We talk about testosterone like it comes from a factory that runs forever. It doesn't. The factory has a construction phase, and it ends around your twenty-second birthday. After that, what you have is what you carry for life.
This article is about the cells that make testosterone — Leydig cells — and the discovery that they are a finite, non-renewable resource with a developmental window that closes. It connects six of my previous articles through a single principle: damage that arrives before the factory finishes building itself is damage that lasts forever.
The Hormone That Measures the Factory, Not the Output
Testosterone tells you how much hormone is being produced right now. It says nothing about the capacity of the cells producing it. A factory running at 100% of a damaged capacity and one running at 50% of a healthy capacity can produce the same output.
Insulin-like peptide 3 (INSL3) is different. It's a constitutive product of mature Leydig cells — secreted steadily, not pulsed like testosterone, not driven by acute LH stimulation. INSL3 reflects the number of functional Leydig cells and their differentiation status. It is a biomarker of capacity, not output.
This distinction matters because testosterone can be normal while the factory is damaged. INSL3 reveals what testosterone hides.
The Construction Timeline
The largest study to track Leydig cell maturation from adolescence into adulthood comes from the ALSPAC birth cohort — 1,781 young men measured at approximately 17 and again at 24 years. Tulumcu and colleagues published their findings in Frontiers in Endocrinology in 2025.
After approximately 22, the Leydig cell population is fixed. No further cell division. The factory is built. What you have is what you carry — through your 30s, 40s, 50s, and beyond — with only very limited attrition into old age.
What Shapes the Factory Before It's Finished
A companion paper from the same ALSPAC cohort, published in Andrology (Ivell et al. 2025), asked a question no study had asked before: what childhood factors determine adult Leydig cell capacity?
This timeline reframes the entire landscape. The adolescent period isn't just when puberty happens and ends — it's when the testosterone-producing infrastructure of the body is being physically assembled. Every exposure during this window acts on a system that's still under construction.
What INSL3 Predicts That Testosterone Doesn't
The European Male Ageing Study (EMAS) — 3,000+ men across 8 European centers, ages 40-79, measured twice over 4.3 years — tested whether INSL3 or testosterone better predicts health outcomes. Anand-Ivell and colleagues (2022) published the results.
| Morbidity Category | INSL3 Predicts? | Testosterone Predicts? |
|---|---|---|
| Cardiovascular disease | Yes | Yes |
| Diabetes | Yes | Yes |
| Bone weakness | Yes | No |
| Cancer | Yes | No |
| Sexual dysfunction | Yes | Yes |
| Respiratory disease | Yes | No |
| Musculoskeletal disorders | Yes | No |
| Gastrointestinal disorders | No | No |
| Neurological disorders | No | No |
| Score | 7 / 9 | 3 / 9 |
INSL3 predicted seven of nine morbidity categories. Testosterone predicted three. After full statistical adjustment, bone mineral density showed the most robust causal association with INSL3 — suggesting that the INSL3/RXFP2 signaling pathway is directly involved in bone metabolism, not merely a proxy for testosterone.
The Missing Variable in the Bone Paradox
In article #21, I described the bone paradox: TRT improves bone mineral density, yet fracture rates don't consistently fall. I proposed a four-layer resolution — compartment specificity, estradiol thresholds, site-specific fracture divergence, and population mismatch.
INSL3 adds a fifth layer that may be the most important.
Ferlin et al. (2008) found that RXFP2 gene mutations — the receptor for INSL3 — caused osteoporosis in 64% of men with normal testosterone. INSL3 stimulates osteoblast proliferation via cAMP in a dose- and time-dependent manner. This is a testosterone-independent mechanism for bone health.
If AAS use or adolescent obesity damages Leydig cells before maturation completes — reducing INSL3 permanently — then bone weakness could persist even when testosterone is fully replaced. TRT restores the output but not the factory. The INSL3-RXFP2 bone pathway stays broken.
This may explain why some men on TRT still fracture despite improved BMD.
The Steroid Damage Is Permanent Because the Organ Was Unfinished
In article #13, I described how INSL3 reveals Leydig cell damage that testosterone alone doesn't show. Rasmussen and colleagues (JCEM 2021) measured the aftermath:
Current users: 15× lower than controls. Former users, even after cessation: still 33% below healthy levels. LH recovered within ~12 months. Inhibin B within ~21 months. But neither testosterone nor INSL3 fully recovered. Longer accumulated AAS duration predicted lower INSL3.
Now add the developmental window. The average age of first AAS use has been declining — and no study has ever stratified INSL3 outcomes by age at first use. But the biology is clear: if you suppress a Leydig cell population that is still actively proliferating and differentiating, you're not just temporarily shutting it down. You're potentially reducing the total number of cells that will ever exist.
This is not a hypothesis that's been tested directly. It's a synthesis from two lines of evidence that no one has connected in a clinical study. That absence is itself a finding.
Six Articles, One Principle
The developmental window reframes much of what I've written:
INSL3 as Diagnostic Instrument
INSL3 can distinguish things testosterone cannot. In congenital hypogonadotropic hypogonadism (CHH), adults with CHH have INSL3 of 0.05 ng/ml versus controls at 1.08 ng/ml — a difference with 100% diagnostic accuracy. In adolescents, INSL3 distinguishes permanent CHH from constitutional delay of puberty (0.15 vs 0.35 ng/ml, 86.7% AUC), a clinical distinction that currently requires years of watchful waiting.
Anand-Ivell (2024) proposed a clinical threshold: INSL3 below 0.4 ng/ml as a "good index for clinical hypogonadism" — one that measures the factory, not the output. Combined with testosterone, it would give clinicians two dimensions: how much hormone is being produced now, and how much capacity exists to produce it in the future.
Almost no one measures INSL3 in clinical practice. It remains a research biomarker.
The Evidence Voids
What I've described is a synthesis — a developmental framework connecting published data from independent research groups who weren't studying the same question. The framework generates specific predictions that no study has tested:
Five Studies That Should Exist But Don't
- Age-at-first-AAS-use vs. adult INSL3. The Rasmussen data shows accumulated duration matters. Does age at first use matter independently? The developmental window predicts it does.
- Adolescent SSRI exposure → adult INSL3. The OR 1.44 for delayed puberty exists. No one has followed these men to age 25+ and measured their Leydig cell capacity.
- Childhood obesity → INSL3 trajectory → adult disease. Ivell showed the cross-sectional association. A prospective cohort tracking INSL3 from childhood through adulthood would test whether early obesity permanently reduces capacity.
- INSL3 as a predictor of TRT non-response. If low INSL3 reflects reduced Leydig cell mass, and INSL3/RXFP2 bone effects are T-independent, then men with low INSL3 might benefit from TRT for testosterone but still suffer bone loss, fractures, and other INSL3-dependent outcomes.
- The generational factory decline. If childhood obesity rates drive the secular testosterone decline through Leydig cell capacity, then INSL3 population data across birth cohorts would show a declining trajectory independent of adult BMI.
The absence of these studies is not surprising. INSL3 measurement requires LC-MS/MS or specific immunoassays, most longitudinal cohorts didn't include it, and the developmental window concept only crystallized with the ALSPAC data in 2025. But the absence means the developmental framework I've described remains a synthesis, not a confirmed model. The data that would falsify it doesn't exist yet.
The Unfinished Organ
We treat the testicle as if it arrives fully formed at puberty's end — defined, historically, as around 16-18. The INSL3 data says otherwise. The Leydig cell population continues to differentiate until approximately 22, with the major variance in adult capacity determined by influences that act before 17.
This means the testicle, in functional terms, is an unfinished organ through most of what we consider young adulthood. Every exposure during this extended construction phase — anabolic steroids, obesity, opioids, SSRIs, infections, endocrine disruptors — acts on a system that hasn't finished building itself. The damage isn't a temporary suppression that recovers when the exposure stops. It's a permanent reduction in the number of cells that will ever exist.
The clinical implications are straightforward. Screen adolescents for hypogonadism before prescribing drugs that suppress the HPG axis. Consider Leydig cell capacity, not just testosterone levels, when evaluating adults with a history of adolescent exposures. Measure INSL3 when testosterone alone doesn't explain the clinical picture.
The research implications are harder but more important: build the longitudinal cohorts that connect childhood exposures to adult Leydig cell capacity. Test whether age-at-first-AAS-use predicts INSL3 outcome independently of duration. Follow adolescents on SSRIs into their 20s and measure what's left.
The factory doesn't run forever. It doesn't even finish building on the schedule we thought. And we've been sending exposures into it during construction without checking what they leave behind.