on the same remodeling surfaces.⁹ Nevertheless, combination therapy of TPT with either alendronate or zoledronate improved hip BMD more than TPT alone,^38,39 which probably reflects the net effects of TPT at endocortical surfaces when intracortical remodeling is concomitantly suppressed (see above). Results from the DATA study (Denosumab And Teriparatide Administration) further demonstrated that simultaneous denosumab and TPT led to additive increases in BMD far exceeding either agent alone, despite low levels of bone turnover.⁴⁰ The rationale is that TPT stimulates new bone formation on existing (pre-treatment) remodeling surface,s while denosumab prevents PTH-induced bone remodeling, thereby preserving the new bone formed. Due to the intrinsic anti-resorptive properties of romosozumab, it remains currently unclear whether combination therapy with an anti-resorptive will potentiate romosozumab effects. However, the diminished effects of romosozumab in subjects previously on anti-resorptives and whose bone turnover is low indicate that the reduction of active bone surfaces may actually blunt romososumab anabolic effects. The reduction of osteo-anabolic effects once resorption is reduced could also be partly explained by the deprivation of clastokines and other matrix-embedded growth factors that normally contribute to osteoblast activation.

Combined anabolic + anabolic therapy

Parathyroid hormone can reactivate bone formation in a skeleton already adapted to sclerostin inhibition—likely by stimulating remodeling-based bone formation and renewing the pool of osteoblast precursors.⁴¹ Preclinical research has explored PTH or PTHrP analogs combined with sclerostin inhibition. Since PTH signaling and Wnt signaling intersect (PTH can increase Wnt activity partly by lowering sclerostin in osteocytes), one might expect diminishing returns from simply combining them. Yet, studies in animal models have shown at least additive effects. In rodent models, combining PTH and Scl-Ab increases bone mass more than either treatment alone. For instance, Li et al. (2014)⁴² reported that the combination elicited greater osteoblast activity and increased osteoprotegerin levels, potentially enhancing bone formation while reducing resorption. Moreover, in mice pre-treated with Scl-Ab to the point where bone formation begins to decline, subsequent administration of PTH was able to rescue the bone forming indices, increase osteoblast numbers, and induce further bone accrual not achieved by Scl-Ab alone.⁴³

Adjusting duration or dosing

Another idea is altering the dosing schedule of anabolic therapy to prevent or delay waning of bone formation. For TPT, daily dosing is standard, but some studies have tried cyclic (eg, daily for 3 mo on, 3 mo off, repeated) or once-weekly regimens to allow periodic recovery. Results have been mixed, and these regimens are not clearly superior in terms of BMD gains.⁴⁴ To note that once weekly administration of a higher TPT dose (56 ug s.q instead of 20 ug) in the TOWER trial⁴⁵ led to significant BMD gains at the spine and hip over 18 m.; however, the profile of BTMs did not suggest an extended period of bone anabolism. In fact, P1NP peaked at 1 m. but then decreased below baseline by month 3, in parallel with the resorption marker NTx, for the duration of the study. The reasons for the apparent inhibition of bone resorption with this regimen, also seen with oral TPT,⁴⁶ remain unclear, but could be related to the suppression of endogenous PTH. For romosozumab, the rapid feedback inhibition by Wnt antagonists raises the question of whether a dosing pause could restore responsiveness. The preclinical study by Holdsworth et al. showed that after a 4-wk holiday from Scl-Ab, the next dose was able to trigger a fresh PINP rise similar to the initial response.²³ In humans, shorter courses and/or lower doses of romosozumab that would still trigger a potent stimulation of bone formation without necessarily activating all bone surfaces at once nor exhausting the pool of osteoblastic precursors could be an interesting approach. In practice, however, withholding an anabolic drug means forfeiting its benefits during that time and risking that any gain could be lost if not protected. Therefore, such schemes would need to be balanced with another therapy during the off-period (for instance, a short antiresorptive cover), although this may blunt the second response to osteoanabolics.

The opposite approach, namely to gradually increase the dose of romosozumab in order to neutralize the higher levels of sclerostin induced by the treatment (see above), has also been tested pre-clinically (in rodents and non-human primates), showing some rescue on P1NP levels at least.¹⁸ Nevertheless, no further gains in bone mass and microstructure were observed, perhaps because of the exhaustion of bone forming surfaces, as explained above.

Mechanical loading and exercise

As hinted earlier, mechanical stimulation of the skeleton (weight-bearing exercise, resistance training, or targeted vibration/loading) is a potent anabolic stimulus in its own right. Combining exercise with drug therapy could amplify bone gains. Animal studies have shown that mechanical loading can synergize with PTH on cortical bone formation³⁵ and counteract the decline in response to Scl-Ab by upregulating mechano-responsive pathways.¹⁸ In humans, clinical trials such as the LIFT study (lifting plus Forteo) are exploring if supervised exercise can augment TPT’s effects. Even outside of trials, it is biologically plausible that patients who engage in regular impact and resistance activities will direct more of the drug-stimulated formation to critical regions (such as the hip) and possibly keep osteocytes from producing more sclerostin to brake on bone formation. Exercise could also recruit dormant lining cells or progenitors via mechanotransduction signals, adding to the pool of bone-building cells.

Looking ahead, emerging strategies like novel PTH1R agonists that avoid desensitization²⁰ and combinations of Wnt pathway inhibitors (eg, simultaneous blockade of sclerostin and DKK1)⁴⁷ might further extend the duration of osteoanabolic effects.

Conclusion

Osteoanabolic therapies have transformed osteoporosis treatment by enabling actual restoration of bone structure and strength, but their benefits accrue primarily during a finite window of time before the effect diminishes. Anabolic effects wane by the coordinated feedback mechanisms at play in the skeleton. As reviewed here, bone’s remodeling system is tuned to maintain homeostasis: intense stimulation of bone formation triggers countermeasures at molecular, cellular, and tissue levels to prevent unchecked bone accrual. The waning of anabolic drug effects should not be viewed as a therapeutic failure, but rather as the bone’s normal adaptation to an extraordinary stimulus. Importantly, many of these limiting factors—whether receptor down-regulation, limited progenitor supply, or Wnt antagonist upregulation—could perhaps be at least partially mitigated. Higher dosing alone is generally not the answer (as seen with Scl-Ab experiments), but adjunct approaches are promising. Combining treatments that address different pathways (or combining pharmacotherapy with mechanical stimulation) has already shown synergistic outcomes in some studies and animal models and should be further clinically tested.

Author contributions

Serge Ferrari (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing), Cyril Thouverey (Investigation, Validation, Visualization, Writing—review & editing), and Maude Gerbaix (Conceptualization, Data curation, Investigation, Methodology, Supervision, Validation, Visualization, Writing—original draft, Writing—review & editing)

Funding

None declared.

Conflicts of interest

Serge Ferrari: Scientific advisory boards and/or educational lectures sponsored by AMGEN, UCB, Agnovos, Fresenius, Flowbone, Gedeon Richter, Alexion, Angita. Cyril Thouverey: None. Maude Gerbaix: None.

Data availability

No new data were generated or analyzed in support of this review article.