Id4 deficiency attenuates prostate development and promotes PIN-like lesions by regulating androgen receptor activity and expression of NKX3.1 and PTEN

Androgen receptor is the key regulator of prostate development including size, branching morphogenesis and differentiation. Quantitation of androgen receptor positive cells (brown nuclei, Figure 3A and B, n = 25 tubules each) followed by statistical analysis revealed that loss of Id4 had no apparent effect on androgen receptor expression (black arrowheads, Figure 3A) as compared to wild type littermates (black arrowheads, Figure 3B and inset) in the glandular epithelium of the prostate. Similar to wild type, AR expression was also present in the stromal cells in Id4-/- prostates (yellow arrowheads, Figure 3A and B). AR was also predominantly nuclear suggesting efficient nuclear translocation in Id4-/- following ligand binding. Thus androgen receptor pathway which is essential to support normal sex differentiation, development of male genital tract and organ development appears to be intact. These results also suggested that Id4 is required to maintain normal prostate development through genetic events downstream of androgen receptor and deficiency of Id4 may attenuate these pathways leading to decreased prostatic secretions and PIN like lesions.

We next investigated the expression of Nkx3.1, a key androgen receptor downstream target. The expression of homeobox gene Nkx3.1 in prostate epithelial cells is rapidly lost after castration, but is quickly restored after androgen dependent prostate regeneration [38]. Nuclear Nkx3.1 expression was clearly observed in prostates from WT mice suggesting a normal prostate developmental program and androgen response (Figure 3C and D). In contrast, Nkx3.1 expression was noticeably absent in the Id4-/- mice (Figure 3E and F and inset, Nkx3.1 positive nuclei counted in n = 25 tubules). Nkx3.1 is also the earliest known marker of prostate development and is a critical regulator of prostate epithelial differentiation in mouse models [39]. Loss of Nkx3.1 leads to significant decreases in prostatic ductal branching and production of secretory proteins [39]. Nkx3.1 knockout mice also frequently display prostate epithelial hyperplasia and dysplasia and often develop PIN [39]. Some of these phenotypes such as reduced ductal branching and diameter (Figure 1D) and PIN like lesions (Figure 2M, O, P, Q and R) were also present in Id4-/- prostates, perhaps due to loss of Nkx3.1 expression.

Androgen-dependent transcription of the mouse Nkx3.1 is conferred through a non-canonical androgen response element (ARE) element within an intron [40] (Figure 3G). Chromatin immuno-precipitation analysis using androgen receptor antibody revealed that AR binding is significantly reduced (P < 0.001) at this site in Id4-/- mice as compared to the levels observed in prostates from WT mice (Figure 3H). These results provided direct evidence that decreased Nkx3.1 expression is not due to loss of androgen receptor (Figure 3A and B) but due to attenuated androgen receptor binding to its cognate response element.

Based on in vitro and in vivo studies, PTEN and its downstream signaling pathways have emerged as major regulators of NKX3.1 expression [41]. As expected, Pten was highly expressed in the wild type prostate epithelium and stroma (Figure 4A). The immuno-histochemical studies shown in Figure 4B and C clearly demonstrated a significant decrease in Pten expression in Id4-/- prostate epithelial cells (black arrowheads). Surprisingly, Pten expression was maintained in non-prostatic tissue such as urethra (Figure 4C, asterisk) in Id4-/- mice suggesting that the decreased Pten expression was specific to prostate. Lack of Id4 expression in the urethra (data not shown) further suggests that Pten expression is influenced by Id4 specifically in the prostate. Since Pten regulates Nkx3.1 expression, the loss of prostatic Pten might be an alternate mechanism by which Nkx3.1 is down-regulated in the Id4-/- prostate [42]. Furthermore, these mechanisms may be independent of AR-regulated Nkx3.1 gene transcription mechanism. The Id4-/- knockout model thus closely mimics the Pten:Nkx3.1 mutant mice [43].

Pten, a phosphatase is involved in the regulation of Akt phosphorylation. We measured the expression of phospho-Akt (p-Akt1, 2 and 3) as readout of Pten expression/activity in Id4-/- mice. High p-Akt activity (nuclear and cytoplasmic) in the dorsal prostate of Id4-/- mice was consistent with decreased Pten expression (Figure 4D). Unexpectedly, low to negligible p-Akt activity was observed in the ventral (Figure 4E) and lateral prostates (Figure 4F) suggesting a lobe specific effect. We reasoned that decreased p-Akt even in the absence of Pten could be due to reduced expression of total Akt. Surprisingly, total Akt expression was undetectable in lateral and ventral prostate (Figure 4G) but was present in dorsal prostate (Figure 4H and I). These results suggested that loss of p-AKt observed in lateral and ventral prostate was likely due to decreased expression of total Akt and not due to loss of Pten. High Akt expression was observed in the wild type prostate but the expression pattern was unanticipated. Akt expression in the glandular epithelium was not uniform but highly localized to few cells (Figure 4J-L) suggesting that Akt expression is not constitutive. The expression of p-Akt was consistent with regions expressing high and low Akt. We next counted p-Akt positive cells in tubules that also stained positive for Akt (see panels D, H and I in Figure 4). A significant (P < 0.001) increase in the fraction of p-Akt positive cells in this analysis (inset in Figure 4L) further supports the lack of Pten in Id4-/- prostates as compared to Id4+/+ prostate. Additional studies will be required to demonstrate the functional significance of the localized Akt expression and its effect on cell function, for example whether Akt expression correlates with pAKt, Pten and downstream effectors of p-Akt in specific Akt positive and negative cell types in a lobe specific manner.

The presence of hyperplastic regions (Figure 2O-R) was associated with increased expression of proliferative marker Ki67 in Id4-/- in prostate ducts (Figure 5A). Marked increase in Ki67 was also observed in growing prostatic projections in the lumen in Id4-/- prostates (Box in Figure 5A). In contrast, Ki67 positive nuclei in Id4+/+ littermates were observed in only few cells per tubule (Figure 5B, P < 0.01, Ki67 positive nuclei counted in n = 25 tubules).

Increased Ki67 was also associated with increased expression of regulators of proliferation such as Myc [44] and Id1 [45]. Myc positive nuclei were more frequently observed in glandular epithelial cells in Id4-/- (Figure 5C) as compared to Id4+/+ prostates (Figure 5D, P < 0.001, Myc positive nuclei counted in n = 25 tubules). Recent studies have also shown an inverse relationship between Myc and Nkx3.1 [46, 47].

Id1, a member of the HLH family of transcription factors was undetectable in the Id4+/+ (Figure 5F) but increased significantly in Id4-/- mice (Figure 5E, P < 0.001, Id1 positive nuclei counted in n = 25 tubules). Id1 promotes cell cycle progression by down-regulating multiple CDKNIs including p21 and p16(Ink4a) [48, 49]. Together with Myc, increased Id1 expression is also associated with increasing grade of prostate cancer [33, 34, 50].

In Pten and Nkx3.1 mutant mice, cells with increased levels of SOX9 are persistently present within prostate epithelia [51]. Immuno-histochemical analysis using Sox9 antibody revealed few Sox9 positive luminal epithelial cells in the wild type prostates (Figure 5H) [52]. In contrast, the epithelial cells from the Id4-/- prostate showed significantly higher Sox9 expression (Figure 5G, P < 0.001, Sox9 positive nuclei counted in n = 25 tubules). Increased Sox9 expression is observed at early stages of prostate hyperplasia and is associated with progression to high grade PIN lesions [53]. Sox9 is part of the prostate developmental pathway that is reactivated in prostate neoplasia where it promotes tumor cell proliferation and correlates with Ki67 expression [51].

The average number of TUNEL positive cells (from 5 different fields) in Id4-/- mice prostate (Figure 5I and inset) was not significantly different from WT mice (Figure 5J and inset, and graph representing TUNEL positive nuclei counted in n = 25 tubules). We speculate that even a small increase in proliferation, without noticeable change in apoptosis could have a dramatic effect on cellular growth.

The molecular changes in the prostate observed in the Id4-/- mouse model were further confirmed in vitro using Id4 gene silencing and gain-of-function approaches in the prostate cancer cell lines LNCaP and DU145. Id4 was silenced in LNCaP cells using gene specific siRNA (Figure 6A) and over-expressed in DU145 cells as previously described [29, 35]. Similar to the Id4-/- studies as described above (Figure 3D and F), silencing of Id4 in LNCaP cells resulted in decreased NKX3.1 expression, whereas ectopic Id4 expression in DU145 increased NKX3.1 expression (Figure 6B). Consistent with lack of androgen receptor binding on NKX3.1 promoter in Id4-/- mice prostate (Figure 3H), a significant decrease (P < 0.001) in androgen receptor binding on consensus ARE in NKX3.1 promoter (-3013 bp relative to transcriptional start site) was observed in LNCaP-Id4 cells as compared LNCaP cells (Figure 6C). These results clearly demonstrated that NKX3.1 is dependent on Id4. Loss of Id4 in LNCaP cells also resulted in increased Sox9 in these cells whereas Sox9 was undetectable in DU145 + Id4 cells (Figure 6B). Due to frame shift mutation, PTEN protein expression is not observed in LNCaP cells (Figure 6D) [54]. However, PTEN expression was higher in DU145 + Id4 cells as compared to DU145 cells alone (Figure 6D). These results not only confirmed the molecular changes observed in our in vivo and in vitro models but strongly support the role of Id4 as a potential tumor suppressor that is required for normal prostate development also.