Search tips
Search criteria 


Logo of moloncolLink to Publisher's site
Mol Oncol. 2014 December; 8(8): 1667–1678.
Published online 2014 July 9. doi:  10.1016/j.molonc.2014.07.001
PMCID: PMC5528598

Patterns of DNA damage response in intracranial germ cell tumors versus glioblastomas reflect cell of origin rather than brain environment: Implications for the anti‐tumor barrier concept and treatment


The DNA damage response (DDR) machinery becomes commonly activated in response to oncogenes and during early stages of development of solid malignancies, with an exception of testicular germ cell tumors (TGCTs). The active DDR signaling evokes cell death or senescence but this anti‐tumor barrier can be breached by defects in DDR factors, such as the ATM‐Chk2‐p53 pathway, thereby allowing tumor progression. The DDR barrier is strongly activated in brain tumors, particularly gliomas, due to oxidative damage and replication stress. Here, we took advantage of rare human primary intracranial germ cell tumors (PIGCTs), to address the roles of cell‐intrinsic factors including cell of origin, versus local tissue environment, in the constitutive DDR activation in vivo. Immunohistochemical analysis of 7 biomarkers on a series of 21 PIGCTs (germinomas and other subtypes), 20 normal brain specimens and 20 glioblastomas, revealed the following: i) The overall DDR signaling (γH2AX) and activation of the ATM‐Chk2‐p53 pathway were very low among the PIGCTs, reminiscent of TGCTs, and contrasting sharply with strong DDR activation in glioblastomas; ii) Except for one case of embryonal carcinoma, there were no clear aberrations in the ATM‐Chk2‐p53 pathway components among the PIGCT cohort; iii) Subsets of PIGCTs showed unusual cytosolic localization of Chk2 and/or ATM. Collectively, these results show that PIGCTs mimic the DDR activation patterns of their gonadal germ cell tumor counterparts, rather than the brain tumors with which they share the tissue environment. Hence cell‐intrinsic factors and cell of origin dictate the extent of DDR barrier activation and also the ensuing pressure to select for DDR defects. Our data provide conceptually important insights into the role of DNA damage checkpoints in intracranial tumorigenesis, with implications for the differential biological responses of diverse tumor types to endogenous stress as well as to genotoxic treatments such as ionizing radiation or chemotherapy.

Keywords: Intracranial germ cell tumors, Gliomas, DNA damage signaling, Testicular germ cell neoplasms, ATM-Chk2-p53 pathway


  • Human PIGCTs show very low DDR activation, similar to TGCTs, and unlike gliomas.
  • ATM‐Chk2‐p53 pathway defects are rare in PIGCTs, in contrast to other solid tumors.
  • Chk2 and ATM show cytoplasmic localization in normal brain and subsets of PIGCTs.
  • DDR barrier activation reflects tumor cell origin, rather than brain environment.

1. Introduction

The cellular DNA damage response (DDR) machinery becomes activated in response to various genotoxic insults, both environmental such as ultraviolet or ionizing radiation, and endogenous, such as reactive oxygen species produced in metabolic processes. The DDR is a complex network of signaling and effector pathways that encompasses cell cycle checkpoints, diverse DNA repair mechanisms, impacts other aspects of cellular function such as transcription and chromatin remodeling, and can induce cell death in cases of irreparable DNA lesions or damage overload (Jackson and Bartek, 2009). At the center of the DDR network are two kinase signaling modules that spread the signal by phosphorylating a plethora of cellular proteins: i) the ATM‐Chk2 axis reacting mainly to DNA double strand breaks (DSBs), and the ATR‐Chk1 pathway activated in response to diverse DNA lesions including replication intermediates under conditions of replication stress (Bartek et al., 2004; Kastan and Bartek, 2004).

One of the more recently discovered biological roles of the DDR machinery, shared by the ATM‐Chk2 and ATR‐Chk1 cascades, is to provide an intrinsic barrier against oncogenic insults that commonly evoke replication stress and DNA breakage (Bartkova et al., 2005a; Gorgoulis et al., 2005; Bartkova et al., 2006; DiMicco et al., 2006; Bartek et al., 2007; Halazonetis et al., 2008). The activated DDR thereby helps prevent or delay progression of the early stages of tumor development to full malignancy, among other mechanisms by triggering p53‐dependent senescence or apoptosis of the nascent tumor cells. In addition to the DDR, also the ARF tumor suppressor is capable of triggering the p53‐mediated ‘anti‐cancer barrier’ (Sherr, 2012), however recent analyses of a spectrum of mouse models, cohorts of human clinical specimens and cell culture models of regulatable oncogenes have revealed that the DDR barrier is more sensitive to oncogenic insults and becomes activated earlier than ARF during multistep tumorigenesis (Evangelou et al., 2013; Velimezi et al., 2013). Consistent with their tumor barrier roles, both the DDR checkpoints and ARF, as well as p53, are often lost or inactivated in those tumors that do progress to advanced stage (Bartkova et al., 2005a; Gorgoulis et al., 2005; Halazonetis et al., 2008; Sherr, 2012).

Given the sensitivity of the DDR machinery, capable of reacting to a single oncogene (Evangelou et al., 2013), it is perhaps not surprising that the DNA damage signaling, such as through the ATM‐Chk2‐p53 pathway, is found activated in the majority of early stages of virtually all types of human solid tumors, including diverse types of epithelial neoplasms (Bartkova et al., 2005, 2005, 2007, 2007, 2013, 2005) or gliomas (Bartkova et al., 2010). There is, however, one notable exception from such widespread DDR activation among solid tumors: namely the very modest, if any DDR activation observed in human testicular germ cell tumors (TGCTs) including their preinvasive stage, the carcinoma in situ (CIS) testis (Bartkova et al., 2007, 2007, 2003). The striking contrast between the high frequency and extent of DDR activation in solid tumors such as gliomas (the tumor type with so far the most prominent DDR activation) on the one hand, and TGCTs with their hardly detectable endogenous activation of the DDR on the other, raises important questions about the biological basis of such a difference, and its potential impact on tumorigenesis and response to treatments including ionizing radiation and genotoxic chemotherapy. Based on the observed correlation between the low extent of DDR activation and low incidence of tumor‐associated aberrations of DDR factors such as p53, MDC1, ATM or 53BP1 among TGCTs, and such DDR defects being common in those types of solid tumors that show the high degree of DDR activation early during tumor development, we proposed that the activation of DDR checkpoints and the ensuing senescence or cell death creates selective pressure for survival of tumor cells with mutated p53 and other loss‐of‐function DDR defects (Bartkova et al., 2007a). Such selective pressure would then be low among TGCTs the DDR checkpoints of which are not significantly activated, hence avoiding the necessity to select for p53 mutations and other DDR defects as a strategy to escape from the activated DDR anti‐cancer barrier. A related, clinically most relevant issue is the exceptionally high degree of curability among TGCTs, again in contrast to often poor response among most other solid malignancies, using genotoxic treatments such as ionizing radiation or platinum drugs. Despite some recent insights (reviewed in Burrell et al., 2013; Koster et al., 2013), the cellular and molecular basis for the exceptionally favourable response of TGCTs to genotoxic treatment modalities are not fully understood. There are at least three major plausible reasons to explain this phenomenon that come to mind: i) As TGCTs originate from germ cells, while other solid tumors are derived from somatic cells, the cell of origin may decisively affect the treatment response; ii) There could be pronounced biological differences due to the local organ/tissue environment, and testes may provide a particularly special niche, distinct from epithelial organs or the central nervous system; iii) The differential treatment responses may reflect distinct spectra of molecular lesions (activated oncogenes and loss of tumor suppressors) that lead to malignant transformation of TGCTs and other solid tumors, respectively. These are conceptually important yet very challenging issues, not easy to address experimentally as both cell culture and xenograft models are inadequate to mimic the human tissue niches and full complement of potentially important factors including the immune system and inflammatory cytokines known to be capable of mediating the DNA damage response at a distance (Bartek et al., 2008; Campisi and d'Adda di Fagagna, 2007), for example.

To assess the relative contributions of tumor cell of origin versus environmental factors in clinically relevant settings, one would ideally like to compare situations in which tumors of the same origin, for example derived from germ cells, can be studied as naturally occurring lesions in different organ environments, not only testes. Indeed, there is such ‘experiment of nature’, a group of rare human tumors known as primary intracranial germ cell tumors (PIGCTs). While germ cell tumors (GCTs) are most common in the testis of young adults, they are also found (albeit rarely) as primary tumors in the central nervous systém (CNS), located in midline structures. PIGCTs have been hypothesised to share a common origin from germ cell progenitors based on similar histopathology between these tumors in different locations. An embryonic cell theory has been proposed for the PIGCTs, which postulates misplacement of embryonic germ cells in the brain due to aberrantly aborted migration of the germ cell precursors during early embryonic development (Teilum, 1965). More recent analyses of various germ cell markers as well as the spectra of tumorigenic defects largely support the notion that gonadal and intracranial GCTs share the origin from germ cell precursors (Hoei‐Hansen et al., 2006; Oosterhuis et al., 2007; Rajpert‐De Meyts et al., 2003; Terashima et al., 2014). The PIGCTs occur primarily in children and adolescents, and account for approximately 0.5% of malignant pediatric tumors (Gobel et al., 2000). As GCTs found elsewhere, there are a variety of histological types among PIGCTs, including germinomas (equivalents of the ovarian dysgerminomas and the testicular seminomas), embryonal carcinomas, yolk sac tumors and teratomas (Rosenblum et al., 2000). Due to the low incidence of PIGCTs, research has been limited, and there have been no analyses of cell cycle checkpoints or DDR factors in PIGCTs published so far.

In this study, we took advantage of the fact that PIGCTs share the germ cell origin with testicular GCTs, yet share the organ location with CNS tumors such as glioblastomas. We collected archival specimens from a cohort of 21 cases of human PIGCTs, and used this material for immunohistochemical analysis of a series of DDR factors, including components of the ATM‐Chk2‐p53 pathway and their activation status, along with the global DDR activation/signaling marker γH2AX (Bartkova et al., 2005a; Gorgoulis et al., 2005) and the standard proliferation marker Ki67. The results from these analyses, and their implications for conceptual debate on the DDR anti‐tumor barriers, role(s) of DDR defects and tissue environment in tumorigenesis and their relevance for responses to genotoxic treatment of cancer, are presented below.

2. Material and methods

2.1. Intracranial tumor material and normal tissue

The material analysed in this study was thoroughly described in (Hoei‐Hansen et al., 2006), and included tumor samples from 13 male and 8 female patients, median age 12 years (range 0–30 years) diagnosed with CNS GCTs in Zeeland, Denmark (or Greenland, patient M12) in 1987–2004. Tumors were localised in the midline of the brain (see Table 1). Our series of formalin‐fixed tissue comprised: 11 germinomas, one embryonal carcinoma, one germinoma with yolk sac tumor and polyembryoma, one mixed malignant GCT, one mature and three immature teratomas, and two malignant teratomas. One tumor was not classifiable with standard immunohistochemical stainings, but had GCT morphology (M12). The Regional Committee for Medical Research Ethics in Denmark approved the project. The control normal human brain tissue specimens, and the glioblastoma samples used for comparative purposes with the PIGCTs, have been described previously (Bartkova et al., 2010).

Table 1

Intracranial GCT tumor cohort and immunohistochemical staining results. For each protein the staining intensity is listed followed by percentage of tumour cells stained and the localisation of the signal (nuclear or cytoplasmic).

2.2. Primary antibodies

Antibodies used in this study included mouse monoclonal antibodies to human phospho‐histone H2A.X (Ser139) (Millipore, clone JBW301, diluted 1:2500), ser1981‐phosphorylated, activated ATM (Rockland, USA, diluted 1:250), p53 (our own antibody DO.1, diluted 1:4000, Bartek et al., 1993), Chk2 kinase (our own clone DCS‐270, diluted 1:5000; Lukas et al., 2001), and rabbit antibodies to ATM kinase (Y170, from Abcam, UK, diluted 1:500); Thr68‐phosphorylated, activated Chk2 (Cat.No.2584 Cell Signaling Technology, Danvers, MA, US, diluted 1:250) and Ki67 (Ki‐S5, M7187, DakoCytomation, Glostrup, DK, diluted 1:200).

2.3. Immunohistochemistry procedures

For the sensitive immunohistochemistry procedure without nuclear counterstaining, the slides were deparaffinized and incubated with the primary mouse or rabbit antibody against the selected human protein or protein modification, incubated overnight, followed by the indirect streptavidin–biotin–peroxidase method using the Vectastain Elite kit (Vector Laboratories, Burlingame, CA, USA) and nickel sulphate‐based chromogen enhancement detection as previously described (Bartkova et al., 2001). Alternatively, to better visualize the subcellular localization, a standard immunoperoxidase method including light nuclear counterstaining with hematoxylin was employed, as previously published (Hoei‐Hansen et al., 2006). Briefly, the dewaxed and rehydrated sections were pre‐treated in a microwave oven in a buffer (for Ki67: TEG buffer = TRIS 1.21 g/L, EGTA 0.19 g/L, pH = 9.0; for CHK2 and pCHK2: citrate buffer (10 mmol/L, pH = 6.0). Subsequently, the sections were incubated with 1.5% H2O2 to inhibit the endogenous peroxidase, followed by non‐immune goat serum to block unspecific binding sites. Incubation with the primary antibody was overnight at 4 °C. A secondary biotinylated goat anti‐mouse or goat anti‐rabbit link antibody was applied, followed by horseradish peroxidase–streptavidin complex. Between all steps sections were thoroughly washed. Visualisation was with aminoethyl carbazole substrate (all reagents from Zymed, S. San Francisco, CA), and light counterstaining with Mayer's hematoxylin. For both methods, positive and negative controls were run in parallel. For negative controls, a serial section was incubated with non‐immune mouse or rabbit sera. For positive controls sections of normal testicular tissue, tissue containing testicular carcinoma in situ, seminoma and a non‐seminoma was analysed.

2.4. Data evaluation and photography

Evaluation of staining patterns was done by at least two independent observers, and the scoring reflected both the percentage of positive cells, and the intensity of staining, and the overall patterns were scored as negative (−), weak (±), moderate (+) or strong (++). More detailed assessment of the immunohistochemical results including subcellular localization and heterogeneity of the staining patterns is provided in the Results section, and numerous examples are shown in in1,1, ,2,2, ,3,3, ,4,4, ,6.6. Both evaluation and photography was performed using microscopes and photography equipment from Zeiss (Oberkochen, Germany).

Figure 1

Examples of immunohistochemical staining patterns for DNA damage response proteins in normal human brain tissue. The names of the proteins (Chk2) or their phosphorylated forms (Chk2‐P, γH2AX) are indicated directly in figure panels. Ki67 ...

Figure 2

Expression of ATM, ATM‐P, γH2AX and p53 in human intracranial germinomas. (A, B) – Sets of images from two different germinoma specimens, with the examined markers indicated in the individual panels. Note the variable expression ...

Figure 3

Expression of ATM, ATM‐P, γH2AX and p53 in an intracranial embryonal carcinoma, detected by immunohistochemistry. Note the aberrant absence of ATM and enhanced nuclear p53 in a fraction of cancer cells. Scale bars: 50 μm. ...

Figure 4

Examples of immunohistochemical staining patterns for DNA damage response markers in an intracranial teratoma, compared to glioblastoma. (A) Upper row, from left to right: Teratoma stained for total ATM, parallel section showing very low/absent ATM‐P; ...

Figure 6

Localisation and activation of Chk2 kinase in intracranial germinomas. (A) A germinoma with tumor cells expressing Chk2 localized in both cytoplasm (arrow) and nucleus (arrowhead). (B, C) Parallel sections from another germinoma, stained for Chk2 and ...

3. Results

The overall results from the immunohistochemical analyses of all the seven markers are presented in Table 1, along with the histological subtype of the individual PIGCTs and the sex of the patient. Examples of the staining patterns seen in a series of control tissues from normal human brain are shown in Figure 1. Consistently, there was no detectable activation of the ATM‐Chk2‐p53 pathway in any of the normal brain samples (n = 20), regardless of the brain area examined. What was striking in the normal brain cells was the predominantly cytosolic localisation of the Chk2 kinase and partly also the ATM kinase (Figure 1A and data not shown).

Among germinomas, the subtype of the PIGCTs that was most common in our cohort, both Chk2 and ATM kinase proteins were expressed, but were either nonphosphorylated (inactive) or activated in only a very minor fraction of tumor cells (Figure 2A, B, Table 1). Consistently, also the γH2AX marker of DDR signaling and the p53 protein were barely detectable, mostly in fewer than 5% of tumor cells (Figure 2, Table 1).

In the non‐germinoma subtypes of PIGCTs, the overall trend for only low extent of DDR activation was also apparent. Arguably the most ‘exceptional’ pattern of the markers examined was apparent in the only case of embryonal carcinoma (case M13), where the relatively low level of γH2AX might reflect the aberrant absence of ATM (the only case of such aberrant lack of ATM in our cohort) and this was accompanied by the highest fraction of tumor cells highly positive for p53 (Figure 3, Table 1). Such p53 pattern is borderline for possibly reflecting a p53 mutation, which otherwise occurs rarely among germ cell tumors (see Discussion for more details). The embryonal carcinoma was also one of the only 3 cases that showed the Ki67‐positive proliferation fraction of 10% or more tumor cells, with most PIGCT cases showing a rather low Ki67 index between 1 and 5% (Table 1).

In the subset of PIGCTs with features of teratoma or mixed germinoma and non‐germinoma components, the general pattern seen was again the one of low extent of DDR activation/signaling, no aberrant loss of ATM or Chk2 and no pronounced p53 stabilisation/protein overexpression. Examples of immunohistochemical staining of teratoma components are shown in Figure 4 (in which representative images from glioblastoma are shown side by side for comparison), and the summary of the results is apparent from the data presented in Table 1.

Returning to the issue of the impact of cell of origin versus tissue environment, we present the global marker of DNA damage response signaling, namely the scoring of fractions of tumor cells positive for γH2AX. This comparative analysis in the present cohort of intracranial PIGCTs is presented in parallel to normal brain tissue and to our recently examined cohort of human glioblastomas (Figure 5). Apart from the lack of DDR signaling in normal brain, what is obvious from this γH2AX summary graph is that the degree of constitutive signaling is restricted to the two low categories of scoring among PIGCTs, in sharp contrast to glioblastomas, whose scores were exclusively in the upper two categories, i.e. in the range that was not reached by even a single case of the PIGCT.

Figure 5

A summary graph documenting pronounced differences between DDR activation (γH2AX) in human intracranial germ cell tumors versus glioblastomas. Whereas there is a high degree of γH2AX among human glioblastomas (all GBM cases show the activation ...

Last but not least, an interesting phenomenon related to the central nervous system is the unusual cytoplasmic localisation of Chk2 and ATM kinases, which is an exception among all normal human tissues examined (see Figure 1 for an example of cytosolic Chk2). Notably, analogous cytosolic (or often both cytosolic and nuclear) expression was found in approximately one quarter of the PIGCT cases for the Chk2 kinase, and approximately one half of the cases for the ATM kinase (Figure 6 and Table 1). The molecular basis for this intriguing finding and its restriction to only some of the PIGCTs is currently unknown, but it is very unlikely to be a result of antibody cross‐reactivities for the following reasons: i) Normal brain tissues show such cytosolic localization invariably, in all 20 cases examined, and regardless of whether the tissue was obtained as a biopsy or necropsy; ii) The cytosolic staining of Chk2 was confirmed by independent antibodies, known to recognize distinct epitopes of the Chk2 protein (our unpublished results), and iii) a somewhat analogous cytosolic localisation may be true for even other DDR factors that are otherwise largely restricted to nuclei in non‐CNS cells (our unpublished results). The implications and interpretation of these staining patterns are considered in the Discussion.

In conclusion, our analysis of a series of DDR markers in a cohort of rare human intracranial germ cell tumors revealed a low overall degree of constitutive DDR activation. Such pattern is reminiscent of that previously seen in testicular germ cell tumors, and differs markedly from the much more pronounced DDR signaling seen in other types of solid human tumors including those of the central nervous system that share with PIGCTs the organ environment (Figure 7).

Figure 7

Schematic representation of the overall patterns of DDR activation and tumor‐associated DDR defects in diverse types of human tumors. Note that testicular germ cell tumors share with the intracranial germ cell tumors the low degree of DDR activation, ...

4. Discussion

The results of this study provide novel insights into mechanisms of tumorigenesis, particularly the relative contributions of cell‐intrinsic factors and cell of origin versus local tissue environment, the constitutive DNA damage response activation as part of a barrier to tumor development and selective pressure for accumulation of DDR defects that facilitate the multistep tumor progression. These results are also instrumental in the light of the conceptual issues of germ cell development and biology, regulatory switch between mitotic and meiotic cell divisions, and responses of diverse types of tumors to genotoxic therapies.

First, our data further support the notion that the intracranial GCTs do indeed most likely originate from a developmentally mis‐migrated precursor – a totipotent primordial germ cell ‘left behind in the inappropriate location’ of the CNS midline during early development. Despite some arguments about a potential origin of PIGCTs from some kind of neural progenitor cells (Scotting, 2006) the bulk of studies on this topic indicate that both the basic morphology and pathology, as well as patterns of embryonic pluripotency genes and germ cell‐specific features are consistent with the concept of PIGCTs being derived from primordial germ cells (reviewed in: Mosbech et al., 2014). Similarities between gonadal and non‐gonadal GCTs have mainly been analysed in pediatric cases, where imprinting status supports a common origin of pediatric gonadal and pediatric non‐gonadal GCTs (Schneider et al., 2001). Our present study supports the emerging view of shared origin between the testicular and intracranial GCTs, through the first analysis of the patterns of constitutive activation of the DNA damage response machinery in a cohort of human PIGCTs. Thus, our results show that in their exceptionally low extent of constitutive DDR signaling, PIGCTs very closely mimic the DDR status reported for testicular GCTs (Bartkova et al., 2005, 2007, 2007, 2003) and sharply contrast with other types of primary brain tumors such as gliomas (Bartkova et al., 2010, and this study) or medulloblastomas (our preliminary unpublished data), despite such brain tumors share with the PIGCTs the intracranial localisation (Figure 7).

Conceptually, our present dataset helps better understand the biology of the DDR activation during tumorigenesis, and its role as an inducible anti‐cancer barrier (Bartkova et al., 2005a; Gorgoulis et al., 2005; Bartek et al., 2007; Halazonetis et al., 2008). Given that the DDR activation occurs rather early during multistep tumorigenesis, as a reaction to the initial oncogenic event(s), and it precedes accumulation of p53 mutations and other DDR defects such as loss of ATM or Chk2 (Bartkova et al. 2005a; Gorgoulis et al. 2005; Bartkova et al. 2006; Evangelou et al., 2013), for example, such early activation of the DDR creates a checkpoint‐imposed barrier to tumor progression. This DDR‐driven scenario limits outgrowth of cancer cells to rare variants capable of bypassing the DDR barrier, such as those with p53 mutations (Halazonetis et al., 2008). This concept is also consistent with the observed high frequency of p53 and other DDR‐related aberrations in types of solid tumors that do show a high degree of DDR activation, such as glioblastomas or carcinomas of the lung or urinary bladder (Bartkova et al., 2005, 2010, 2005, 2013, 2005, 2013). It is furthermore also consistent with the very low frequency of p53 and DDR defects in testicular germ cell tumors that generally show very modest DDR activation during tumor development (Bartkova et al., 2005, 2007, 2007). In the context of the present study, the observed overall low DDR activation among PIGCTs would predict that the primary intracranial GCTs should also harbor relatively few defects in the DDR machinery including p53 mutations. Indeed, such prediction is consistent with our present findings of only one PIGCT tumor with a clear aberrant absence of ATM, embryonal carcinoma that also showed the highest degree of p53 expression, the only candidate for a possible case of p53 mutation (Figure 3 and Table 1). In addition, this overall pattern seen in PIGCTs is again highly reminiscent of the gonadal GCTs, in that embryonal carcinomas of the testes were the subtype of TGCTs that showed the somewhat higher degree of DDR activation as compared to e.g. seminomas or teratomas (Bartkova et al., 2005, 2007, 2007).

Another interesting aspect of our present results is related to diverse tissue environment factors, as opposed to cell‐intrinsic features, as decisively impacting the overall pattern of DDR signaling activation. As testes and the central nervous system provide two rather different tissue environments, for example with regard to high oxidative stress due to the much higher oxygen consumption in the brain compared to other organs, the intracranial location of PIGCTs allowed us to reach some interesting conclusions in this regard. The overall pattern of constitutive DDR signaling in PIGCTs was shared with testicular GCTs, rather than brain tumors such as gliomas, strongly indicating that cell‐intrinsic features, such as the cell of origin and/or the oncogenic events underlying the transformation, rather than local tissue environmental cues, represent the major factor determining the pattern of DDR activation in the intracranial tumors (Figure 7). While diverse extracellular messengers such as cytokines and chemokines are clearly capable of evoking DDR activation in cells through so‐called bystander effects, including γH2AX signaling and cellular senescence, such scenario requires strong DDR activation in some cells in the first place, as the initial wave of the cytokine/chemokine production depends on the DDR signaling (Bartek et al. 2008; Hubackova et al., 2012; Campisi and d'Adda di Fagagna, 2007; Hodny et al., 2013). An intriguing observation related to tissue‐associated peculiarities, is the unorthodox cytosolic localisation of the DDR kinases Chk2 and ATM in large subsets of the PIGCTs (reminiscent of normal brain tissue), in contrast to exclusively nuclear ATM and Chk2 seen in testicular germ cell tumors and other neoplasms. This issue is distinct from the question of DDR activation per se, and seems to involve some brain‐related influence. However even in this case the major factor, admittedly currently obscure, seems to be cell‐intrinsic rather than caused by some extracellular cues, as only fraction of the PIGCTs showed this phenotype, while the surrounding normal brain tissue showed the cytosolic Chk2 in all cases. The cell‐intrinsic cause of the cytosolic location of some DDR proteins is further supported by our analyses of peripheral nerve ganglia from both human and baboon enteric tissue biopsies. Thus, we found strong cytoplasmic Chk2 and both cytosolic and nuclear ATM staining in neurons of peripheral nerve ganglia, i.e. tissue distant from brain, yet sharing the neuronal cell origin with intracranial neurons (data not shown). In addition, some PIGCT tumors examined here showed cytosolic location of only Chk2 or only ATM, while occasionally both, again inconsistent with an overall unified messenger signaling network operating in the brain tissue environment. Future studies will need to address this interesting phenomenon and its biological significance.

One of the key questions that emerges from our present work is why do the germ cell tumors, be it the gonadal or the intracranial GCTs, differ so much from the other solid tumors in terms of extent of their constitutive DDR activation. Although discussion about this crucial question is still largely speculative, several arguments come to mind. First, given that the major factor causing the DDR activation in non‐germ cell tumors is replication stress, one possibility would be that replication stress is simply much lower among germ cell tumors, and so would be also the cellular response to it. While the proliferation rate of the PIGCT tumors in our present cohort was indeed rather low in general, the extent of DDR activation in other types of tumors does not strictly correlate with proliferation, and for example overall γH2AX signaling was greater among grade II gliomas compared to glioblastomas, despite the proliferation rate was higher among the latter (Bartkova et al., 2010). Relevant to this point, virtually all TGCTs show transcriptional repression of the RB tumor suppressor, a feature shared by the early gonocytes from which these tumors likely arise (reviewed in Bartkova et al., 2003, 2003, 2003). Importantly, as loss or viral‐mediated inactivation of RB causes replication stress and the ensuing DDR activation in human somatic cells (Tort et al., 2006), it is apparent that the germ cell tumors likely are exposed to replication stress, due to their RB defects. To explain this conundrum, we propose that primordial germ cells and early gonocytes, and by extension the germ cell tumors, may be equipped with a so far unidentified physiological mechanism that allows these cells (and germ cell tumors) to cope with replication stress in a unique way, more efficient than the common reaction that occurs in somatic cells and tumors. Another phenomenon unique to germ cells and possibly linked to lower DDR activation is the regulation of mitosis‐meiosis switch, which is a germ cell‐specific feature related to the fact that DNA double strand breaks and DDR activation with a transient upregulation of γH2AX occur physiologically during spermatogenesis‐associated meiotic recombination. Our results show that intracranial germ cell tumors maintain their intrinsic ability to suppress DDR activation (and meiosis), regardless of their ‘ectopic’ position in brain environment. We have recently performed a systematic study of mitosis‐meiosis switch in the testicular CIS (and overt TGCT), where we showed dysregulation of this switch, with the presence of some early pre‐meiosis markers but also inhibitors in place, and no signs of real meiotic entry (Jørgensen et al., 2013). This challenging issue and its relationship to the observed low degree of DDR activation among GCTs in general clearly deserves further investigation.

Yet another possibility, not completely mutually exclusive with those above, may reflect the extreme importance of preserving the genome integrity of germ cells, to minimize the danger of genetic aberrations being transferred through germ‐line to the next generation. Indeed, CHK2 and ATM have been implicated in germ cell development (Yamada and Coffman, 2005), arguing that primordial germ cells may utilise DNA damage‐induced signaling mechanisms to select against germ cells that are genetically defective, for proper maintenance of germline integrity.

To this end, germ cells are known to preferentially undergo apoptosis when challenged by DNA damage and it is plausible that germ cells and tumors derived from them trigger cell death more readily when exposed to endogenous DNA damage. As a result, there will be fewer cells with active DNA damage signaling, and rather low degree of such signaling within the cell population at any given time, since those cells that were exposed to significant DNA damage are promptly eliminated from the proliferating pool, thereby resulting in the overall pattern of only modest DDR activation. This property of germ cells and germ cell tumors may actually underlie, at least partly, the exceptional curability of germ cell tumors by DNA damaging treatment modalities such as platinum drugs or ionizing radiation that are used so successfully to treat the germ cell malignancies. Such extreme sensitivity, due to the propensity to sacrifice damaged germ cells is therefore important phylogenetically, to limit potentially hazardous genetic variation in the germ‐line, and at the same time beneficial in treatment of germ cell neoplasms. Better understanding of DDR activation mechanisms in germ cells and germ cell tumors may therefore be exploited also in future strategies not only to further optimize the therapy of germ cell tumors, but also to sensitize the naturally more resistant somatic tumors to standard‐of‐care genotoxic treatments. Last but not least, this line of research may also provide ways to better protect normal germ cells against cell death and excessive genetic instability under conditions of natural or man‐made catastrophic scenarios such as nuclear reactor disasters, for example.

Disclosure statement

Nothing to declare.


The authors wish to thank Dr H. Laursen (Department of Neuropathology, Rigshospitalet) for sections of intracranial GCT, as well as H. Kistrup, L. Andersen and A. Meisler for skilful technical assistance. This work was supported by grants from the Lundbeck Foundation (R93‐A8990), the Danish Cancer Society, the Novo Nordisk Foundation, the Danish Child Cancer Foundation, the Danish Medical Research Council (projects: 1331‐00262B/FSS, and ID4765/11‐105457), the Danish National Research Foundation, the Grant Agency of the Czech Ministry of Health (NT11065‐5) and the European Commission (projects DDResponse, TransMedChem: CZ.1.07/2.4.00/17.0015, and Biomedreg: CZ.1.05/2.1.00/01.0030).


Bartkova Jirina, Hoei-Hansen Christina E., Krizova Katerina, Hamerlik Petra, Skakkebæk Niels E., Rajpert-De Meyts Ewa, Bartek Jiri, (2014), Patterns of DNA damage response in intracranial germ cell tumors versus glioblastomas reflect cell of origin rather than brain environment: Implications for the anti‐tumor barrier concept and treatment, Molecular Oncology, 8, doi: 10.1016/j.molonc.2014.07.001.

Contributor Information

Jirina Bartkova, kd.recnac@bij.

Ewa Rajpert-De Meyts,


  • Bartek J., Bartkova J., Lukas J., Staskova Z., Vojtesek B., L .D., 1993. Immunohistochemical analysis of the p53 oncoprotein on paraffin sections using a series of novel monoclonal antibodies. J. Pathol. 169, 27–34. [PubMed]
  • Bartek J., Bartkova J., Lukas J., 2007. DNA damage signalling guards against activated oncogenes and tumour progression. Oncogene 26, 7773–7779. [PubMed]
  • Bartek J., Hodny Z., Lukas J., 2008. Cytokine loops driving senescence. Nat. Cell Biol. 10, 887–889. [PubMed]
  • Bartek J., Lukas C., Lukas J., 2004. Checking on DNA damage in S phase. Nat. Rev. Mol. Cell Biol. 5, 792–804. [PubMed]
  • Bartkova J., Bakkenist C.J., Meyts E.R., Skakkebæk N.E., Sehested M., Lukas J., Kastan M.B., Bartek J., 2005. ATM activation in normal human tissues and testicular cancer. Cell Cycle 838–845. [PubMed]
  • Bartkova J., Falck J., Rajpert-De Meyts E., Skakkebaek N.E., Lukas J., Bartek J., 2001. Chk2 tumour suppressor protein in human spermatogenesis and testicular germ-cell tumours. Oncogene 20, 5897–5902. [PubMed]
  • Bartkova J., Hamerlik P., Stockhausen M.-T., Ehrmann J., Hlobilkova a, Laursen H., Kalita O., Kolar Z., Poulsen H.S., Broholm H., Lukas J., Bartek J., 2010. Replication stress and oxidative damage contribute to aberrant constitutive activation of DNA damage signalling in human gliomas. Oncogene 29, 5095–5102. [PubMed]
  • Bartkova J., Horejsí Z., Koed K., Krämer A., Tort F., Zieger K., Guldberg P., Sehested M., Nesland J.M., Lukas C., Ørntoft T., Lukas J., Bartek J., 2005. DNA damage response as a candidate anti-cancer barrier in early human tumorigenesis. Nature 434, 864–870. [PubMed]
  • Bartkova J., Horejsí Z., Sehested M., Nesland J.M., Rajpert-De Meyts E., Skakkebaek N.E., Stucki M., Jackson S., Lukas J., Bartek J., 2007. DNA damage response mediators MDC1 and 53BP1: constitutive activation and aberrant loss in breast and lung cancer, but not in testicular germ cell tumours. Oncogene 26, 7414–7422. [PubMed]
  • Bartkova J., Lukas C., Sørensen C.S., Rajpert-De Meyts E., Skakkebaek N.E., Lukas J., Bartek J., 2003. Deregulation of the RB pathway in human testicular germ cell tumours. J. Pathol. 200, 149–156. [PubMed]
  • Bartkova J., Rajpert-De Meyts E., Skakkebaek N.E., Lukas J., Bartek J., 2003. Deregulation of the G1/S-phase control in human testicular germ cell tumours. APMIS 111, 252–265. discussion 265–6 [PubMed]
  • Bartkova J., Rajpert-De Meyts E., Skakkebaek N.E., Lukas J., Bartek J., 2007. DNA damage response in human testes and testicular germ cell tumours: biology and implications for therapy. Int. J. Androl. 30, 282–291. discussion 291 [PubMed]
  • Bartkova J., Rezaei N., Liontos M., Karakaidos P., Kletsas D., Issaeva N., Vassiliou L.-V.F., Kolettas E., Niforou K., Zoumpourlis V.C., Takaoka M., Nakagawa H., Tort F., Fugger K., Johansson F., Sehested M., Andersen C.L., Dyrskjot L., Ørntoft T., Lukas J., Kittas C., Helleday T., Halazonetis T.D., Bartek J., Gorgoulis V.G., 2006. Oncogene-induced senescence is part of the tumorigenesis barrier imposed by DNA damage checkpoints. Nature 444, 633–637. [PubMed]
  • Burrell R., McGranahan N., Bartek J., Swanton C., 2013. The causes and consequences of genetic heterogeneity in cancer evolution. Nature 501, 338–345. [PubMed]
  • Campisi J., d'Adda di Fagagna F., 2007. Cellular senescence: when bad things happen to good cells. Nat. Rev. Mol. Cell Biol. 8, 729–740. [PubMed]
  • Di Micco R., Fumagalli M., Cicalese A., Piccinin S., Gasparini P., Luise C., Schurra C., Garre' M., Nuciforo P.G., Bensimon A., Maestro R., Pelicci P.G., d'Adda di Fagagna F., 2006. Oncogene-induced senescence is a DNA damage response triggered by DNA hyper-replication. Nature 444, 638–642. [PubMed]
  • Evangelou K., Bartkova J., Kotsinas a, Pateras I.S., Liontos M., Velimezi G., Kosar M., Liloglou T., Trougakos I.P., Dyrskjot L., Andersen C.L., Papaioannou M., Drosos Y., Papafotiou G., Hodny Z., Sosa-Pineda B., Wu X.-R., Klinakis a, Ørntoft T., Lukas J., Bartek J., Gorgoulis V.G., 2013. The DNA damage checkpoint precedes activation of ARF in response to escalating oncogenic stress during tumorigenesis. Cell Death Differ. 20, 1485–1497. [PubMed]
  • Gobel U., Schneider D.T., Calaminus G., Haas R.J., Schmidt P., Harms D., 2000. Review germ-cell tumors in childhood and adolescence. Ann. Oncol. 11, (3) 263–271. [PubMed]
  • Gorgoulis V.G., Vassiliou L.F., Karakaidos P., Zacharatos P., Kotsinas A., Liloglou T., Venere M., D R.A., Kastrinakis N.G., Levy B., Kletsas D., Yoneta A., Herlyn M., Kittas C., Halazonetis T.D., 2005. Activation of the DNA damage checkpoint and genomic instability in human precancerous lesions. Nature 434, 907–913. [PubMed]
  • Halazonetis T.D., Gorgoulis V.G., Bartek J., 2008. An oncogene-induced DNA damage model for cancer development. Science 319, 1352–1355. [PubMed]
  • Hodny Z., Hubackova S., Bartek J., 2013. Cytokine-induced “bystander” senescence in DDR and immuno-surveillance. Oncotarget 4, 1552–1553. [PubMed]
  • Hoei-Hansen C.E., Sehested a, Juhler M., Lau Y.-F.C., Skakkebaek N.E., Laursen H., Rajpert-de Meyts E., 2006. New evidence for the origin of intracranial germ cell tumours from primordial germ cells: expression of pluripotency and cell differentiation markers. J. Pathol. 209, 25–33. [PubMed]
  • Hubackova S., Krejcikova K., Bartek J., Hodny Z., 2012. IL1- and TGFβ-Nox4 signaling, oxidative stress and DNA damage response are shared features of replicative, oncogene-induced, and drug-induced paracrine “bystander senescence”. Aging (Albany. NY) 4, 932–951. [PubMed]
  • Jackson S.P., Bartek J., 2009. Europe PMC Funders Group the DNA-damage response in human biology and disease. Nature 461, 1071–1078. [PubMed]
  • Jørgensen A., Nielsen J.E., Almstrup K., Toft B.G., Petersen B.L., Rajpert-De Meyts E., 2013. Dysregulation of the mitosis-meiosis switch in testicular carcinoma in situ. J. Pathol. 229, 588–598. [PubMed]
  • Kastan M.B., Bartek J., 2004. Cell-cycle checkpoints and cancer. Nature 432, 316–323. [PubMed]
  • Koster R., van Vugt M.A., Timmer-Bosscha H., Gietema J.A., de J.S., 2013. Unravelling mechanisms of cisplatin sensitivity and resistance in testicular cancer. Expert Rev. Mol. Med. 15, 12 [PubMed]
  • Lukas C., Bartkova J., Latella L., Falck J., Mailand N., Schroeder T., Sehested M., Lukas J., Bartek J., 2001. DNA damage-activated kinase Chk2 is independent of proliferation or differentiation, yet correlates with tissue biology. Cancer Res. 61, 4990–4993. [PubMed]
  • Mosbech C.H., Rechnitzer C., Brok J.S., Rajpert-De Meyts E., Hoei-Hansen C.E., 2014. Recent advances in understanding the etiology and pathogenesis of pediatric germ cell tumors. J. Pediatr. Hematol. Oncol. 36, 263–270. [PubMed]
  • Oosterhuis J.W., Stoop H., Honecker F., Looijenga L.H.J., 2007. Why human extragonadal germ cell tumours occur in the midline of the body: old concepts, new perspectives. Int. J. Androl. 30, 256–263. discussion 263–4 [PubMed]
  • Rajpert-De Meyts E., Bartkova J., Samson M., Hoei-Hansen C.E., Frydelund-Larsen L., Bartek J., Skakkebaek N.E., 2003. The emerging phenotype of the testicular carcinoma in situ germ cell. APMIS 111, 267–278. [PubMed]
  • Rosenblum M.K., Matsutani M., VanMeir E.G., K. P., C. W., 2000. Pathology and Genetics of Tumours of the Nervous System, Chapter: CNS Germ Cell Tumours IARC Press (International Agency for Research on Cancer) 208–214.
  • Schneider D.T., Schuster A.E., Fritsch M.K., Pediatric N., Cell G., Hu J., Olson T., Lauer S., Go U., 2001. Multipoint imprinting analysis indicates a common precursor cell for gonadal and nongonadal pediatric germ cell tumors. Cancer Res. 7268–7276. [PubMed]
  • Scotting P.J., 2006. Are cranial germ cell tumours really tumours of germ cells?. Neuropathol. Appl. Neurobiol. 32, 569–574. [PubMed]
  • Sherr C.J., 2012. Ink4-Arf locus in cancer and aging. Wiley Interdiscip. Rev. Dev. Biol. 1, 731–741. [PubMed]
  • Teilum G., 1965. Classification of endodermal sinus tumour (mesoblatoma vitellinum) and so-called “embryonal carcinoma” of the ovary. Acta Pathol. Microbiol. Scand. 64, 407–429. [PubMed]
  • Terashima K., Yu A., Chow W.T., Hsu W.J., Chen P., Wong S., Hung Y.S., Suzuki T., Nishikawa R., Matsutani M., Nakamura H., Ng H., Allen J.C., Aldape K.D., Su J.M., Adesina A.M., Leung H.E., Man T., Lau C.C., 2014. Genome-wide analysis of DNA copy number alterations and loss of heterozygosity in intracranial germ cell tumors. Pediatr. Blood Cancer 593–600. [PubMed]
  • Tort F., Bartkova J., Sehested M., Orntoft T., Lukas J., Bartek J., 2006. Retinoblastoma pathway defects show differential ability to activate the constitutive DNA damage response in human tumorigenesis. Cancer Res. 66, 10258–10263. [PubMed]
  • Velimezi G., Liontos M., Vougas K., Roumeliotis T., Bartkova J., Sideridou M., Dereli-Oz A., Kocylowski M., Pateras I.S., Evangelou K., Kotsinas A., Orsolic I., Bursac S., Cokaric-Brdovcak M., Zoumpourlis V., Kletsas D., Papafotiou G., Klinakis A., Volarevic S., Gu W., Bartek J., Halazonetis T.D., Gorgoulis V.G., 2013. Functional interplay between the DNA-damage-response kinase ATM and ARF tumour suppressor protein in human cancer. Nat. Cell Biol. 15, 967–977. [PubMed]
  • Yamada Y., Coffman C.R., 2005. DNA damage-induced programmed cell death: potential roles in germ cell development. Ann. N. Y. Acad. Sci. 1049, 9–16. [PubMed]

Articles from Molecular Oncology are provided here courtesy of Wiley-Blackwell