rs2235091 — KLK4

Most people think of tooth enamel as something that forms passively in the womb and during early childhood. What is less well known is that forming enamel passes through two fundamentally different phases — and the second phase, maturation, is where the bulk of mineral is deposited and where enamel either becomes hard or stays soft. The gene KLK4, encoding kallikrein-related peptidase 4¹¹ kallikrein-related peptidase 4
a serine protease secreted by transition and maturation-stage ameloblasts that aggressively degrades residual enamel matrix proteins to clear space for mineral expansion, controls this finishing step. rs2235091 is an intronic variant in KLK4 that has been linked to caries susceptibility and molar hypomineralization in multiple populations.

During the secretory stage of enamel formation, ameloblasts lay down a protein-rich organic matrix — predominantly amelogenin — that templates the growth of long, thin hydroxyapatite crystals. At the start of maturation, this matrix must be almost entirely removed so that the crystals can thicken, coalesce with adjacent crystals, and reach the high mineral density (~96% by weight) that gives mature enamel its extraordinary hardness. KLK4 is the enzyme primarily responsible for this protein clearance: it is secreted at high levels by maturation-stage ameloblasts and cleaves amelogenin and other enamel proteins at numerous sites, enabling their endocytic removal.

[Klk4 knockout mice | Hu et al., 2008, PubMed 19132006] develop enamel that retains a large amount of residual organic matrix: the crystals fail to fully expand, the enamel remains hypomineralized and pigmented (from retained matrix proteins), and it chips and abrades readily. In humans, homozygous loss-of-function mutations in KLK4 (e.g., the W153X nonsense mutation identified by Hart et al., 2004) cause autosomal recessive pigmented hypomaturation amelogenesis imperfecta — enamel of normal thickness that is radiographically soft and clinically fragile.

rs2235091 is an intronic variant whose precise molecular effect has not been characterized. Intronic variants can influence pre-mRNA splicing, alter intronic enhancer sequences, or affect RNA secondary structure and stability. Because KLK4 expression is tightly regulated in a narrow developmental window (transition and early maturation stage), even subtle reductions in protease expression or activity could impair matrix clearance and leave enamel marginally less mineralized than average — not dramatically enough to cause amelogenesis imperfecta, but enough to shift the threshold for acid-mediated demineralization.

The most comprehensive evidence comes from the Czech ELSPAC cohort Broukal et al., Clin Oral Investig, 2022²² Broukal et al., Clin Oral Investig, 2022
Polymorphisms in genes expressed during amelogenesis and their association with dental caries: a case-control study, which analysed 611 children with permanent dentition. In that cohort, the A allele of rs2235091 was significantly more common among caries-affected children than caries-free controls (any caries: OR 1.37, 95% CI 1.05–1.80, p=0.014; severe caries DMFT ≥ 6: OR 1.76, 95% CI 1.20–2.58, p=0.002). The AA genotype reached an OR of 4.15 (95% CI 1.54–11.23, p=0.006) compared to GG homozygotes. Haplotype analysis incorporating four KLK4 SNPs showed the GAGA haplotype (containing the rs2235091 A allele) strongly pro-carious (p=0.001 for DMFT > 0; p < 0.001 for DMFT ≥ 6), while the GAGG haplotype (protective G at rs2235091) was protective (p=0.003). Notably, no significant effect was detected in primary dentition (150 children), suggesting the association is specific to permanent enamel maturation.

The Iowa Fluoride Study³³ Iowa Fluoride Study
Wang et al., 2012; family-based association test in 333 Caucasian parent-child trios found the G allele protective against caries across pit/fissure surfaces (p=0.004) and smooth surfaces (p=0.02), consistent with the Czech findings on direction of effect. A Polish case-control study of 96 children (aged 20–42 months) identified rs2235091 as one of five SNPs significantly associated with caries incidence (p=0.0085). Most recently, a Brazilian preliminary study (118 children) found that the A allele of rs2235091 was associated with molar hypomineralization (MH) with an OR of 3.75 (95% CI 1.65–7.81, p=0.001), with the association amplified when combined with childhood antibiotic exposure — pointing toward a gene-environment interaction in the maturation window.

The overall picture is moderate evidence: three independent populations show the A allele conferring risk or the G allele conferring protection, with effect sizes in the range of OR 1.4–1.8 for any caries and up to OR 4.15 for the AA genotype in severe caries. The variant does not appear in ClinVar as clinically significant, and no functional characterization of its molecular effect is yet published. The evidence is primarily observational in pediatric European and Brazilian populations.

Because rs2235091 is intronic and its functional effect is not yet established, we cannot prescribe a specific molecular intervention. The actionable implication is that carriers of the A allele — particularly AA homozygotes — have a modestly elevated baseline risk for caries in permanent dentition and molar hypomineralization. This shifts the cost-benefit calculus for preventive dentistry in favor of more intensive protocols: higher-strength fluoride, remineralizing agents, and increased check-up frequency are the evidence-backed tools for people with genetically suboptimal enamel maturation.

Calcium and phosphate adequacy during the enamel maturation window (roughly birth to age 12) is an important cofactor — even normal KLK4-mediated protein clearance requires sufficient mineral substrate to fill the cleared space. The combination of reduced KLK4 efficiency and low dietary calcium/phosphate during childhood is likely worse than either alone.

rs2235091 has been studied as part of KLK4 haplotype blocks alongside rs198968, rs2242670, and rs2978642. The GAGA haplotype incorporating all four SNPs shows stronger association than any single SNP, supporting a polygenic model within the KLK4 locus. Additionally, AMELX rs17878486 and KLK4 rs2235091 appear to act as part of a broader enamel-gene cluster (also including MMP20 and MMP13) with joint association with caries risk (p < 10⁻⁵ in gene-cluster analyses). The AMELX–KLK4 interaction is biologically coherent: AMELX shapes the secretory matrix, and KLK4 clears it — defects in either step can impair final enamel quality. A compound action for AA+TT (rs2235091 AA + rs17878486 TT) is a reasonable candidate for future research once both variants are fully characterized.