4 Discussion


4.1  Regulation of intestinal homeostasis by Brk

The human and mouse genomes encode 32 non-receptor protein tyrosine kinases (PTKs) that mediate a vast number of very diverse, highly specific and finely regulated molecular events in the cells. Moreover, only a fraction of these PTKs are typically expressed in an individual cell or a specific tissue. This signaling diversity is achieved not only by the differential specificity of non-receptor PTKs, but primarily through the multitude of the interactions of non-receptor PTKs with multiple non-kinase proteins that can modify the effects of these PTKs and regulate their functions in the cell.


Brk is a novel, non-receptor tyrosine kinase involved in various signal transduction pathways. Although Brk shares approximately 44% sequence identity with Src, it belongs to a distinct family of Src-like intracellular tyrosine kinases (Serfas and Tyner, 2003). Like Src family kinases and the majority of signal transducers, Brk performs its biological function through interaction with other cellular components. This depends on multiple functional domains – the SH3 and SH2 protein binding domains as well as its tyrosine residues (Qiu and Miller, 2004). Furthermore similar to Src, the SH3 and SH2 domains of BRK engage in intramolecular interactions with the kinase domain to form an autoinhibited conformation (Qiu and Miller, 2002). Disruption of the autoinhibited conformation leads to BRK kinase activation and freeing of the SH3 and SH2 domains to recruit substrates or other interacting proteins. In contrast to Src tyrosine kinases, Brk lacks an amino-terminal myristoylation signal, and although an apparent nuclear localization signal is absent, this kinase is flexible in its intracellular localization. Brk is one of a few tyrosine kinases found in the nucleus (Derry, et al., 2000;Haegebarth, et al., 2004), and like c-Abl (Zhu and Wang, 2004), the intracellular localization of Brk may influence its protein-protein interactions and the signaling pathways that it regulates.

Brk expression is epithelial-specific and developmentally regulated (Llor, et al., 1999;Vasioukhin, et al., 1995). Brk is expressed throughout the skin and the gastrointestinal tract with highest levels in neonatal colon and adult ileum. Brk mRNA and protein expression are restricted to differentiating intestinal epithelial cells, and it is excluded from proliferative cells of intestinal crypts, suggesting that Brk may play a role in the migration or terminal differentiation in these tissues (Vasioukhin, et al., 1995). Overexpression of Brk in mouse keratinocytes results in increased expression of the differentiation marker filaggrin during calcium-induced differentiation (Vasioukhin and Tyner, 1997), further supporting a role in differentiation for this kinase.

To begin to understand the biological role of Brk in the regulation of differentiation in vivo, mice carrying a disruption in the brk gene were generated. Since Brk is expressed at highest levels in the gastrointestinal tract it was chosen as the model system to characterize. The intestine is remarkable in that there is a high turnover of cells, arranged in order, undergoing proliferation in the crypt, and then differentiation to achieve the mature functions of absorptive enterocytes, mucus-producing goblet cells and enteroendocrine cells on the villus, or in case of the small intestine of Paneth cells at the base of the crypt (Sancho, et al., 2004).


Genetic ablation of brk in mice resulted in a range of low-penetrant phenotypes, including chronic inflammation (discussed in 4.2) and increased epithelial cell turnover in the gastrointestinal tract. Brk is not essential for embryonic development, as viable Brk null mice are obtained at the predicted Mendelian frequency in different genetic backgrounds. Homozygous deletions of the Brk family members Frk and Srms in mice have been generated as well (Chandrasekharan, et al., 2002, Kohmura, 1994 #2388), and similar to the Brk null mice, these homozygous mutants had no or only a mild phenotype. All mice were viable, fertile and not prone to developing spontaneous tumors. In contrast, homozygous deletion of Src42A in Drosophila caused lethality (Lu and Li, 1999). The lack of comparable effects for the individual mammalian Brk family tyrosine kinases suggests that they may share redundant functions. A high level of functional redundancy has been reported between Src family kinases. Evidence for compensatory changes in expression, activity, and subcellular localization of other Src family kinases has been reported in mice bearing single mutations, possibly accounting for the minimal phenotypes observed (Lowell, et al., 1996;Thomas, et al., 1995). Interestingly, the gastrointestinal tract expresses the highest endogenous levels of Brk and Frk (Sunitha and Avigan, 1996;Thuveson, et al., 1995;Vasioukhin, et al., 1995). It would be interesting to analyze effects of mutations in both Brk and Frk kinases on the homeostasis of the gastrointestinal tract. Clearly, further investigations are needed to test this hypothesis.

However, Brk signaling seems to be required for maintaining the balance between proliferation and differentiation in the intestinal epithelium. Homeostasis of the intestinal epithelium strongly depends on the balance existing among cell proliferation, cell cycle arrest, cell differentiation, and cell migration (Simon and Gordon, 1995). Loss of Brk in mice resulted in the phenotypic appearance of longer villi and increased proliferation measured by BrdU incorporation and PCNA staining. Proliferation in the intestinal epithelium is usually restricted to cells in the lower base of the crypt compartment. However, in Brk knockout mice proliferating cells were present in the middle and upper regions of the crypt, where cells are normally fully differentiated and non-proliferating. This expanded proliferative zone and increased number of proliferating cells in Brk deficient mice suggests a deregulated balance in proliferation and differentiation in the absence of Brk signaling.

Deregulation of crypt homeostasis and the formation of aberrant crypt foci (ACF) is a feature of neoplastic transformation, and is evident in the earliest stage of colon cancer (Renehan, et al., 2002). In addition, it is appreciated that the process of apoptosis is vital for normal crypt homeostasis and its impairment may be an early event in the neoplastic process. In this work I present evidence that in normal tissue, such as the intestinal epithelium, Brk is required to restrict cell proliferation and promotes apoptosis in mice, features it shares with tumor suppressor proteins. Tumor suppressors normally act as inhibitors of cell proliferation or activators of apoptosis and use a variety of mechanisms in tissue growth suppression (Macleod, 2000;Sherr, 2004). Brk expression is restricted to non-proliferating cells in the intestinal epithelium, suggesting a function in differentiation. Loss of this tyrosine kinase in mice resulted in increased epithelial cell turnover, pro-survival signaling and chronic inflammation, features which have all been correlated with the promotion of cancer development (Philip, et al., 2004;Sancho, et al., 2004). Preliminary experiments utilizing the AOM mouse model of colon carcinogenesis confirmed this hypothesis. Whereas very few aberrant crypt foci (ACF) were visible in wild-type animals after treatment with the AOM carcinogen, knockout mice exhibited a high incidence of aberrant crypt foci formation (ACFs) within the distal colon (Heap and Bie, unpublished). Aberrant crypt foci are considered the smallest identifiable lesions proposed to lead to colorectal carcinoma (Cheng and Lai, 2003), and their rapid appearance in Brk deficient mice strongly suggests a tumor suppressor function for Brk in the intestinal epithelium. However, due to the small number of animals used in these pilot experiments, further studies are currently underway.


Wnt signaling plays a major role in the regulation of intestinal cell proliferation and stem cell maintenance (Polakis, 2000;Sancho, et al., 2004), and its activation has been shown to result in hereditary and sporadic colon cancers (Bienz and Clevers, 2000). All intestinal malignancies share the characteristic of genomic instability and aberrant cell signaling pathways. One protein playing a crucial role in these diseases is the APC tumor suppressor protein. The APC protein is expressed throughout the intestine, where it functions as part of the Wnt pathway downregulating cytoplasmic levels of β-catenin (Hinoi, et al., 2000). Mutations in the Apc tumor suppressor gene initiate a majority of human colon cancers, and mice heterozygous for Apc mutations develop intestinal polyps. In APC-mutant mice, the wild-type allele is always lost in tumors. Inactivating mutations in APC have been placed at the beginning of the adenoma-carcinoma sequence in the development of colorectal cancer.

In search of signaling pathways being involved in the increased epithelial cell turnover in Brk deficient mice, the expression of β-catenin was analyzed. β-catenin in complex with members of the Lef/Tcf transcription factor family mediates a proliferation/ differentiation switch along the crypt-villus axis (Pinto, et al., 2003, van de Wetering, 2002 #3550) through activation/ repression of various target genes (Sancho, et al., 2004). These target genes play critical roles in the physiology of the intestine, as shown for the EphB2 and EphB3 in cell position in the intestine (Batlle, et al., 2002).

Loss of Brk resulted in nuclear accumulation of β-catenin not only in cells occupying basal positions of the crypt, as was observed in wild-type controls, but also in cells in higher positions of the crypt that usually lack nuclear β-catenin (Batlle, et al., 2002;van de Wetering, et al., 2002). Furthermore, this increased nuclear accumulation of β-catenin in the intestinal epithelium of Brk deficient mice resulted in increased expression of the β-catenin target gene c-myc in the intestinal crypts. c-Myc has been identified as one of the main mediators of the switch between proliferation and differentiation along the crypt-villus axis (He, et al., 1998). Thus, the upregulated expression of c-myc in the intestine of knockout mice is most likely the molecular mechanism causing the increased proliferation. Furthermore, the c-Myc protein is also playing a central role in the proliferative capacity of many cancers, including colorectal carcinoma (Grandori, et al., 2000). Thus, loss of Brk expression and subsequent signaling promotes activation of the Wnt pathway, implying that the activation of Wnt signaling is responsible for the observed intestinal phenotype in knockout mice.


But by which mechanism does the Brk tyrosine kinase negatively regulate the expression of β-catenin? Interestingly, the loss of Brk resulted in activation of Akt in the intestine. Brk deficient mice exhibited significantly upregulated levels of Akt kinase activity when compared to wild-type controls. Supporting these data, analysis by Carol A. Lange and colleagues showed that Brk associates with Akt resulting in inhibition of Akt activity (Zhang, et al., 2004). Association of Akt with Brk results in tyrosine phosphorylation of Akt, which in turn has been shown to result in reduction of Ser473 phosphorylation and inhibition of Akt kinase activity (Conus, et al., 2002). Akt, in turn, has been shown to facilitate the stabilization and nuclear accumulation of β-catenin either indirectly through inhibition of GSK-3β or through direct phosphorylation of β-catenin (Fukumoto, et al., 2001;Persad, et al., 2001;Tian, et al., 2004). These data would leave us with the following model (Fig. 20): in normal intestinal epithelial cells Brk might be positioned to inhibit Akt kinase activity, resulting in the stabilization of GSK-3β, which in turn phosphorylates β-catenin contributing to the degradation of β-catenin. Loss of Brk on the other hand would result in increased Akt kinase activity and increased nuclear accumulation of β-catenin, thus contributing to increased epithelial cell turnover. However, it remains to be addressed as to how exactly Brk interacts with Akt. Furthermore, it has to be investigated whether Brk actually colocalizes with Akt in epithelial cells of the intestine.

Fig. 20: Illustration of a putative regulatory role of Brk tyrosine kinase signaling. Brk represses Akt kinase activity, which is required to fully activate β-catenin. Thus, Brk inhibits nuclear-accumulation of β-catenin and transcription of β-catenin target genes in differentiated cells. Wnt signaling is also required but not sufficient to fully activate β-catenin.

4.2 Brk signaling and inflammation

The intestinal lining contains more lymphoid cells and produces more antibodies than any other organ in the body, and it has been proposed that approximately 25% of the intestinal mucosa consists of lymphoid tissue (Neutra, 1998). Intestinal lymphoid tissue is present both diffusively as intraepithelial lymphocytes, scattered lymphocytes and plasma cells in the lamina propria, and in organized mucosa-associated lymphoid tissue consisting of single or multiple lymphoid follicles known as Peyer’s patches. Intense immunological interactions occur between these immune cells and intestinal epithelial cells, which are known to interact with mesenchymal cells, such as myofibroblasts, and immunologically active macrophages, dendritic cells and lymphocytes, both intraepithelial and in the lamina propria (Berin, et al., 1999).


The regulation of proliferation of the stem cells in intestinal crypts is critical for maintenance of the structure of the mucosa and the capacity for effective absorption and defensive barrier functions. Although the intestinal epithelium is only one cell thick, it is sealed by tight and adherens junctions, which exclude peptides and antigenic macromolecules and mediate intercellular adhesion and “contact inhibition”. Brk wild-type and knockout mice exhibited no differences in expression of the tight junction marker ZO-1 and the adherens type junction marker β-catenin (membrane-bound) in the intestinal villus epithelium. However, loss of Brk expression resulted in the increased presence of Peyer’s patches in the gastrointestinal tract of outbred knockout mice compared with wild-type controls. Peyer’s patches are the sites for transport, processing and presenting of foreign antigens, which is a main part of the effective immune surveillance of the mucosal surface. Their high number suggests increased immune activity in the intestine of Brk knockout mice.

Furthermore, intestinal epithelial cells have been shown to act not merely as a passive barrier but also as sensitive indicators of infection that initiate defense responses. Noninvasive as well as invasive organisms have been demonstrated to elicit production of chemoattractants, suggesting that receptor-mediated signaling pathways may be involved. Epithelial cells can produce cytokines and chemokines that attract and activate immune cells with potentially important effects on the immediate and long-term host defense functions. For example, Paneth cells secrete potent antimicrobial molecules at the crypt base (Quellette, et al., 1989).

In addition to the observed increase in lymphoid tissue, loss of Brk expression resulted in the upregulated expression of the cytokines IL-6 and IL-18 in outbred mice (2.8). However, no differences in baseline cytokine expression were observed in inbred Brk deficient mice (2.8 G12). It has been reported that host genetic determinants play a major role in the sensitivity to inflammation and the following tumor risk (Rogers and Fox, 2004). In general, genetically engineered mice on a 129/Sv strain background appear to be especially susceptible to IBD-like disease. On the other hand, mice on a C57BL background are sometimes resistant to inflammation compared with equivalent mutant mice on other genetic backgrounds.


Cytokines are soluble mediators of cell-to-cell communication and their upregulation has been correlated to inflammation and cancer. Both, IL-6 and IL-18 have been shown to be involved in the immune response and are considered pro-inflammatory cytokines. IL-6 plays a major role in the acute phase response, B cell maturation and macrophage differentiation (Diehl and Rincon, 2002). It furthermore counteracts apoptosis and can promote cancer development. Increased IL-6 production has been shown to correlate with colon tumor formation and growth (Becker, et al., 2004). IL-18 is a novel cytokine produced by various cells including intestinal epithelial cells. However, within the gut mucosa IL-18 is primarily produced by intestinal epithelial cells (Pizarro, et al., 1999). It has been shown to induce interferon-γ (IFN-γ), and to play a crucial role in proliferation and maintenance of the intestinal epithelium (Okazawa, et al., 2004). Increased levels of IL-18 have been found in intestinal mucosal biopsies from inflammatory bowel disease (IBD) patients (Monteleone, et al., 1999;Pizarro, et al., 1999).

Inflammatory bowel disease (IBD) denotes chronic inflammatory disorders of the gastrointestinal tract of unknown etiology that comprises 2 major groups: ulcerative colitis (UC) and Crohn's disease (CD). Deregulation of the intestinal immune system, both at the humoral and cellular level, constitutes an important element in the multifactorial pathogenesis of IBD (Rogler and Andus, 1998). The expression of pro-inflammatory cytokines, most notably IL-1, IL-6, IL-18, TNF-alpha, and immunoregulatory cytokines, such as IL-2 and IFN-γ, in intestinal mucosa from IBD patients is markedly enhanced; however, it is not always accompanied by increases in cytokines' serum levels. Patients with inflammatory bowel disease (IBD) have an increased risk of developing colorectal cancer (Langholz, et al., 1992). The release of pro-inflammatory cytokines by enterocytes has been shown to be a major part in the immune response, ensuring that cells are not only called into epithelial/mucosal sites from the circulation but are activated and sustained once they reach this site. For example, cytokines are important activators of local macrophages, which also participate in host protection against infection. Furthermore, these cytokines can themselves induce epithelial cells to produce chemokines so that the initial stimulus by a microorganism can have an important multiplier effect. Thus, the observed upregulation of cytokine production in Brk deficient mice indicates an increased immune activity and suggests the occurrence of inflammation in their intestinal epithelium. Up-regulation of both IL-6 and IL-18 expression in Brk deficient mice supports the hypothesis of increased inflammation in knockout mice as indicated by the large amount of Peyer’s patches.

Growth factors and cytokines, produced either by epithelial or immune cells, can positively or negatively regulate the replication of the epithelium. These interactions frequently involve members of the cytokine superfamily, and the signaling from these receptors is in turn linked to tyrosine kinases expressed in the epithelium. Altered local cytokine production has been shown to be critical for inducing increased rates of epithelial cell turnover in inflammation (Podolsky, 2002). One could propose that the induced cytokine production through loss of Brk tyrosine kinase signaling in knockout mice further contributes to the observed phenotype of increased epithelial cell turnover. It furthermore has long been appreciated that small intestinal injury, for example in celiac disease, results in hypertrophy of the crypt compartment, but the mechanisms responsible for regulation of proliferation in the crypts have not been defined. However, it has also been proposed that inflammatory cytokines have anti-proliferative effects on epithelial cells (Booth and Potten, 2001;Neta and Okunieff, 1996).


The observation of increased epithelial cell turnover, high amount of lymphoid tissue and upregulated cytokine production in the intestine of mice deficient for Brk strongly suggests a role for this tyrosine kinase in maintaining intestinal homeostasis including a functional host defense barrier. Inflammation is the physiological response to injury caused by wounding, chemical irritation/damage, or infection. Whereas acute inflammation has been shown to lead to tumor regression and counteract cancer development, chronic inflammation leads to cellular proliferation thus strongly promoting cancer development (Philip, et al., 2004). Chronic inflammation has been shown to promote carcinogenesis indirectly in a non-cell-autonomous fashion. Upregulation of inflammatory mediators creates a tumor-promoting environment in which transformed epithelial cells thrive more optimally (Oshima, et al., 1996). Recently a more direct model has been proposed in which inflammatory signaling in the epithelial cells results in their inappropriate survival and transformation (Rakoff-Nahoum, et al., 2004, Greten, 2004 #3568). Thus, one could conclude that the apparent chronic inflammation in knockout mice predisposes these mice to cancer.

In addition to the reported baseline defects of immune response in Brk deficient intestinal epithelium, ablation of Brk signaling in the intestinal epithelium resulted in increased susceptibility to DSS, while the production of many inflammatory mediators was actually enhanced. Dextran sulfate sodium (DSS) has been shown to induce colitis by chemical injury in the intestinal epithelium accompanied by mild inflammation (Cooper, et al., 1993;Okayasu, et al., 1990). After subjecting wild-type and Brk knockout mice to this irritant, the mutant epithelium appeared more susceptible to DSS-induced histological damage than control epithelium. DSS treatment caused severe and acute inflammation accompanied with severe loss of epithelium in Brk deficient mice. Brk mutant mice showed severe colonic erosions, loss of crypts and an extent amount of inflammatory infiltrate in their submucosa and lamina propria following DSS treatment. In addition to the histological differences, Brk deficient mice exhibited an elevated expression of pro-inflammatory cytokines, such as IL-1 and IL-6, suggesting the occurrence of acute inflammation in these tissues. Acute inflammation is a self-limiting process that starts a cascade of cytokines and chemokines that attract immune and non-immune cells, mainly neutrophils, to infiltrate disrupted and damaged tissue. However, inflammation, even though to a much lesser degree, is also observed in wild-type animals. Furthermore, loss of IL-18 expression was observed in mutant mice. IL-18 is mainly produced by epithelial cells in the intestinal mucosa and its loss further confirmed the overall loss of intestinal epithelium in Brk deficient mice after treatment. Thus, Brk signaling seems to be required for epithelial cell integrity and loss of Brk signaling results in increased epithelial susceptibility to injury.

Interestingly, Brk protein expression was upregulated in the intestinal epithelium after DSS treatment. Nuclear Brk protein was present not only in differentiating cells but also in the proliferating cells of the crypt epithelium. This induction of Brk protein strongly suggests a role for Brk tyrosine kinase signaling in the protection against DSS-induced injury. Brk may protect the intestinal epithelium against injury by ensuring epithelial homeostasis. Consistent with this notion is the increased proliferation of intestinal epithelial cells in Brk deficient mice. This by itself would make them more susceptible to damage (Booth and Potten, 2001;Neta and Okunieff, 1996). The exact molecular mechanism of this effect remains to be detailed. Furthermore, upregulation of Brk protein expression might influence the steady-state production of protective factors, such as COX-2 or TGF-β1, which have been shown to be crucial in protecting the gut from injury (Dignass, 2001;Podolsky, 1999). However, it needs to be delineated, whether upregulation of nuclear Brk upon epithelial damage may actually induce the expression and production of cytoprotective factors. It has long been appreciated that mesenchymal-epithelial crosstalk and following cell signaling is crucial to the orchestration of responses to tissue injury (Clark, 2003).


Of special interest is the strong nuclear expression of Brk throughout the intestinal epithelium in response to DSS. Due to the lack of an amino-terminal myristoylation signal, the intracellular localization of Brk is flexible and it is one of a few tyrosine kinases found in the nucleus (Derry, et al., 2000;Haegebarth, et al., 2004). Nuclear localization of tyrosine kinases has been shown to be important for connecting extracellular signaling with the direct regulation of gene expression. Signal tranduction to the nucleus and the subsequent regulation of gene expression is of main importance for cells to react to external stimuli (Brivanlou and Darnell, 2002). Furthermore, tyrosine phosphorylation of nuclear proteins regulates many cellular processes, including growth, differentiation, and apoptosis (Cans, et al., 2000;Wang, 2000). Thus, Brk’s nuclear localization provides a means to direct communication between the cytoplasm and the nucleus, and even provides a platform for the integration of external signals. It is of importance to identify how this tyrosine kinase shuttles between cellular compartments and if its nuclear localization is necessary in response to growth factors. Furthermore, nuclear localization of PTKs, which possess various protein-protein binding domains, can also support the transport of other molecules to the nucleus. A similar function has been proposed for ErbB-1 in transporting STAT-1, a tyrosine-phosphorylated transcription factor, to the nucleus (Bild, et al., 2002). However, evidence supporting this notion for the Brk non-receptor tyrosine kinase needs to be found.

Non-receptor tyrosine kinases, such as Frk, have been found in the nucleus, but relevant substrates are still unknown or poorly understood (Cance, et al., 1994, Derry, 2000 #2416;Haegebarth, et al., 2004). One of the first Brk substrates identified was the RNA-binding protein Sam68 (Derry, et al., 2000). Brk has been shown to phosphorylate Sam68 in the nucleus resulting in the inhibition of its RNA binding and transport functions. Several studies support roles for Sam68 in the regulation of RNA metabolism and utilization (McLaren, et al., 2004;Reddy, et al., 1999;Soros, et al., 2001). Interestingly, Sam68 has also been found to colocalize and associate with RNA splicing factors, thus regulating alternative splicing (Denegri, et al., 2001;Hartmann, et al., 1999;Matter, et al., 2002). More importantly, this regulation of alternative splicing, and therefore gene expression, is dependent on extracellular signaling through Erk kinases (Matter, et al., 2002). It can be hypothesized that nuclear localization of Brk, and the subsequent phosphorylation and inhibition of its nuclear substrate Sam68, might influence splicing associated functions of Sam68 and hence regulate gene expression.

The increased susceptibility of Brk deficient mice to inflammation and injury was further supported by the results of the DSS/AOM colitis-associated tumor model reported in this study. Wild-type and Brk deficient mice were subjected to a single injection of AOM followed by a 1-week treatment with 2% DSS in drinking water. This protocol has been shown to effectively and specifically induce colonic adenocarcinoma and colitis in a relatively short time (Tanaka, et al., 2003). It is well known that increased cell turnover and chronic inflammation promotes and greatly predisposes to cancer development (Itzkowitz and Yio, 2004;Sancho, et al., 2004). Since loss of Brk resulted in increased epithelial cell proliferation and chronic inflammation in the intestine, it was hypothesized that Brk deficient mice would be more prone to tumor development in a mouse colon tumor model, further supporting putative tumor suppressor functions of Brk in the gastrointestinal tract. This was confirmed in a pilot study utilizing the traditional AOM mouse model of colon carcinogenesis, where mice receive weekly AOM injections for the duration of 6 weeks and are sacrificed 4 or 24 weeks after the last injection (Boivin, et al., 2003). Ten weeks after subjecting wild-type and knockout mice to the AOM pro-carcinogen, knockout mice exhibited a considerably higher incidence of aberrant crypt foci (ACF) formation within the distal colon compared to wild-type controls (Heap and Bie, unpublished). At 30 weeks after AOM treatment, knockout mice developed colonic tumors, which were absent in wild-type controls.


In stark contrast to this study were the results obtained in the accelerated DSS/AOM tumorigenesis model. In this carcinogenesis model Brk deficient mice showed a reduced colitis-associated tumor incidence when compared to wild-type animals. Furthermore, wild-type mice showed increased mortality upon treatment. Combination of the initial injection of the carcinogen AOM with 2% DSS treatment to induce acute inflammation and thus promoting tumor development actually had the opposing effect in knockout animals. The reason for these seemingly paradoxical results lies in the response to the DSS treatment itself. Previously an immediate acute inflammatory response with severe loss of crypt epithelium to DSS treatment has been observed in Brk deficient mice. Thus, the cells carrying oncogenic mutations caused by AOM might actually be ablated during the process of acute inflammation resulting in the lower tumor incidence observed in Brk deficient mice. In wild-type mice on the other hand, after the initial low dose of AOM, tumor formation is promoted by the induction of inflammation through treatment with 2% DSS. The "AOM-initiated cells" carrying oncogenic mutations are supported by the inflammation driving these mutant cells to proliferate and giving preneoplastic and neoplastic cells a growth advantage. This hypothesis may explain the high tumor incidence in wild-type mice whereas Brk deficient mice are seemingly more resistant to tumor formation. Due to the increased susceptibility of Brk deficient mice to inflammation and injury, the accelerated DSS/AOM tumorigenesis model appears to be a “bad” choice to confirm putative tumor suppressor functions for Brk in the intestine. To test this hypothesis, further studies with lower doses of DSS (0.5%) as well as the long-term “AOM only” tumor model are underway.

Our findings reveal a new role for Brk tyrosine kinase signaling in the maintenance of epithelial homeostasis and protection from intestinal injury. Brk kinase signaling might play a crucial role in signaling pathways involved in the response to injury. Exactly which pathways Brk interacts with remains to be determined.

4.3 Regulation of apoptosis by Brk signaling

The intestinal epithelium is a rapidly renewing tissue in which cells undergo topographically organized proliferation and differentiation. Production of epithelial cells occurs in the crypts, which then differentiate and migrate toward the apex of the villus. Eventually, cells are shed at the tip of the villus or the surface epithelium. This cell loss is precisely balanced in steady state by cell division occurring in the crypt compartment. Intensive studies of cell kinetics of this system revealed that approximately 1000 cells are being shed per small intestinal villus each day. It has been shown that apoptosis or programmed cell death is required for maintaining intestinal tissue homeostasis, the defense against certain pathogens and elimination of unwanted cells (Vachon, et al., 2001;Vachon, et al., 2000). Furthermore, mutations affecting critical genes that regulate cell proliferation and survival cause fatal cancers.


As previously discussed, loss of Brk results in increased epithelial cell turnover and the physiological appearance of longer villi in the gastrointestinal tract of mice. Apoptosis is a crucial mechanism required to counterbalance the crypt cell division (Potten, 1992), and the appearance of longer villi in Brk deficient mice suggested an impaired balance between cell production and cell loss. In keeping with this hypothesis, no statistically significant differences in spontaneous apoptosis within the crypt epithelia of untreated wild-type and Brk knockout mice were observed. However, Brk deficient mice showed severe inability to respond to genotoxic stress, such as radiation, suggesting a role for Brk in DNA-damage induced apoptosis. Cellular damage can initiate apoptotic cell death, both in vitro and in vivo. Whether a cell will survive or undergo apoptosis after DNA damage greatly depends on patterns of gene expression determining a survival threshold and modulating the engagement of apoptosis (Dive and Hickman, 1991, Salvesen, 1997 #3488). The crypts of the small intestine provide a readily accessible and quantifiable model to study the apoptotic response to genotoxic damage in vivo (Merritt, et al., 1997). Following DNA-damage, the stem cells in the small intestinal crypt are committed to an altruistic cell suicide, which is required to avoid the accumulation of oncogenic mutations, resulting in the protection against cancer (Potten, 1992).

Whole-body γ-irradiation of mice with doses of 8 Gγ cause p53-dependent apoptosis in the small intestinal crypts with 6 apoptosis-susceptible cells in each crypt (Potten, 2004). Wild-type and Brk deficient mice were subjected to whole-body γ-irradiation and the amount of apoptotic cells was microscopically identified by caspase 3 activation. Interestingly, Brk deficient mice were more resistant to apoptosis induced by γ-irradiation than their wild-type counterparts. They exhibited a twofold reduction in their response to radiation, with only 2-3 apoptotic cells in each crypt. In addition, after ionizing radiation, induction of Brk protein was observed in wild-type mice. After DNA damage, Brk expression was not only detected in non-proliferating cells of the villus but also in the crypt compartment of the intestinal epithelium, which usually excludes expression of this tyrosine kinase. These data support the notion that Brk is required for the altruistic cell suicide of the crypt stem cell following DNA damage. Thus, Brk is not only protecting the small intestine against injury, as has been shown in the DSS model, but also against oncogenic mutations caused by irradiation. These facts support a tumor suppressor function for Brk in the intestine, since loss of this kinase results in impaired apoptosis.

Brk signaling might contribute to the induction of apoptosis by regulating the PI3-kinase/Akt survival pathway. Brk has been shown to phosphorylate and inhibit Akt kinase activity and downstream signaling in unstimulated cells (Zhang, et al., 2004). Furthermore, untreated Brk deficient mice exhibited increased Akt kinase activity. Similar to untreated conditions, Brk mutant mice showed constitutively higher levels of phospho-Ser473 Akt and Akt kinase activity after γ-irradiation when compared to wild-type controls. The serine/threonine kinase Akt has been reported to mediate cell survival by various growth factors and cytokines in a variety of cell types and to block and protect against apoptosis induced by multiple apoptotic stimuli (Aikawa, et al., 2000;Datta, et al., 1999;Fujio, et al., 2000;Kandel and Hay, 1999;Matsui, et al., 1999). Brk seems to be acting as a repressor of Akt activity, thus regulating growth and survival in the intestinal epithelium. Negative regulation of Akt, as shown in the case of the PTEN phosphatase tumor suppressor, results in the promotion of apoptosis, whereas loss of this repression has been implicated in a wide range of human cancers (Haas-Kogan, et al., 1998;Li, et al., 1998;Stambolic, et al., 1998;Wu, et al., 1998). The observed induction of Brk protein expression in the crypt stem cells of wild-type animals after irradiation, together with increased Akt kinase activity in Brk deficient mice, suggest that repression of Akt pro-survival signaling by Brk in the crypt might be one mechanism contributing to DNA-damage induced apoptosis in the intestine.


Furthermore, loss of Brk resulted in increased phosphorylation and therefore activation of extracellular signal-regulated kinases (Erk1/2) following irradiation. Differential involvement of the Erk signaling pathways in the regulation of epithelial cell survival through regulation of balanced transcription of anti-apoptotic and pro-apoptotic genes has been reported (Gauthier, et al., 2001;Gauthier, et al., 2001). Activation of Erk1/2 has been shown to promote cell survival, whereas activation of the stress-activated c-Jun N-terminal kinases (p46/p54JNK) and p38MAPK induces apoptosis (Wang, et al., 1998;Wang, et al., 1998;Xia, et al., 1995). The activation of Erk kinases in Brk deficient mice, together with activated Akt kinase activity, indicates that these pro-survival signaling pathways might contribute to the resistance of Brk deficient mice against apoptosis. Brk signaling appears to be required for the induction of altruistic cell death upon cellular or DNA damage. Therefore, signaling mediated by this tyrosine kinase is protecting the integrity of the small intestine against genetic alterations that would otherwise cause cell transformation and the development of cancer. This is confirmed by the previously discussed AOM tumorigenesis pilot study, in which Brk deficient mice showed increased formation of ACFs (Heap and Bie, unpublished). The resistance of Brk mutant mice to apoptosis, and the resulting continued survival of cells with DNA damage and/or oncogenic mutations, presents the potential for carcinogenic transformation. Thus, Brk deficient mice would be more prone to tumor development, supporting putative tumor suppressor functions of Brk in the gastrointestinal tract.

It should also be noted, that Brk can shuttle between the nuclear and cytoplasmic cell compartments, and therefore could conceivably transmit the DNA damage signal from the nucleus to the cytoplasm. This has been shown to be the fact for the pro-apoptotic function of the c-Abl tyrosine kinase (Taagepera, et al., 1998). In other words, upon DNA-damage nuclear Brk tyrosine kinase may be activated and exit the nucleus, resulting in further inhibition of Akt kinase activity and thus pro-survival signaling.

In addition to these in vivo observations, Brk is also involved in regulating apoptosis in an in vitro model system. Overexpression of Brk in Rat1A fibroblasts sensitizes these cells to apoptosis induced by serum starvation or a combination of UV irradiation/ serum starvation in vitro. Rat1a cells are immortal nontransformed rat embryo fibroblasts that are susceptible to oncogenic transformation and exhibit serum- and Akt-dependent susceptibility to a variety of apoptotic stimuli (Topp, 1981). Increased susceptibility to apoptosis of fibroblasts expressing active Brk further supports a significant role for Brk tyrosine kinase in the regulation of the apoptotic response to DNA-damage. In addition, apoptosis in these fibroblasts is dependant on Akt signaling, and therefore it would be interesting to explore if the introduction of dominantly active Akt into Brk stable Rat1A fibroblasts is able to rescue cells from Brk-mediated apoptosis.

4.4 Brk signaling in tumor suppression and cancer development


Tumor suppressors normally act as inhibitors of cell proliferation or activators of apoptosis and use a variety of mechanisms in tissue growth suppression (Macleod, 2000;Sherr, 2004). Brk is positioned to inhibit cell proliferation by its restriction to differentiated cells in the small intestinal villus or colon. Loss of Brk resulted in increased epithelial cell turnover most likely due to deregulated Wnt signaling. Increased accumulation of nuclear β-catenin and upregulation of the β-catenin target gene c-myc were observed in Brk deficient mice. In addition, these mice exhibited increased pro-survival signaling in terms of increased Akt kinase activity. Brk has been shown to inhibit Akt kinase activity in vitro (Zhang, et al., 2004), and thus it is conceivable that loss of Brk results in increased Akt activity in the intestine. This in turn, could contribute to the increased nuclear accumulation of β-catenin in Brk deficient mice. It could be proposed that Akt activation and β-catenin signaling can be effectively controlled by the availability and activity of Brk protein through differential distribution of Brk in various tissues, cells, or subcellular localization. However, co-localization of Brk with Akt in intestinal epithelial cells still needs to be confirmed.

Furthermore, Brk deficient mice exhibited chronic inflammation, which, as shown in inflammatory bowel disease, significantly increases the risk of developing intestinal cancer. When subjected to DSS, a colon injury model, Brk deficient mice undergo severe epithelial injury and cell loss. This could in part be contributed to the deregulated intestinal tissue homeostasis observed in Brk deficient mice. In this injury model, wild-type animals showed significant upregulation of Brk protein levels and strong nuclear expression of this tyrosine kinase throughout the intestinal epithelium in response to DSS, suggesting a role for Brk in protecting the epithelium from injury. In addition, mice carrying a loss of function mutation in the brk gene are more resistant to DNA damage induced apoptosis in the small intestinal crypt. No differences in base-line apoptosis were observed in untreated animals. However, when subjected to γ-irradiation, Brk deficient animals are significantly more resistant. Thus, these mice are impaired in the altruistic cell suicide, a crucial mechanism protecting the stem cells in the small intestinal crypt against the accumulation of oncogenic mutations and therefore cancer (Potten, 1992). Wild-type mice on the other hand, displayed apoptosis in response to irradiation at levels consistent with the literature (Potten, 2004). In addition, they exhibited a rapid radiation-induced upregulation of Brk protein levels and expression of Brk in crypt cells, which usually exclude its presence. These data suggest that Brk functions as a determinant of cellular sensitivity to genotoxic stress in the intestinal epithelium.

Overall one could propose, that in normal tissues such as the intestinal epithelium, Brk has tumor suppressor functions such as the regulation of differentiation and cell cycle exit and the protection of the epithelium against injury and oncogenic mutations (Fig. 21).


Fig. 21: Brk tumor suppressor qualities. Brk is required for intestinal homeostasis. It is positioned to inhibit cell proliferation in the intestinal epithelium. Loss of Brk results in increased epithelial cell turnover and chronic inflammation in the intestine. In addition, Brk protects the intestine against injury and oncogenic mutations. The Brk tyrosine kinase is required for the induction of intestinal apoptosis following genotoxic stress.

However, in contrast to the gastrointestinal tract, Brk is expressed in a significant proportion of primary breast tumors and breast tumor cell lines, although it is not expressed in normal mammary gland epithelial cells (Barker, et al., 1997). Furthermore, Brk protein was detected in metastatic melanoma cell lines (Easty, et al., 1997), as well as head and neck squamous cell carcinomas (Lin, et al., 2004), and it is capable of promoting cell motility, migration and invasion through phosphorylation of paxillin (Chen, et al., 2004). These seemingly paradoxical roles of Brk during differentiation and tumorigenesis are still poorly understood. However, protein mislocalization has been shown to play an important role in the development of cancer and has been observed for a number of signaling proteins (Kau, et al., 2004). Differential intracellular localization of Brk has been observed in several epithelial cancers of varying malignancy grades (Derry, et al., 2003;Derry, et al., 2000;Petro, et al., 2004;Serfas and Tyner, 2003). Thus, the ability of Brk to associate with distinct sets of substrates in different cellular compartments in normal tissues and cancer cells may lead to the activation of divergent signaling pathways. Its role in biology is likely to be determined by its relative kinase activity in coordination with its interaction with other proteins as part of regulated signaling complexes. While functions of most Brk family kinases are poorly understood, a common feature of Brk family proteins is to limit receptor kinase signaling in untransformed cells (possibly by functioning as co-inhibitors of Akt or inhibitors of Ras pathway signaling) (Lu and Li, 1999;Zhang, et al., 1999), or during differentiation of skin or gut epithelial cells (Anneren and Welsh, 2000;Oberg-Welsh, et al., 1998;Vasioukhin, et al., 1995;Vasioukhin and Tyner, 1997). In cancer cells on the other hand, association of these kinases with de-regulated signaling complexes might contribute to cancer development.

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated: