Regulation of endometrial receptivity by the highly expressed HOXA9, HOXA11 and HOXD10 HOX-class homeobox genes.

STUDY QUESTION: Are other HOX genes, in addition to HOXA10, involved in endometrial receptivity?

SUMMARY ANSWER: The highly expressed HOXA9, HOXA11 and HOXD10 genes also appear to be involved in endometrial receptivity.

WHAT IS KNOWN ALREADY: Within the HOX family of homeobox transcription factor genes are the leading candidates for the regulation of embryonic implantation. A crucial role of HOXA10 in endometrial receptivity has been well established.

STUDY DESIGN, SIZE, DURATION: To identify HOX candidate genes, we performed data mining on all 39 human HOX genes in the 'Human body index' gene expression database of normal human tissue. The temporal and spatial expression pattern of four highly expressed HOX genes in the human endometrium was determined. To further investigate the function of these Hox genes, we used a robust in vivo mouse model in which we blocked maternal Hox gene expression.

PARTICIPANTS/MATERIALS, SETTING AND METHODS: Analysis of a gene expression profile set in the public domain consisting of 504 samples representing 95 different normal human tissues, showed that in addition to HOXA10, also HOXA9, HOXA11, HOXB6 and HOXD10 mRNA showed increased expression in the human endometrium (16 samples). The temporal and spatial expression pattern of these four HOX genes throughout the menstrual cycle was determined in the endometrium from 27 female patients eligible for IVF-embryo transfer with a normal cycle by quantitative real-time PCR (qRT-PCR), western blot and immunohistochemistry. The role of maternal Hoxa9, Hoxa11 and Hoxd10 was assessed in a mouse implantation model by expression knockdown using RNA interference. Forty mice were transfected with Hoxa9-, Hoxa11- or Hoxd10-specific small hairpin RNA (shRNA) constructs or a vector control by injection into the uterine horn at Day 2 after vaginal plug detection (Day 1) (160 mice in total). The effects were examined by qRT-PCR and western blot at Day 4 and litter sizes counted at Day 9 of pregnancy.

MAIN RESULTS AND THE ROLE OF CHANCE: HOXA10, HOXA9, HOXA11 and HOXD10 all showed increased expression during the mid-secretory phase of the menstrual cycle (P < 0.01). Knockdown of Hoxa9, Hoxa11 and Hoxd10 in the murine uterus resulted in significantly reduced average implantation rates (P < 0.01) and, with regard to four Hox target genes, also correlated with a significantly increased empty spiracles homolog 2 (Emx2) and insulin-like growth factor binding protein-1 (Igfbp1), and decreased integrin β3 (Itgb3) and leukemia inhibitory factor (Lif), expression (P < 0.01).

LIMITATIONS, REASONS FOR CAUTION: Menstrual cycle stage was not confirmed by serum hormone analysis. We verified the absence of significant differences in stage-specific expression of the reference genes used in our study (ACTB/Actb and GAPDH/Gapdh) and therefore possible limitations of this approach were minimized. In addition, the translatability of our data from a mouse model to patients needs to be investigated further.

WIDER IMPLICATIONS OF THE FINDINGS: We provide evidence that three other HOX genes in addition to HOXA10 are involved in endometrial receptivity, and that part of their function is asserted through several known HOX target genes, suggesting the presence of a central HOX signal transduction pathway.