What is PGD₂?

Prostaglandin D₂ is a lipid mediator. It belongs to a family of small molecules called prostanoids, all of which derive from arachidonic acid, a fatty acid released from cell membranes. Cells make prostanoids on demand, in response to local conditions, and they act locally. PGD₂ circulates briefly in tissue, binds to receptors on nearby cells, and is then enzymatically degraded. It is not a hormone in the classical sense. It is closer to a short-range chemical message that tunes cell behavior.

PGD₂ was first studied in the context of inflammation, allergy, and sleep regulation. It is the dominant prostanoid released by mast cells during an allergic response. Its role in hair biology was identified much later, and that role turned out to be central. Hair follicles synthesize PGD₂ through an enzyme called prostaglandin D₂ synthase (PTGDS), and they express the receptors required to respond to it. The follicle is both a source of PGD₂ and a target of it.

The 2012 University of Pennsylvania study

In March 2012, researchers at the Perelman School of Medicine at the University of Pennsylvania, led by George Cotsarelis, published a paper in Science Translational Medicine that reframed how researchers think about androgenetic alopecia. The team compared scalp biopsies from bald and haired regions of the same men with male-pattern hair loss. Using gene expression analysis and direct measurement of prostaglandins, they found that prostaglandin D₂ synthase (PTGDS) and its product PGD₂ were strikingly elevated in the bald regions.

PGD₂ was approximately three times higher in bald scalp than in adjacent haired scalp on the same person. PGE₂ and PGF_2α, the prostaglandins associated with follicle growth, were lower in the bald areas. The team then tested PGD₂ directly: when applied topically to mice, PGD₂ inhibited hair growth. Knockout mice lacking the PGD₂ receptor GPR44 were resistant to this suppression. Cultured human hair follicles exposed to PGD₂ shrank visibly within a few days.

The finding mattered for two reasons. First, it identified a measurable biochemical signature of androgenetic alopecia, not just a downstream consequence. Second, it produced a druggable target. Within two years, pharmaceutical companies were testing GPR44 antagonists as potential hair-loss therapies. The trajectory of that work is part of the rest of this article.

The prostaglandin balance theory

The literature now supports a model in which different prostaglandins exert opposing, complementary effects on the hair growth cycle. PGE₂ and PGF_2α are pro-anagen, meaning they support the active growth phase. PGD₂ and PGI₂ are pro-catagen and pro-telogen, meaning they push follicles toward regression and rest. The follicle is regulated not by any one of these molecules in isolation, but by their relative concentrations at the follicle bulb at any given moment.

This explains several otherwise odd observations. Female scalps tend to produce more PGE₂ and PGF_2α than male scalps, which may partly account for the lower lifetime incidence of androgenetic alopecia in women. The biochemical signature of androgenetic alopecia, an inverted PGD₂:PGE₂ ratio, is more pronounced in bald scalp than in scalp from the same person where hair is still present. And minoxidil, one of two FDA-approved treatments for androgenetic alopecia, achieves part of its effect by stimulating COX-1 and thereby increasing PGE₂ production in dermal papilla fibroblasts. Even when researchers had not yet identified the underlying mechanism, the most effective interventions for hair growth were already nudging the prostaglandin balance toward the pro-growth side.

GPR44, the receptor that lets PGD₂ shut follicles down

PGD₂ can bind two receptors, DP1 (PTGDR) and DP2 (also called GPR44 or CRTH2). Hair growth inhibition is mediated through DP2, not DP1. GPR44 is strongly expressed in the outer root sheath of the hair follicle, the cellular layer that surrounds and supports the growing hair shaft, and is more weakly expressed in dermal papilla cells, the signaling hub at the base of the follicle that orchestrates the growth cycle.

When PGD₂ activates GPR44, it triggers a Gi-protein-coupled signaling cascade that inhibits adenylyl cyclase. Adenylyl cyclase is the enzyme that makes cyclic AMP (cAMP). Lower cAMP means less downstream signaling through protein kinase A (PKA), less phosphorylation of CREB, less upregulation of growth factors like VEGF, and less of the metabolic activity required to sustain a hair in its growth phase. The follicle is pushed toward catagen, the regression phase.

The inhibitory effects of PGD₂ extend beyond cAMP suppression. PGD₂ is spontaneously converted to a metabolite called 15-deoxy-Δ^12,14-prostaglandin J₂ (15-dPGJ₂), which induces apoptosis (programmed cell death) in follicular keratinocytes at concentrations starting around 5 μM. PGD₂ also selectively reduces proliferation markers in the K15-positive stem cell population at the follicle bulge, the regenerative compartment that fuels each successive growth cycle. In mouse models tracking the natural follicle cycle, PTGDS expression and PGD₂ levels peak immediately before the onset of catagen, suggesting PGD₂ functions as an endogenous catagen-triggering signal.

Why setipiprant failed

Setipiprant is an oral, selective antagonist of the DP2/GPR44 receptor. It was originally developed for asthma and allergic conditions, where blocking PGD₂ signaling has clear therapeutic logic. After the 2012 Penn study, the same logic was applied to hair loss: if elevated PGD₂ drives follicle miniaturization, then blocking the receptor through which PGD₂ acts should restore growth.

A phase 2a trial enrolled men aged 18 to 49 with androgenetic alopecia. The study ran for 32 weeks. Setipiprant was well tolerated, with no significant safety signals. It did not, however, produce statistically significant hair growth compared to placebo. The trial established the safety of the molecule and the feasibility of the receptor target, but it did not establish efficacy.

The failure clarified something important about the biology. Blocking PGD₂ at the receptor removes an inhibitory signal, but removing an inhibitor is not the same as installing an activator. A follicle that has been suppressed by PGD₂ for years has lost more than active inhibition. The local concentrations of pro-growth prostaglandins, especially PGE₂, are reduced. The follicle bulge stem cells may be quieted. The vascular supply may have regressed. Releasing the brake does not press the accelerator.

Several reviews now argue that effective intervention in the prostaglandin axis likely requires both arms: suppression of PGD₂-driven inhibition and active stimulation of the PGE₂/PGF_2α pathway, or at least direct stimulation of the downstream cAMP machinery that PGD₂ would otherwise be suppressing. The setipiprant outcome remains the strongest pharmacological evidence that single-target PGD₂ blockade is not enough.

cAMP, Wnt/β-catenin, and the follicle's growth program

To understand why blocking PGD₂ alone falls short, it helps to look at what PGD₂ is suppressing. Inside dermal papilla cells, cyclic AMP (cAMP) acts as a master signal for growth-related activity. Elevated cAMP activates protein kinase A (PKA). PKA in turn phosphorylates a transcription factor called CREB, which upregulates the production of VEGF and other growth factors that sustain the follicle's metabolic demands during anagen. Independently, PKA phosphorylates and inactivates an enzyme called GSK-3β.

GSK-3β is the enzyme responsible for tagging β-catenin for destruction. When GSK-3β is inactivated by PKA, β-catenin accumulates inside the cell, then translocates to the nucleus, where it activates a set of transcription factors (LEF/TCF) that switch on the Wnt/β-catenin pathway. The Wnt/β-catenin pathway is the master switch for anagen initiation in follicle bulge stem cells. It controls whether dormant follicles re-enter the growth cycle. It also helps activate melanocyte stem cells, which is part of why prostaglandin-pathway activation is associated with darker, more pigmented lashes and brows in addition to longer ones.

What PGD₂ does, through GPR44, is sit on the top step of that chain and prevent it from starting. Blocking PGD₂ with setipiprant means the chain is no longer being actively suppressed. It does not mean a pro-growth signal is arriving to start the chain. That distinction explains the trial outcome.

Why eyelashes and eyebrows respond so strongly to prostaglandins

Of all the follicles in the body, lash and brow follicles are the most prostaglandin-responsive. The clearest evidence is accidental. When ophthalmologists began prescribing prostaglandin-analog eye drops (latanoprost, bimatoprost, travoprost) for glaucoma in the late 1990s, patients started reporting that their lashes were growing longer, thicker, darker, and more curved. The effect was consistent enough that bimatoprost was eventually reformulated and FDA-approved as Latisse® specifically for eyelash hypotrichosis.

Several biological factors converge to make lash and brow follicles unusually sensitive to prostanoid signaling:

• Short anagen phase. Eyelash follicles spend roughly 30 to 45 days in active growth, compared to two to eight years for scalp follicles. Each prostaglandin-signaling event therefore represents a much larger fraction of a lash follicle's biology than the same event in a scalp follicle.
• Receptor expression that mirrors growth phase. EP3 and EP4 receptors are expressed specifically during anagen in dermal papilla and outer root sheath cells of lash follicles. The FP receptor (the target of PGF_2α analogs like latanoprost) is also expressed in lash anagen tissue and not in resting follicles. The follicle is biochemically primed to receive prostaglandin signals during the brief window in which it is actually growing.
• Vellus-to-terminal conversion. Lash and brow regions contain large numbers of vellus hairs, the fine, lightly pigmented hairs that surround the more visible terminal hairs. Sustained prostaglandin-pathway activation can convert vellus hairs to terminal hairs, which increases visible density. This conversion requires a high-amplitude, sustained growth signal, which the lash follicle's biology is unusually well configured to receive.
• Melanocyte stem cell activation. The Wnt/β-catenin pathway, activated downstream of cAMP, also activates melanocyte stem cells, which explains why prostaglandin-pathway activation tends to darken lashes and brows in addition to lengthening them.

The shorthand: lash and brow follicles are tuned to listen for prostaglandin signals during their brief anagen windows, and they respond more dramatically than scalp follicles to any intervention that meaningfully shifts the prostaglandin balance.

Lash serums and prostaglandin biology

Topical lash serums fall into three categories based on how they engage the prostaglandin axis.

Prescription prostaglandin analogs (Latisse, latanoprost-based products)

These are direct receptor agonists. Bimatoprost is an analog of PGF_2α. It binds the FP receptor on lash dermal papilla cells and stimulates growth directly through that pathway. The efficacy is robust, with multiple randomized controlled trials demonstrating significant increases in lash length, thickness, darkness, and density. The side-effect profile is also well documented: iris hyperpigmentation, periorbital fat atrophy (sometimes called periorbital fat loss or prostaglandin-associated periorbital syndrome), eyelid skin darkening, conjunctival hyperemia, and dry eye. These are mechanism-linked, not formulation artifacts, which is why they appear across the prostaglandin-analog drug class.

Over-the-counter serums containing prostaglandin analogs (PGAs)

Several over-the-counter lash serums have, at various points, contained prostaglandin analog ingredients, most commonly isopropyl cloprostenate, isopropanol phenyl-hydroxy-pentene dihydroxycyclopentyl-heptenate, or dechloro dihydroxy difluoro ethylcloprostenolamide. These compounds activate the same prostaglandin receptor family as Latisse, often through the FP receptor, sometimes through dual or modified receptor targeting. The mechanistic story is similar to Latisse, but the regulatory and labeling story is messier: most are marketed as cosmetic ingredients despite acting on prescription-equivalent pathways. The same mechanism-linked side effects appear in this category.

Prostaglandin-free serums

A third category aims to engage the same pro-growth biology without using prostaglandin analogs. The strategies vary. Peptide-based serums attempt to stimulate keratin production or dermal papilla activity through different pathways. Plant-derived serums target the cAMP cascade through compounds that activate adenylyl cyclase directly or that bind native prostaglandin receptors as natural agonists. The mechanistic story here connects to the cAMP and Wnt/β-catenin chain described earlier: any intervention that drives cAMP upward in dermal papilla cells without engaging the prostaglandin-analog side-effect profile is, in principle, working with the same downstream growth biology.

Plume Science's Elite Lash & Brow Enhancing Serum is in the third category. It uses a patented composition called the C² Complex (US Patent 11,045,444) that combines ricinoleic acid, the principal fatty acid in castor oil, with forskolin, a labdane diterpene from Coleus forskohlii. Ricinoleic acid has been identified in a 2012 PNAS paper as a direct agonist at EP3 and EP4 prostanoid receptors, the same receptor subtypes engaged by endogenous PGE₂. Forskolin activates adenylyl cyclase directly, raising cAMP independently of receptor binding. The mechanistic logic is to lift cAMP through two non-redundant routes, while avoiding the prostaglandin-analog drug class and its associated side-effect profile.

Related Plume Science explainers: Prostaglandin-free lash serum guide · Latisse side effects, explained · Orbital fat loss and lash serums · Does GrandeLash contain a prostaglandin analog?

Research limitations and open questions

Most of what is reliably known about prostaglandin biology in hair follicles comes from cell culture, mouse models, and observational clinical data from glaucoma drug use. Several important questions remain.

• EP3 isoform ambiguity. The EP3 receptor can couple to multiple G-proteins (Gi inhibitory, Gs stimulatory) depending on which splice variant is expressed in a given tissue. The dominant pro-growth effect in follicles appears to be mediated by EP4, but the specific EP3 splice variants in eyelash dermal papilla cells have not been fully characterized.
• PTGDS inhibition is investigational. Several natural compounds, including ricinoleic acid, have been identified in computational screens as predicted inhibitors of prostaglandin D₂ synthase. Confirmation of these predictions in human follicle tissue remains outstanding. The current evidence is suggestive, not definitive.
• Lack of definitive human trials for non-prostaglandin-analog lash serums. Bimatoprost has multiple randomized controlled trials. Most over-the-counter serums, including those with mechanistically coherent compositions, rely on smaller, often open-label studies. Head-to-head trials against prescription standards do not exist in the published literature.
• Pathway complexity. The prostaglandin axis interacts with androgen signaling, Wnt signaling, growth factor pathways, and the local follicular immune environment. Reducing hair loss biology to any one pathway, including prostaglandins, is a simplification. The 3× elevation of PGD₂ in bald scalp is a real finding. It is not the only finding that matters.
• The scalp-versus-lash distinction. Most of the foundational prostaglandin-follicle research was done in scalp tissue or in mouse models. The unusually strong response of eyelash follicles is well documented clinically, but the molecular fine print of why lashes respond so much more than scalp is still being characterized.

Treat the prostaglandin balance theory as a coherent, well-supported framework with meaningful open questions, not as a closed scientific case.

Frequently asked questions