Spatio-temporal expression of HOX genes in human hindgut development



Background: Hox genes belong to a highly conserved subgroup of the homeobox gene superfamily. Studies of animal models have emphasized their role in defining the body plan by their coordinated expression along the body axis during ontogeny. Although an important role of HOX genes in human development is assumed, little is known about their expression during human ontogenesis. Therefore, we investigated the expression of the nine most posterior members of the HOXA, HOXB, HOXC, and HOXD clusters in embryonic hindgut between weeks 6 to 12 and in adult rectal tissue. Results: Applying in situ hybridization and immunohistochemistry, we observed expression of HOXA11, HOXA13, HOXD12, and HOXD13 in developmental week 6. However, expression of HOXD12 faded during weeks 7 and 8, and then became increasingly re-expressed during week 9 in humans. With the exception of HOXD13, all expressed HOX genes dropped below detection limits in week 11. Adult rectal tissue displayed distinct HOXA11, HOXA13, HOXD12, and HOXD13 expression patterns within the rectal layers. Conclusions: Our data suggest a strict spatio-temporal regulation of HOX gene expression during human development, supporting the idea of their role as key regulators. Nonetheless, the expression pattern of distinct HOX genes differs markedly from animal models. Developmental Dynamics 242:53–66, 2013. © 2012 Wiley Periodicals, Inc.


Ontogenic development is a complex process that needs precise control mechanisms. The regulation of well-timed events in restricted areas of the body is currently not fully understood. However, there is evidence suggesting that transcription factors are involved in the management of developmental processes. The family of hox genes contains a group of transcription factors and is known as one of the main players in the regulation of development. Hox genes are organized in clusters and display distinct spatial and temporal expression patterns during development (Hart et al., 1985; Harding et al., 1985; Duboule et al., 1986; Gehring, 1987). Hox genes are highly conserved over a broad spectrum of species and show similar characteristics of expression gradients along their body axes (Krumlauf, 1994).

Based on animal studies, three hypotheses for the regulation of hox gene expression have been proposed: (1) spatial, (2) temporal collinearity, and (3) posterior prevalence.

Spatial collinearity was first described in Drosophila (Lewis, 1978), demonstrating a direct association of distinct hox mRNAs to segments of the bithorax. Spatial collinearity was proposed by Dolle et al. (1989) for mouse limb development and in the gonads (Dolle and Duboule, 1989). Hox gene position on the chromosome follows a reciprocal sequence to the expression of their mRNAs along the body axes. Thus, genes located on the 5′ end of clusters code for mRNAs important for distal/caudal regions, while those at the 3′ end code for mRNAs involved in proximal/rostral regions (Duboule and Dolle, 1989; Galliot et al., 1989; Dolle et al., 1989, 1991a, b; Graham et al., 1989; Wilkinson et al., 1989; Gaunt, 1991; Hunt et al., 1991; Izpisua-Belmonte et al., 1991a; Kessel and Gruss, 1991; Hunt and Krumlauf, 1992). In line with data obtained from animal models, HOXB13 appears important in human prostate development (Dhanasekaran et al., 2005) while HOX paralogous groups 1–3 apparently play a role in foregut development (Yahagi et al., 2004).

The hypothesis of temporal collinearity proposes that Hox clusters are expressed from 5′ to 3′ in temporal sequence. Supporting evidence for a “hox-clock” (Duboule, 1994; Kmita and Duboule, 2003) was reported from mouse (Dolle et al., 1989), mouse and Drosophila (Izpisua-Belmonte et al., 1991b) and chicken (Dolle et al., 1991a).

The hypothesis of posterior prevalence is closely related to temporal collinearity as it also postulates a fetal dominance of posterior (i.e., those located 5′ on the chromosome) over anterior genes expressed in the same domain. Such a posterior dominance was proposed for Drosophila (Gonzalez-Reyes and Morata, 1990) and for mouse development (Duboule, 1991; Duboule and Morata, 1994).

The functional importance of hox genes during development also became evident through several loss-of-function mutations (Carpenter et al., 1993; Condie and Capecchi, 1993; Favier and Dolle, 1997; Kondo et al., 1998). Loss of Hoxd13 resulted in severe limb malformations in mice (Davis and Capecchi, 1996; Kmita et al., 2005). Hoxd12 and Hoxd13 were shown to be important for the development of the mouse anal sphincter (Kondo et al., 1996) and terminal gut (Mark et al., 1997), and Hoxa13 is involved in hindgut development in chicken (Roberts et al., 1995, 1998; de Santa Barbara et al., 2002). Dysfunction of Hoxd11 in mouse induced malformations of the forelimbs and os sacrum, and incomplete development of muscles of the iliocecal transition zone (Davis and Capecchi, 1994; Davis et al., 1995; Zakany et al., 1996, 1997). Deletion of Hoxd8 to Hoxd10 resulted in an entire loss of the “anterior region”, containing the proximal parts of fore- and hind-limb and the entire abdominal region of mice (Noordermeer et al., 2011). Knockout of the entire hox paralogous groups 3, 4 and 5 resulted in malformations of the neck, of groups 5–9 in malformations of the thorax, of group 10 of the lumbar column and of group 11 of the sacrum and tail (Mallo et al., 2010). In humans, mutations of HOXA13 and/or HOXD13 induce developmental anomalies in hands, feet and genitals (Goodman, 2002).

Although there are several studies in animal models, little information is available on the functional role of HOX genes in humans. Most prominent are the genetic association of the hand-foot-genital syndrome with HOXA13 (Mortlock et al., 1996; Goodman et al., 2000) and Guttmacher syndrome (Guttmacher, 1993; Innis et al., 2002) or brachydactyly and synpolydactyly with HOXD13 and the HOXD13 association with VACTERL (Akarsu et al., 1996; Goodman et al., 1997, 1998; Johnson et al., 1998; Kan et al., 2003; Zhao et al., 2007; Garcia-Barcelo et al., 2008). However, it is not known whether HOX gene expression follows the proposed spatial or temporal collinearity in human development. Due to ethical and legal limitations, studies on human development are restricted to a short time-window of developmental weeks 6 to 12, confining functional developmental studies to a few organs. During this time, the hindgut undergoes important developmental changes. Endodermal differentiation starts in week 6 and is completed in week 9, followed by the mesenchymal differentiation of the muscle layers in week 7 (Fritsch et al., 2007). In week 10 muscular layers further differentiate and are completed in week 12 when the muscles reach their final position (Fritsch et al., 2010). The regulation of these developmental steps is currently not fully understood; however, a functional involvement of HOX genes appears highly likely. Detailed knowledge of molecular processes (as summarized by Grapin-Botton, 2005, concerning animals) is necessary to answer open questions in human development (Lappin et al., 2006). Therefore, studying the expression of HOX mRNAs in distinct cell groups during distinct developmental stages will provide helpful information.

For a better understanding of the processes involved in the development of the human hindgut, we collected specimens covering the age of developmental week 6 to 12. We recently developed a reliable method to investigate the expression of mRNAs in newly collected as well as archival samples (Illig et al., 2009, 2010). Applying this method, we present data demonstrating a clearly restricted temporal and spatial expression pattern of nine HOX gene members of the posterior clusters in human hindgut development and in adulthood.


Spatio-temporal Alterations of HOX Gene and Protein Expression in the Developing Hindgut

HOXA11 and HOXA13 mRNAs were already expressed in the developing epithelial and mesenchymal rectal tissue in week 6 (the earliest stage investigated). This expression pattern persisted up to week 11 (Figs. 1, 2). Faint staining for HOXD12 and HOXD13 mRNAs was observed in rectal epithelial cells in week 6 (Figs. 1(9,10); 2(4,5)). HOXD13 mRNA expression extended to the surrounding mesenchyme between week 7 and 8 and remained up to week 12 (the latest time investigated; Figs. 1, 2). In contrast, HOXD12 mRNA levels dropped below detection limits during week 7 and 8 (Figs. 1(19,29); 2(9,14)). After its reappearance in week 9, marked labeling of epithelial as well as mesenchymal cells was evident in week 10 (Figs. 1(39); 2(29)). During week 11, HOXD12 mRNA appeared reduced in the epithelia but still markedly expressed in the mesenchyme (Figs. 1(59), 2(34)). After this time HOX gene expression was generally reduced and vanished with the exception of HOXD13 (Figs. 1(70), 2(35)). HOX protein expression determined by immunohistochemistry on neighboring sections revealed identical staining for HOXA11 (Fig. 1(1, 6, 11, 16, 21, 26, 31)) and HOXA13 (Fig. 2 insert in (18)), while HOXD12 and HOXD13 protein distribution was detectable only in the mesenchyme at the same stage (shown in week 9, Fig. 2 inserts in (19, 20)).

Figure 1.

Spatio-temporal alterations in HOX gene expression during hindgut development. Photomicrographs obtained from neighboring sections after in situ hybridization for HOXA11 (2,12,22,32,42,52,62), HOXA13 (3,13,23,33,43,53,63), HOXB13 (4,14,24,34,44,54,64), HOXC11 (5,15,25,35,45,55,65), HOXC12 (6,16,26,36,46,56,66), HOXC13 (7,17,27,37,47,57,67), HOXD11 (8,18,28,38,48,58,68), HOXD12 (9,19,29,39,49,59,69), and HOXD13 (10,20,30,40,50,60,70) mRNAs are depicted for developmental week 6 to 12 (as indicated on left border). The region of interest was identified by hematoxylin and eosin (HE 1,11,21,31,41,51,61) staining. m, mesenchyme; e, epithelium. Scale bar in (70) for 1 to 70 represents 100 μm.

Figure 2.

Spatio-temporal alterations in expression of the four most prominent HOX genes during hindgut development. Labeling for HOXA11 (2,7,12,17,22,27,32), HOXA13 (3,8,13,18,23,28,33), HOXD12 (4,9,14,19,24,29,34), and HOXD13 (5,10,15,20,25,30,35) mRNAs expression are depicted after in situ hybridization on neighboring sections for developmental week 6 to 12 (as indicated on left border). Immunohistochemical detection of the respective proteins is depicted for HOXA11 (1,6,11,16,21,26,31) from week 6 to 12; and for HOXA13 (insert in 18), HOXD12 (insert in 19) and HOXD13 (insert in 20) in week 9. The arrow-tip indicates the border between mesenchyme (left or above) and epithelium (right or below). Scale bar in (35) for (1-35) and in insert of (20) for (all inserts) represent 50 μm.

Expression of HOXB13, HOXC11, HOXC12, HOXC13, and HOXD11 mRNAs was detected neither in the epithelia nor in the surrounding mesenchyme at any developmental stage investigated (Fig. 1).

HOX Gene and Protein Expression in the Adult Rectum

To examine the spatial expression pattern in adults we dissected adult rectal specimens and took samples from the cranial, middle and caudal part (on the border of the anorectal zone) of the rectum, covering the mucosal, muscular, and mesorectal layers (Figs. 3, 4, 5, 3–5).

Figure 3.

Spatial HOX mRNA expression in adult rectal tissues. Photomicrographs obtained from neighboring sections after in situ hybridization for HOXA11 (2,12,22), HOXA13 (3,13,23), HOXB13 (4,14,24), HOXC11 (5,15,25), HOXC12 (6,16,26), HOXC13 (7,17,27), HOXD11 (8,18,28), HOXD12 (9,19,29), and HOXD13 (10,20,30) are depicted for the cranial, middle, and caudal part of adult rectum. Mucosal (MUC), muscular (MUS), and mesorectal (MES) layers (as indicated on left border) were identified by hematoxylin and eosin staining (HE; 1,11,21). Inserts in 3,12,23,29 and 30 display level dependent expression pattern of HOXA11, HOXA13, HOXD12, and HOXD13 mRNAs at higher magnification. Scale bar in (30) (for 1-30) represents 100 μm; scale bar in insert of (30) for all inserts represent 50 μm.

Figure 4.

Expression of the four most prominent HOX genes in adult rectum. Photomicrographs obtained from neighboring sections after in situ hybridization for HOXA11 (1,5,9,13,17,21,25,29,33), HOXA13 (2,6,10,14,18,22,26,30,34), HOXD12 (3,7,11,15,19,23,27,31,35), and HOXD13 (4,8,12,16,20,24,28,32,36) mRNAs are depicted for the cranial, middle, and caudal part of entire adult rectums covering mucosal (MUC), muscular (MUS), and mesorectal (MES) layers (as indicated on left border). Scale bar in (36) for all photomicrographs represents 50 μm.

Figure 5.

Layer-dependent protein expression of HOXA11, HOXA13, HOXD12, and HOXD13 on neighboring sections of the cranial part of the adult rectum. Photomicrographs depict HOXA11 (1,5,9), HOXA13 (2,6,10), HOXD12 (3,7,11), and HOXD13 (4,8,12) immunoreactivity from sections obtained after immunohistochemistry. Scale bar in (12) for all photomicrographs represents 50 μm.

In the middle and caudal position of the rectum, marked HOXA13 mRNA expression was observed in cells of all three layers of the mucosa, in the two layers of the muscularis and in the mesorectum (Figs. 3(3, 13, 23) and inserts in 3(3, 23); 4 second column). Interestingly, within the mucosa of the cranial part of the rectum, no HOXA13 mRNA expression was detectable (Fig. 4(2)).

HOXD13-labeled cells were observed in the muscularis of the cranial part (Fig. 4(8)) and in the mesorectum of the cranial and middle parts (Fig. 4(12,24)), but not in the caudal part (Fig. 4(36)). Of interest, an expression of HOXD13 mRNA in the mucosa was observed only in the caudal part of the rectum (compare Fig. 4(4, 16, 28)).

The caudal part of the rectum displayed an equal distribution of HOXA13- and HOXD13-labeled cells in the mucosa and of HOXA13 in the muscularis including the mesorectum, accompanied by HOXD12-positive cells within the epithelium (Fig. 4).

HOXA11 mRNA expression was restricted to some cells in the epithelium of the middle rectal part (Figs. 3(12) and insert therein, 4(13)). An analogous distribution of HOXA11, HOXA13, HOXD12, and HOXD13 protein expression on neighboring sections confirmed the labeling obtained by in situ hybridization (depicted for the cranial rectal part in Fig. 5(1–12)).

No expression of HOXB13, HOXC11, HOXC12, HOXC13, and HOXD11 mRNAs was detected in any layer of the entire rectum (Fig. 3).

Early Developmental Spatio-temporal Alterations in HOX Gene Expression in Other Body Regions

We detected HOXA11 mRNA in pancreatic exocrine acinar cells during weeks 9 and 10 (depicted for week 10 in Fig. 6(2)) and in the urogenital sinus from week 9 to 11 (Fig. 6(5)). HOXA13 mRNA was expressed in the interlobular duct system of the exocrine pancreas exclusively during week 8 (Fig. 6(1)) and in the urethra and urinary bladder from week 9 to 12 (Fig. 6(6)). In addition, HOXA13 mRNA was observed in the stomach and mesangial cells of the kidney in week 11 (data not shown). HOXD12 mRNA was observed in exocrine acinar cells including the loose surrounding connective tissue of the pancreas (Fig. 6(3)) and in the urinary bladder during week 10 (Fig. 6(7)). Moreover, HOXD12-positive cells were observed in the glomeruli of the kidney during week 11 (Fig. 6(8)).

Figure 6.

Spatio-temporal alterations of HOX mRNA expression in ontogenesis. Photomicrographs depict labeling for different HOX mRNAs in distinct cell populations outside the hindgut after in situ hybridization. Expression of HOXA13 (1), HOXA11 (2), and HOXD12 (3) was observed in pancreatic tissues while staining for HOXC11 (4) was negative on neighboring section. Within the urinary system labeling for HOXA11 in UGS = urogenital sinus (5), for HOXA13 and HOXD12 in urinary bladder (6,7); and HOXD12 in the kidney (8) was observed. HOXA13 and HOXD13 expression was observed in the vertebral column (9-12) and in developing limbs: HOXD13 in lower limbs (14); HOXA13 in upper limbs (15). HOXD12 expression was also detected in lower limbs (13,16). Positive staining for the remaining distal paralogues was revealed in the endothelia of blood vessels for HOXC11 (17) and within mesenchymal cells of connective tissue for HOXD11 mRNA (18). HOXB13 mRNA expression was restricted to more proximal organs like the esophagus (19). HOXC12 mRNA expression emerged distally in the skin of the plantar sole (20). Developmental stages of the tissues are indicated within the figure. Scale bar in (4) for (1–4; 9–12) and scale bar in (8) for (5–8; 13–16; and 17–20) represent 50 μm.

During the development of the vertebral column, HOXA13 mRNA expression was observed from week 8 to 10 (Fig. 6(9–11)) followed by HOXD13 in week 10 (Fig. 6(12)). Additionally, HOXD12 mRNA expression appeared in week 8 in the hipbone (Fig. 6(13)) and in the perichondral zone of the lower limbs in week 11 (Fig. 6(16)). At week 9, HOXD13-positive cells were detected in the lower limbs (Fig. 6(14)) whereas HOXA13 mRNA expression was restricted to the upper limbs one week later (Fig. 6(15)).

Despite the lack of expression of HOXB13, HOXC11, HOXC12, HOXC13, and HOXD11 mRNAs in the hindgut or other abdominopelvic organs or tissues (note the lack of labeling for HOXC11 mRNA in the pancreas in week 10, Fig. 6(4)), we occasionally observed positive labeling for most HOX mRNAs in other tissues, thereby providing evidence for the suitability of our probes. For example, HOXC11-positive cells appeared in week 9 in blood vessel walls (Fig. 6(17)), HOXD11-positive staining for fibrocytes in the surrounding connective tissue (Fig. 6(18)) and HOXB13 in the epithelial cells of the developing esophagus in week 10 (Fig. 6(19)) and the trachea (data not shown). Additionally, HOXC12 was observed in the plantar dermis at the beginning of week 11 (Fig. 6(20)) and HOXC13 mRNA in fibroblasts of the skin during week 9 (data not shown). For a summary of the observed expression of posterior HOX genes, refer to Figure 7A.

Figure 7.

A: Cumulative overview of spatio-temporal expression of HOX genes during ontogenesis: ++ marked expression; + moderate expression; - no detectable expression in the hindgut. Ectopic expression in other organs is indicated. B: Schematic summary of spatial HOX expression in sub-regions of adult rectal tissue.


Our data regarding the expression pattern of the most posterior members of the posterior HOX paralogues groups (HOXPGs) reveal selective temporal and spatial HOX gene expression profiles during embryonic development and adulthood in the human hindgut. Our data are in line with the proposed spatial and temporal collinearity as well as posterior prevalence models. However, beside several similarities, marked differences compared with animal models were assessed.

Comparability of Animal Models and Human Development

Our data strongly support the idea of a tight regulation of temporal and spatial expression of HOX genes during human development (Fig. 7A). We observed an orchestrated expression of several HOX members of the posterior cluster in distinct cell populations of the hindgut (Fig. 7A). HOXA11 and HOXA13 mRNA expression was already established in the epithelium and mesenchyme in the earliest specimens (week 6). The expression of HOXA11 and HOXA13 mRNAs remained stable until week 11. Marked Hoxa13 expression was also observed in rodent hindgut epithelia, but not in the mesenchyme (Warot et al., 1997; Zakany and Duboule, 1999a, b) and chicken hindgut (de Santa Barbara and Roberts, 2002) at comparable developmental stages. In contrast, the expression of HOXA11 in the hindgut appears to be a typical human feature, which also persists in adulthood (Figs. 3, 4). Despite restricted expression in the cecum of chicken (Yokouchi et al., 1995), Hoxa11 expression was primarily associated with skeletal and urogenital development (Small and Potter, 1993; Davis et al., 1995; Hsieh-Li et al., 1995; Zhao and Potter, 2001, 2002; Wellick et al., 2002; Wagner and Lynch, 2005; Wellick, 2011).

Faint staining for HOXD12 and HOXD13 mRNAs was observed in the hindgut during week 6. HOXD12 dropped below detection limits between week 7 and 8 and reappeared in week 9 in epithelial and mesenchymal cells, and subsequently persisted up to week 11. We observed expression in all five specimens derived from week 6, but in none derived from weeks 7 and 8 (in total nine specimens). From week 9–11, all samples yielded highly comparable results. So we can be certain that our observation indeed reflects alterations in mRNA levels. However, we cannot judge how close to the detection limit the levels are. Thus, the actual fluctuation in mRNA expression might be much less dramatic than it looks. While Izpisua-Belmonte et al. (1991b) observed no expression of Hoxd12 in the terminal part of the murine hindgut, studies by Dolle et al. (1991a) and Kondo et al. (1996) revealed its expression restricted to the surrounding mesenchyme. HOXD13 expression started to extend to the surrounding mesenchyme between week 7 and 8 and remained until week 12, which is in line with findings reported in animal models (Dolle et al., 1991a, b; Kondo et al., 1997; Roberts et al., 1998). In our study, we observed a widespread spatial and temporal coexistence of HOXA13 and HOXD13. This is in line with reports in mice, for which a Hoxa13-dependent expression of Hoxd13 was suggested (de Santa Barbara and Roberts, 2002). In contrast, studies in rats favored Hoxd13 expression before Hoxa13 (Mandhan et al., 2006). As we observed both mRNAs from the earliest stage investigated, we cannot address a potential inter-dependence in humans.

Potential Role of HOX Genes in Development

In our study, we found that HOXA13 and HOXD13, which appear to be highly important in a broad variety of species, are also widely expressed in humans (Fig. 7A). The lack-of-function mutations in HOXA13 were shown to be causal to the hand-foot-genital (OMIM 140000; Stern et al., 1970; Mortlock and Innis, 1997; Goodman et al., 2000) and Guttmacher (OMIM 176305; Guttmacher, 1993; Innis et al., 2002) syndromes in humans. Both syndromes also include severe developmental disturbances of the limbs and urogenital system. HOXD13-related developmental malformations mainly affect limb development, but were also correlated with hindgut malformation in a single case (Garcia-Barcelo et al., 2008). Studies in vertebrates suggest a normal Hoxd13 expression as a prerequisite for the correct development of structures in the most posterior regions along the primary and secondary body axis (van der Hoeven et al., 1996; Johnson and Tabin, 1997). In line with this, hetero- and homozygous deletion of Hoxd13 induces malformations in the muscular layers and internal sphincter of the mouse hindgut (Kondo et al., 1996; Mark et al., 1997; Warot et al., 1997).

HOXD12 mRNA displayed strong fluctuations in human hindgut development, which may correspond to the unequivocal data reported for Hoxd12 expression in mouse epithelial and mesenchymal layers (Izpisua-Belmonte et al., 1991b; Kondo et al., 1996). Kondo et al. (1996) and Zakany and Duboule (1999a) suggested that Hoxd12-induced expression of Hoxd13 is necessary for the development of sphincter muscles in mice. In humans, the expression of HOXD13 appears to be independent of HOXD12. In contrast, a high expression of HOXA11 and HOXA13 in epithelial and mesenchymal cells in week 6 precedes the appearance of HOXD12 and HOXD13 mRNAs, which were not observed in the mesenchyme before week 8.

Of interest, during development HOXA11 mRNA is prominently expressed in the human, but not in the murine hindgut. In contrast to the hindgut, a marked expression in the axial skeleton, limbs, and urogenital development was described in chicken, eutherian mammals (Burke et al., 1995; Davis et al., 1995; Nelson et al., 1996; Taylor et al., 1997; Zhao and Potter, 2002; Lynch et al., 2004) and therian mammals (Wagner and Lynch, 2005). In line with a study by Connell and colleagues (2009) who proposed a promoting role for HOXA11 in human fibroblast proliferation in uterosacral ligaments, the knockdown of Hoxa11 in mice resulted in an altered development of the uterosacral ligament and the uterus (Ma et al., 2012). Lack of function of HOXA11 was associated with radioulnar synostosis with amegakaryocytic thrombocytopenia in humans (OMIM 605432; Thompson and Nguyen, 2000); no association with hindgut developmental disorders have yet been reported.

In contrast to mouse development (Zeltser et al., 1996; Cillo et al., 2001) where spatial expression of Hoxb13 along the primary axis in hindgut, urogenital tract, and posterior extent of the spinal cord was reported, no HOXB13 mRNA was observed in human abdominopelvic organs. HOXB13 expression was detected in the epithelia of the developing trachea and esophagus in our study, and in proliferating fibroblasts of the fetal dermis by Stelnicki et al. (1998) and by Cazal and colleagues (2006) in human minor salivary glands. However, a recent study concerning germ-line mutations in HOXB13 and an associated higher prostate-cancer risk (Ewing et al., 2012) suggests a potential involvement of HOXB13 in developmental stages not investigated here.

HOXC11 expression was observed from fetal week 16 to birth in the human kidney, skeletal muscle and small intestine by Mitchelmore et al. (1998), but was not observed in the hindgut during developmental weeks 6–12 or adult tissue (present study). However, HOXC11 mRNA expression was observed in the epithelia of blood vessels in week 9. In mice, early Hoxc11 mRNA expression was found in the neural tube and later in presumptive regions of the urogenital tract, femur and fibula (Hostikka and Capecchi, 1998), and in the prevertebral column of the snake (Di-Poi et al., 2010).

HOXC12 and HOXC13 mRNAs were not observed in our abdominal samples. Instead, we observed HOXC12-positive cells in the plantar skin at the beginning of week 11, HOXC13-positive fibroblasts of the thoracic skin in week 9. Expression of these mRNAs also appeared highly restricted in animal models (Suemori and Noguchi, 2000). Experiments in mice showed that deletion of Hoxc13 results in loss of hair and sweat glands (Godwin and Capecchi, 1998), but did not result in alterations of internal organs.

Hoxd11 has been proposed to be expressed in the caudal segment of the lumbar spine modulating Hoxd10 expression during vertebral column development in chicken (Misra et al., 2009). In mice, temporal and spatial Hoxd11 expression was observed following the body axis during axial skeleton and limb development (Davis et al., 1995; Favier et al., 1995, 1996; Kondo et al., 1997; Zakany et al., 1997; Zakany and Duboule, 2007), and in the genital and distal limb buds (Spitz and Duboule, 2005). Both findings could not be confirmed for the developmental weeks 6–12 in humans or in adult rectal tissue in this study. In contrast, we observed some HOXD11-mRNA-expressing fibroblasts within the loose connective tissue of 9-week-old fetuses.

Potential Role(s) of HOX Genes in Adulthood

Beside the well-accepted functions of Hox genes during development with its established Hox code, the role of Hox genes in adulthood remains unclear. Recent studies emphasized that HOX genes might play an enduring role in maintaining positional identity and regional regenerative processes throughout the lifetime of an organism (Yahagi et al., 2004). Chang et al. (2002) as well as Shah and Sukumar (2010) suggested that one function of the ongoing HOX expression might be tissue homeostasis. A potential functional relationship between sustained HOX expression and positional identity of cells in adulthood was shown for human smooth muscle cells (Chi et al., 2007) and for skeletal muscle cells (Donoghue et al., 1992; Grieshammer et al., 1992). Wang and colleagues (2009) supported this theory by showing that ongoing HOXA13 expression in adult palmoplantar fibroblasts results in a positional memory of newly generated epidermal cells. HOX genes were also demonstrated to specify distal epidermal differentiation in adulthood (Rinn et al., 2007). In line with this, in our study, the same four posterior HOX genes, namely HOXA11, HOXA13, HOXD12, and HOXD13, which were expressed during hindgut development, remained until adulthood (Fig. 7B). It is noteworthy that the expression of HOX genes was suppressed in close vicinity to rectal cancer samples (Illig et al., 2009), thereby outlining the difference between CDX2-positive tumor from healthy HOX-gene-positive tissue by a thin layer of cells that were negative for both HOX and CDX2. In contrast, other studies in human rectal tissues reported the lack of HOXD13 expression in normal tissue but an increase in neoplastic rectal tissue (Cantile et al., 2009). In addition, divergent results were reported in two studies regarding HOXB13 expression in human colonic tissues. While Jung and colleagues (2005) detected HOXB13 expression in normal colorectal tissues and diminished or lost expression in tumors, observations by Edwards and colleagues (2005) yielded only a marginal expression of HOXB13 in normal colonic tissues but a convincing overexpression in colorectal adeno-carcinomas. Moreover, divergent Hoxb13 and HOXB13 expression patterns were also observed in prostate tissue (Economides and Capecchi, 2003; Takahashi et al., 2004; Edwards et al., 2005). However, further studies concerning these divergent findings will be necessary.

Methodical Considerations

mRNA detection in archival tissue is not an easy issue and might be prone to false negative results and consequently misleading comparisons to animal models. In this respect, it is important that we observed positively labeled cells for all mRNAs investigated, a fact suggesting that areas devoid of labeling are actually lacking detectable amounts of the respective mRNA. Moreover, cellular mRNA distribution patterns of HOXA11, HOXA13, HOXD12, and HOXD13 were reproduced by immunohistochemistry at the protein level, which is especially important to exclude potential false positive results.


Based on our data, we suggest an initial expression of HOXA11 and HOXA13, which precede the expression of HOXD13 and potentially HOXD12 during human hindgut development (Fig. 7A). It is noteworthy that the same HOX genes were expressed in adult rectal tissue (see Fig. 7A,B), which may support the proposed function of the ongoing HOX expression in tissue homeostasis (Chang et al., 2002).


Tissue Preparation and Histological Staining

Altogether, 1,371 sections obtained from 42 embryonic/fetal and nine adult specimens were included in this study. All samples belong to the collection of the Division of Clinical and Functional Anatomy (Department of Anatomy, Histology and Embryology, Innsbruck Medical University). Sampling was done in accordance with the Austrian laws including parental or patients' consent, respectively (McHanwell et al., 2008). All experiments were conducted in agreement with the guidelines of Austrian Ethic Laws and the principles of the Declaration of Helsinki.

All specimens were fixed in 4% paraformaldehyde (PFA) or formalin solutions (3–7%) in phosphate buffered saline (PBS) and paraffin-embedded in a routine histological infiltration processor (Miles Scientific Inc, Naperville, IL). Freshly collected samples were fixed for 4 hr; existing tissue samples taken from the collection were mostly stored longer in PFA. Paraffin-embedded tissues were cut sagittally except one specimen of week 9 that was cut coronary to 4 μm (ERGO Star Microm, Walldorf, Germany), and mounted on SuperFrost Plus microscope slides (Menzel, Braunschweig, Germany). Sections were dried overnight for storage. For histological stains, sections were dewaxed with xylene, rehydrated in a graded alcohol series, and every 10th section of a series was stained with hematoxylin and eosin (HE) applying standard histological protocols (Mulisch and Welsch, 2010).

In Situ Hybridization

In situ hybridization was carried out using an improved chromogenic detection protocol (Illig et al., 2009, 2010). Sections were dewaxed in xylene overnight at room temperature. After rehydration in a descending series of alcohol, they were refixed for 10 min in 2% PFA in PBS (0.9% NaCl; 10 mM Na-Phosphate buffer pH 7.4) at 4°C. Target retrieval was achieved by 10 min incubation in 2 M HCl at 30°C followed by proteinase K 50 μg/ml in TES (10 mM EDTA; 10 mM NaCl; 50 mM Tris-HCl, pH 8.2; 30 min at 37°C) treatment. Basic proteins were acetylated with 0.5% acetic anhydrite in TEA (0.1 M triethylamine HCl, pH 7.2). Sections were delipidated with chloroform. Prehybridization was carried out either with HybriBuffer ISH (Biognostik, Goettingen, Germany) or an optimized hybridization buffer (Illig et al., 2010) supplemented with Triton X-100 (0.4%) over 3 hr (50°C). Hybridization was performed overnight (18 hr) with hybridization buffer supplemented with the respective FAM-labeled single strand DNA probes (three to four per mRNA; Table 1) and Triton X-100 (0.4%) at 50°C for the first 15 min and reduced to 30°C afterward. Controls for specificity were performed by the addition of 1,000-fold excess of nonlabeled oligonucleotide probes or by not including FAM-labeled probes.

Table 1. Probes Used in This Study (End-Numbers Indicate Their Position on mRNA)

Incubation with a rabbit anti-fluorescein isothiocyanate antibody conjugated to alkaline phosphatase (1:200; Q 67848; Sigma-Aldrich, Vienna, Austria) was performed over night at 4°C. Sections were stained 2 × 20 min in NBT/BCIP staining solution (Roche, Mannheim, Germany). The staining reaction was terminated by 1 min washing in tap water and subsequent dehydration in a series of graded alcohols. After clearing the sections in xylene, slides were coverslipped with SPURR (Spurr, 1969) and incubated at 35°C for 48 hr for polymerization.


Immunohistochemistry applying HOXA11, HOXA13, HOXD12, and HOXD13 specific antibodies was performed on sections adjacent to those used for in situ hybridization. Pretreatment of the sections was done in accordance with the in situ method until the target retrieval step with HCl. For the detection of human HOXA11, HOXA13, HOXD12, and HOXD13, rabbit polyclonal antisera (HOXA11/ab30697; HOXA13/ab26084, HOXD13/ab19866, Abcam, Cambridge UK; HOXD12/ARP32647_T100, Aviva Systems Biology, CA) were used at dilutions of 1:50. Horseradish-peroxidase conjugated goat anti-rabbit (P 0448, DakoCytomation, Vienna, Austria) secondary antibody was applied at 1:400 dilutions. Sections were stained with Ni-DAB (Wouterlood et al., 1988) and coverslipped similar to the in situ protocol. Controls excluding the primary antibodies were included. All HOX antibodies used yield a single band on Western blots and the secondary antibody produced no labeling on sections incubated without primary antibody (data not shown).


We thank Hans Joachim Wolf, MD for constant support of our tissue collection and to Stefano Longato, MD and Dieter Urbas, MD for their assistance in preparing the figures. Medical University Innsbruck, Austria provided the funding for research conduct. Funding resources had no influence on study design, analysis, interpretation, and publication of obtained data. The authors declare that no conflicts of interest exist.