PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of jcinvestThe Journal of Clinical Investigation
 
J Clin Invest. 2010 October 1; 120(10): 3594–3605.
Published online 2010 September 1. doi:  10.1172/JCI41939
PMCID: PMC2947220

Cytoplasmic p21 expression levels determine cisplatin resistance in human testicular cancer

Abstract

Platinum-based chemotherapies such as cisplatin are used as first-line treatment for many cancers. Although there is often a high initial responsiveness, the majority of patients eventually relapse with platinum-resistant disease. For example, a subset of testicular cancer patients still die even though testicular cancer is considered a paradigm of cisplatin-sensitive solid tumors, but the mechanisms of chemoresistance remain elusive. Here, we have shown that one key determinant of cisplatin-resistance in testicular embryonal carcinoma (EC) is high cytoplasmic expression of the cyclin-dependent kinase (CDK) inhibitor p21. The EC component of the majority of refractory testicular cancer patients exhibited high cytoplasmic p21 expression, which protected EC cell lines against cisplatin-induced apoptosis via CDK2 inhibition. Localization of p21 in the cytoplasm was critical for cisplatin resistance, since relocalization of p21 to the nucleus by Akt inhibition sensitized EC cell lines to cisplatin. We also demonstrated in EC cell lines and human tumor tissue that high cytoplasmic p21 expression and cisplatin resistance of EC were inversely associated with the expression of Oct4 and miR-106b seed family members. Thus, targeting cytoplasmic p21, including by modulation of the Oct4/miR-106b/p21 pathway, may offer new strategies for the treatment of chemoresistant testicular and other types of cancer.

Introduction

In testicular cancer (TC), even in cases of extensive metastatic disease, an enormous decrease in mortality has been observed with the introduction and use of highly effective cisplatin-containing chemotherapy schemes (1, 2). Therefore, TC is considered the paradigm for curative disease. Despite the overall treatment success, about 20%–50% of the TC patients with extensive metastatic spread belonging to the intermediate or poor risk group will not achieve a durable complete remission after initial treatment and will eventually die from this disease (3). The molecular basis for resistance, however, remains obscure.

A major role in the response to chemotherapeutic drugs and the execution of apoptosis has been ascribed to wild-type p53 (4). In only a small proportion of TCs, mutations in p53 affect the downstream apoptotic pathway and lead to resistance (5), in particular, as wild-type p53 is expressed at high levels in the majority of TCs (69). p53 is a tumor suppressor protein with a dual role in stress response by transactivation of genes that induce apoptosis, such as FAS, as well as genes that induce cell-cycle arrest, such as CDKN1A (encoding p21cip1/waf1). Interestingly, the cyclin-dependent kinase (CDK) inhibitor p21cip1/waf1 (p21) has been shown to inhibit apoptosis (1013). Remarkably, several studies have demonstrated that p21 protein and CDKN1A mRNA expression are abundantly expressed in more differentiated TCs, such as mature teratoma (14, 15). Worthy of note, these teratomas are resistant to cisplatin-based chemotherapy (8, 1517). In contrast, p21 is almost not detectable in seminomas and embryonal carcinomas (EC) that are predominantly sensitive to cisplatin (9, 14, 18).

In human TC/EC cell lines, similar to patients, cisplatin proved to be an extremely cytotoxic drug, inducing massive apoptosis (1922). Cisplatin treatment of EC cells resulted in enhanced levels of p53 and Mdm2, activation of the Fas apoptotic pathway, and induction of apoptosis, while the expression levels of p21 were almost not affected (7, 20). In contrast, γ irradiation induced p53 and Mdm2 levels and a massive induction of cytoplasmic p21 without inducing apoptosis or cell-cycle arrest in EC cells (7). These results suggested an important role for cytoplasmic p21 in preventing DNA damage–induced apoptosis in EC cells.

Results

High cytoplasmic p21 expression is associated with cisplatin resistance in EC cells.

A panel of cisplatin-sensitive and -resistant EC cell lines was used in this study to compare cisplatin responses (Figure (Figure1A1A and Supplemental Figure 1A; supplemental material available online with this article; doi: 10.1172/JCI41939DS1) with the levels of (induced) p21 (Figure (Figure1,1, B and C). Resistance to cisplatin was predominantly associated with high p21 expression levels, since the intrinsically resistant EC cell lines Scha and 2102EP showed higher basal and cisplatin-induced p21 and CDKN1A levels compared with the cisplatin-sensitive EC cell lines Tera and 833KE (Figure (Figure1,1, A–C). Low levels of p21 were not associated with cisplatin response in Tera-CP, a subline with acquired cisplatin resistance due to p21 unrelated mechanisms (23).

Figure 1
High cytoplasmic p21 expression is associated with cisplatin resistance in EC cell lines.

The subcellular localization of p21 was investigated with immunofluorescence microscopy (IF) and Western blot analysis (WB) using fractionated protein samples. We stained the EC cell lines Tera, Tera-CP, 833KE, Scha, and 2102EP for basal levels of p21 and p21 levels after treatment with cisplatin (or γ irradiation). All untreated EC cell lines showed a predominant cytoplasmic localization of p21, albeit with large differences in p21 intensity (Figure (Figure1,1, D and E). Cytoplasmic localization was not affected by cisplatin treatment (Figure (Figure1,1, D and E, and Supplemental Figure 1B). γ irradiation was used as positive control for p21 induction in EC cells (7). Despite almost a similar p53 increase, γ irradiation, in contrast with cisplatin treatment, strongly induced accumulation of p21 in the cytoplasm in both cisplatin-resistant Scha and cisplatin-sensitive Tera cells (ref. 7, Figure Figure1E,1E, and Supplemental Figure 1C). We next hypothesized that high levels of cytoplasmic p21 have a causal role in cisplatin resistance in EC cell lines.

Cytoplasmic p21 protects EC cells against cisplatin-induced apoptosis.

To investigate the importance of high levels of cytoplasmic p21, both cisplatin-sensitive and cisplatin-resistant EC cell lines were depleted of p21, using siRNA. Downregulation of p21 led to an increase in apoptosis upon cisplatin treatment in the intrinsically resistant cell lines Scha and 2102EP as compared with cells transfected with scrambled siRNA (Figure (Figure2A).2A). The loss of p21 in Scha and 2102EP resulted in enhanced PARP cleavage and higher caspase 3 activation compared with control cells following cisplatin treatment (Figure (Figure2,2, B and C). In the cisplatin-sensitive 833KE and Tera cells and the acquired cisplatin-resistance Tera-CP cells, all expressing low endogenous cytoplasmic p21 levels, no significant effect of p21 downregulation on apoptosis levels was observed (Supplemental Figure 2, A–C).

Figure 2
Increased cisplatin sensitivity after p21 downregulation in intrinsically resistant EC cell lines.

Next, we stably transduced the cisplatin-sensitive cell line Tera with a retroviral construct containing both p21-ΔNLS and GFP, or GFP only (Tera-mock) to study more extensively whether enhanced levels of cytoplasmic localized p21 are involved in suppressing apoptosis. The p21-ΔNLS protein lacks its bipartite nuclear localization signal and is therefore maintained in the cytoplasm (Supplemental Figure 2D). Treatment with cisplatin resulted in less apoptosis and reduced caspase 3 activity in Tera-p21-ΔNLS compared with Tera-mock (Supplemental Figure 2, E and F). We downregulated p21-ΔNLS in Tera-p21-ΔNLS cells using p21 siRNA, and we showed that these cells became as sensitive as Tera-mock cells to cisplatin. No effect of p21 siRNA on cisplatin-induced apoptosis was observed in Tera-mock cells (Figure (Figure3,3, A and B, and Supplemental Figure 2, G and H).

Figure 3
Overexpression of cytoplasmic p21 protects against cisplatin-induced apoptosis via complex formation of p21 with CDK2 and ASK1.

Cytoplasmic p21 forms a complex with proapoptotic CDK2 in cisplatin-resistant EC cells.

It has been described that caspase 3–mediated cleavage of p21 activates the proapoptotic role of CDK2 (2426). Therefore, we analyzed the presence of CDK2 in p21 IP. In Tera, Tera-p21-ΔNLS, and Scha, CDK2 coprecipitated with p21, indicating that CDK2 is in complex with cytoplasmic p21 in untreated cells and in cells treated with cisplatin or γ irradiation (Figure (Figure3C).3C). In addition, we stained for the presence of apoptosis signaling kinase 1 (ASK1), since cytoplasmic p21 can bind to ASK1, thereby inhibiting ASK1-mediated apoptosis via the intrinsic mitochondrial apoptotic pathway (1013). In fact, more CDK2 and ASK1 was coprecipitated with p21 in untreated and cisplatin-treated Scha and Tera-p21-ΔNLS, which have higher p21 levels compared with Tera. In irradiated Tera and Scha expressing equal p21 levels, high levels of p21 were precipitated, leading to coprecipitation of almost similar levels of CDK2 and ASK1 (Figure (Figure3C).3C). We further established the important proapoptotic role of CDK2 by downregulating CDK2 with siRNA in cisplatin-sensitive cell lines, which led to a decrease in cisplatin-induced apoptosis and PARP cleavage compared with scrambled siRNA in Tera and 833KE (Figure (Figure3D3D and Supplemental Figure 3A). Next, we investigated the proapoptotic role of CDK2 in the high p21-expressing 2102EP and Scha. Therefore, we sensitized these cells to cisplatin by suppression of p21 and, additionally, transfected these cells with CDK2 siRNA. The observed increase in apoptosis in p21-suppressed 2102EP and Scha cells upon cisplatin treatment could be dramatically reduced by cotransfection with CDK2 siRNA (Supplemental Figure 3B).

Interfering in p-Akt–mediated cytoplasmic localization of p21 sensitizes cisplatin-resistant EC cells.

The cytoplasmic localization of p21 can be caused by phosphorylation of Thr145 located in the NLS of p21 (11, 27). Upon phosphorylation of the NLS domain by p-Akt, phosphorylated p21 (p-p21) becomes more stable and dynamically shuttles between the nucleus and the cytoplasm (11, 27, 28). In untreated and cisplatin-treated EC cells, p-p21 was detectable with WB (Supplemental Figure 4A). IF indicated that p-p21 was also localized in the cytoplasm (Figure (Figure4A).4A). In order to find out whether phosphorylation of Thr145 could occur via p-Akt, we stained the EC cells for p-Akt and its natural inhibitor phosphatase and tensin homolog (PTEN). In Tera, Scha, and 2102EP, p-Akt is present (Figure (Figure4B4B and Supplemental Figure 4B), whereas PTEN is not expressed in these cells (data not shown) (29).

Figure 4
Dephosphorylation of p-p21, nuclear localization of p21, and sensitization for cisplatin after combined treatment with LY294002 or triciribine.

Treatment with the PI3K inhibitor LY294002 or with the specific Akt inhibitor triciribine (currently in phase I trial) resulted in dephosphorylation of Akt in Tera, Scha, and 2102EP and concomitant dephosphorylation of p21 (Thr145), which is consistent with a more pronounced nuclear localization of p21 (Figure (Figure4,4, A and B, and Supplemental Figure 4, B–E). In the cisplatin-sensitive cell line Tera, either LY294002 or triciribine did not affect cisplatin sensitivity (Supplemental Figure 5A), though shuttling of the weakly detectable cytoplasmic p21 to the nucleus was observed (Supplemental Figure 4, C and D). Interestingly, a strong sensitization to cisplatin was observed in Scha and 2102EP with the combined treatment for 24 hours with cisplatin and either LY294002 or triciribine (Figure (Figure4C4C and Supplemental Figure 5A). Shuttling of p21 to the nucleus following LY294002 treatment resulted in a loss of CDK2 in complex with p21 in Scha cells (Supplemental Figure 5B). Release of CDK2 was instrumental in the enhanced induction of apoptosis, since the sensitization to cisplatin combined with LY294002 in 2102EP and Scha cells was almost completely blocked following suppression of CDK2 (Supplemental Figure 5C). Downregulation of Akt using siRNA targeting Akt1–3 reduced the overall levels of p-p21 and resulted in a more pronounced nuclear localization of p21 (Supplemental Figure 5D). In addition, Akt1–3 suppression sensitized 2102EP and Scha for cisplatin-induced apoptosis (Figure (Figure4D). 4D).

Enhanced cytoplasmic p21 positivity and low levels of apoptosis were found in Tera and Scha 24 hours after γ irradiation (Figure (Figure1E1E and Supplemental Figure 4D). The combined treatment of γ irradiation with either LY294002 or triciribine for 24 hours resulted in less p-p21, predominantly nuclear localization of p21, and augmented apoptosis in both Tera and Scha (Supplemental Figure 4, B and D, and Supplemental Figure 5E). In Tera-p21-ΔNLS, lacking the phosphorylation site Thr145, treatment with the PI3K/Akt inhibitors, however, had no effect on the cytoplasmic localization of p21-ΔNLS (Supplemental Figure 4C) and treatment with the inhibitors did not lead to sensitization to cisplatin (data not shown). These results show that the sensitizing effect of PI3K/Akt inhibitors on cisplatin-induced apoptosis is mediated via dynamic changes in p21 localization.

Cytoplasmic expression of p21 and p-p21 in the EC component of chemorefractory TC patients.

Next, we investigated the p21 and p-p21 (Thr145) expression patterns in EC containing TC tissue from 23 patients with metastatic disease who were cured and considered sensitive to cisplatin-containing chemotherapy and from 7 patients with metastatic disease not cured and considered refractory to cisplatin-containing chemotherapy. Histological analysis after H&E staining was used to distinguish the various TC components. In addition, Oct4 (also known as POU5F1 and Oct3/4) immunohistochemistry (IHC) was used as marker for embryonal pluripotency (16). Mature teratoma showed an intense p21 and p-p21 staining, whereas no staining of Oct4 was noticed (Supplemental Figure 6A). EC components from TC patients, sensitive to cisplatin-containing chemotherapy, were negative for p21 in 23 (100%) and negative for p-p21 in 21 (91%) of the patients (Figure (Figure5A5A and Supplemental Figure 6A). In contrast, the EC component of refractory TC patients stained positive for p21/p-p21 in 6 out of 7 (86%) patients, including 2 patients with EC as the only component (Figure (Figure5B).5B). Positive staining for p21/p-p21 in the EC component was significantly more often observed in refractory TC patients compared with TC patients sensitive to cisplatin-containing chemotherapy (P < 0.0003, Fisher’s exact test). Moreover, IF showed that the localization of p21/p-p21 in the EC component of refractory patients was mainly cytoplasmic, whereas the localization of p21/p-p21 in teratomas is more nuclear (Figure (Figure5C).5C). In addition, presence of Ki-67 staining indicated proliferation instead of cell-cycle arrest in abundantly p21/p-p21–positive EC (Supplemental Figure 6B). Taken together, our results show that cytoplasmic p21 and p-p21 positivity of EC components in TC patients is associated with the response (poor outcome) to cisplatin-containing chemotherapy.

Figure 5
Cytoplasmic p21 expression in ECs of nonresponding TC patients.

Reduced levels of Oct4 and miR-106b family members cause high cytoplasmic p21 expression in cisplatin-resistant EC cells.

Involvement of miRNAs belonging to the miR-106b seed family in regulating p21/CDKN1A and cell-cycle control has previously been demonstrated (30, 31). Therefore, we examined the expression levels of the miR-106b seed family in the EC cell lines. The intrinsically resistant Scha and 2102EP cell lines, with high p21 levels, showed lower levels of the miR-106b seed family compared with the cisplatin-sensitive Tera and 833KE cell lines (Figure (Figure6A6A and Supplemental Figure 6C). Subsequently, we used synthetic anti-miRNA to specifically inhibit the miR-106b seed family and to demonstrate its relation with localization and expression level of p21 in EC cell lines. In Tera cells, transfection with synthetic anti-miRNA against miR-17-5p, miR-20a, miR-93, miR-106a, or miR-106b resulted in a massive increase in cytoplasmic p21 levels compared with control (Figure (Figure6,6, B and D). Although treatment with synthetic anti–miR-17-5p resulted in enhanced p21 levels in Tera, Tera-CP, Scha, and 2102EP, the strongest enhancement of p21 was found in Tera and Tera-CP (Figure (Figure6C),6C), which may at least in part be explained by the higher miR-17-5p levels in these cell lines (Figure (Figure6A6A and ref. 32). The strongly enhanced cytoplasmic p21 levels in Tera following synthetic anti–miR-17-5p treatment completely blocked cisplatin-induced apoptosis (Figure (Figure6,6, D and E). Suppression of cytoplasmic p21 with siRNA or shuttling of p21 to the nucleus with LY294002 resensitized synthetic anti–miR-17-5p–treated Tera cells to cisplatin (Figure (Figure6F6F and Supplemental Figure 6D). Overexpression of pre–miR-17-5p resulted in lower p21 levels and an increase in cisplatin-induced apoptosis in Scha and 2102EP cells (Figure (Figure6G),6G), demonstrating the important causal relation among the presence of miR-17-5p, reduced p21 expression, and high cisplatin sensitivity.

Figure 6
Positivity for p21 and p-p21 protein is inversely associated with the expression of the miR-106b seed family in ECs.

We performed in situ hybridization (ISH) for miR-106a and miR-17-5p on paraffin-embedded formalin-fixed tumor tissue from patients with multicomponent (primary) TC to define in vivo the relation between p21/p-p21 expression and the presence of miR-106b seed family members in the EC component. A strong positive staining for miR-17-5p, miR-106a, and Oct4 but not for p21/p-p21 was found in the EC components of chemosensitive patients. In contrast, mature teratomas showed no staining of miR-17-5p, miR-106a, and Oct4 while staining intensively positive for p21/p-p21 (Figure (Figure5A5A and Supplemental Figure 6A). On the contrary, in chemorefractory TC patients, almost no miR-17-5p staining was observed in EC components, while p21/p-p21 was clearly detectable. Moreover, the staining intensity of Oct4 in the EC component was less pronounced in refractory patients compared with chemotherapy-sensitive patients (Figure (Figure5B).5B). Additionally, the expression levels of both Oct4 and miR-106b seed family members were lower in the intrinsically resistant cell lines Scha and 2102EP compared with the cisplatin-sensitive cell line Tera (Figure (Figure6A).6A). These results suggest a relation between the reduced expression of Oct4 and the miR-106b seed family member in refractory ECs. In murine embryonic stem cells, Oct4 has been implicated in controlling expression of several miRNA families, including miR-17~92 and miR-106a~363 cluster of the miR-106b seed family (33). We downregulated Oct4 in Tera cells with siRNA in order to establish the direct relation among Oct4, miR-106b, and p21 expression in ECs. Suppression of Oct4 lowered the expression level of miR-106a, miR-17-5p, and miR-20a compared with control Tera cells (Figure (Figure7A7A and Supplemental Figure 6E) and resulted in enhanced p21-3′ UTR dependent luciferase translation (Figure (Figure7B)7B) and robust induction of p21 (Figure (Figure7C7C and Supplemental Figure 6F). A similar enhancement of p21-3′ UTR–dependent luciferase translation was observed following transfection of Tera cells with synthetic anti–miR-17-5p (Figure (Figure7B).7B). Furthermore, Oct4 suppression in Tera, Scha, and 2102EP resulted in an upregulation of p21 and a concomitant resistance to cisplatin, which could be reverted by p21 downregulation (Figure (Figure7,7, C and D). Sustained miR-17-5p expression using pre–miR-17-5p precluded the induction of p21 protein even in the context of Oct4 knockdown in Tera and 2102EP cells (Figure (Figure7E).7E). Some p21 expression, however, was still observed in 2102EP cells, albeit at a lower level compared with control 2102EP cells, indicating that not all effects of Oct4 on p21 expression are mediated by the miR-106b seed family. Overall, these results reveal an important role of Oct4 in regulating p21 expression via the miR-106b seed family in ECs.

Figure 7
Oct4 regulates the expression level of the miR-106b seed family in ECs.

Discussion

The functionality of many proteins is thought to be related to their intracellular localization. Nuclear localized p21 is known to control the cell cycle and DNA replication, whereas cytoplasmic p21 has been implicated in the inhibition of apoptosis. Here, we have demonstrated that high levels of cytoplasmic localized p21 protect EC cells against cisplatin-induced apoptosis. Furthermore, we dissected the pathway involved in the regulation of p21 expression levels, i.e., OCT4 and the miR-106b family, and identified the key protein causing the cytoplasmic localization of p21. Our results indicate that p-Akt–mediated p21 phosphorylation is essential for p21 localization in the cytoplasm. Inhibition of p-Akt retained p21 in the nucleus, resulting in less p21 complex formation with CDK2 and sensitized EC cells to cisplatin-induced apoptosis. In cisplatin-sensitive EC cell lines and EC components of chemosensitive TC patients, p21 is not commonly expressed, which is related to the expression levels of miR-106b seed family members and Oct4. On the contrary, in cisplatin-resistant EC cell lines and EC components of refractory TC patients, indicating clinical cisplatin resistance, cytoplasmic p21 expression was clearly detectable and related to reduced Oct4 and miR-106b expression. Therefore, we conclude that high levels of cytoplasmic localized p21 are important determinants of resistance to cisplatin-based chemotherapy in ECs (Figure (Figure8). 8).

Figure 8
Proposed simplified model showing the mechanisms through which cytoplasmic p21 can inhibit cisplatin-induced apoptosis in EC cells.

We recently demonstrated that cisplatin induces apoptosis mainly in a Fas-dependent manner in EC cells (20), while enhanced p21 expression protects EC cells against Fas-mediated apoptosis (7). This protection might be mediated via the inhibition of caspase 3 (34), CDK2, and ASK1 by cytoplasmic p21. We have no indication that p21 binds to caspase 3 and inhibits activation in irradiated EC cells (7) or cisplatin-treated EC cells (data not shown). In the present paper, we show that endogenous cytoplasmic p21 forms a complex with CDK2 and ASK1 in EC cells. In EC cells expressing low p21 levels, CDK2 downregulation strongly reduced the cisplatin-induced apoptotic response. Moreover, relocalization of p21 to the nucleus using LY294002, which completely abolished p21-CDK2 complexes, as well as suppression of p21 made initially cisplatin-resistant cells sensitive to cisplatin in a CDK2-dependent manner, indicating the importance of cytoplasmic CDK2 as a proapoptotic factor in cisplatin-treated EC cells. The proapoptotic signaling of CDK2 is not clear yet. It has been reported that cytoplasmic CDK2 activity is induced in a caspase-dependent way following activation of the Fas pathway (35) and leads to depolarization of the mitochondrial membrane potential (36) by cytoplasmic CDK2-mediated translocation of Bax to the mitochondria (36, 37). Alternatively, functional interaction of CDK2 with FOXO1 after DNA damage can be an important mechanism to enhance, among others, Fas ligand expression (38). Taken together, these results indicate that the high cytoplasmic p21 levels, in complex with cytoplasmic CDK2, protect against cisplatin-induced apoptosis in EC cells.

In the present study, in vivo relevance of p21 expression was demonstrated in the EC component of TC patients with refractory disease, being positive for p21, whereas no p21 was detected in the EC component of patients with chemosensitive tumors. Lack of p21 staining in the EC component but strongly positive p21 staining in mature teratoma components has been frequently observed (9, 14, 15, 18). Teratomas, however, predominately show a nuclear localization of p21. Moreover, teratomas in contrast with ECs express Rb (17, 39), suggesting a putative role for p21 in cell-cycle control in these nongrowing or rarely growing tumors, also known as growing teratoma syndrome (17). Furthermore, teratomas are, possibly via p21-mediated cell-cycle arrest, not responsive to treatment with cisplatin (8, 1517). Interestingly, p21 staining in the EC component of refractory patients was mainly localized in the cytoplasm of EC cells. In addition, Ki-67 staining revealed no G1 arrest, but proliferation in ECs with abundant expression of p21, indicating a similar role in vivo for high levels of cytoplasmic p21 in the protection against cisplatin-induced apoptosis, as we have demonstrated for cisplatin-resistant EC cell lines. Cytoplasmic p21 expression has been related to drug resistance in several other types of cancer (11). Previous studies in non-TC cell types have shown that phosphorylation of Thr145 in the p21-NLS by p-Akt is important for dynamic shuttling of p-p21 between the nucleus and the cytoplasm (11, 27). PTEN can prevent phosphorylation of Akt in EC cells (29). Virtually all ECs (cell lines and TC patients), however, lack PTEN expression (29), which is shown to be important for the development of invasive TCs (29, 40). Here, we show that relocalization of p21 to the nucleus by dephosphorylation of p21, using the PI3K inhibitor LY294002, the Akt inhibitor triciribine, or siRNA against Akt, sensitized cells for cisplatin-induced apoptosis. Phosphorylation of Akt might be inhibited by targeting tyrosine kinase membrane receptors such as KIT and ERBB2, that have been implicated in TC (8, 4146). An even more direct approach to targeting Akt phosphorylation, besides triciribine, is the use of PI3K/Akt inhibitors that are currently in clinical development (47). However, the efficacy of the various subclasses of PI3K or Akt inhibitors to relocalize p21 and subsequently enhance cisplatin sensitivity in refractory TC patients needs to be further elucidated.

The role of miRNAs in tumorigenesis is attracting increasing attention, and miRNAs have actually been implicated in TC/EC carcinogenesis (48, 49). The miR-106b seed family is potentially oncogenic (50) and has been implicated in embryonal stem cells in their maintenance and control of differentiation (51). ECs share many similarities with embryonic stem cells (16, 32) including their miRNA expression profile (32). Interestingly, our results indicate differences in expression levels of p21-3′ UTR regulating miR-106b seed family members between intrinsically cisplatin-resistant EC and cisplatin-sensitive EC cell lines. Additionally, we report that high levels of the miR-106b seed family members are causing low levels of (cytoplasmic) p21 in EC cells and consequently sensitivity to cisplatin. In contrast, low levels of the miR-106b seed family members were associated with high levels of (cytoplasmic) p21 in EC cells and ECs of refractory TC patients. In murine embryonic stem cells, miR-17~92 and miR-106a~363 clusters of the miR-106b seed family are under control of embryonic stem cell factor Oct4 (33). ECs, like embryonic stem cells, also express a cassette of pluripotency genes (16, 32, 52), where Oct4 is supposed to be the key pluripotency regulator (52). Here, we show that expression of members of the miR-106b seed family colocalizes with Oct4 expression in TC. Our results indicate that Oct4 regulates expression of miR-17-5p, miR-106a, and miR-20a, belonging to the miR-106b seed family, via the p21-3′ UTR. Furthermore, in EC cells expressing high basal levels of Oct4, overexpression of miR-17-5p completely precluded p21 induction when Oct4 was suppressed. Similar results were obtained with EC cells expressing lower basal levels of Oct4; however, some p21 expression was still detectable in these cells after combined transfection of Oct4 siRNA and pre–miR-17-5p. This suggests that besides the miR-106b seed family, the regulation of p21 expression through Oct4 may be more complex, involving other factors as well. In summary, we provide evidence that Oct4 regulates expression of p21-3′ UTR targeting miR-106b seed family members and thus indirectly p21 levels and cisplatin sensitivity in EC cells.

In conclusion, we demonstrate an important mechanism in ECs that causes elevated cytoplasmic p21 levels and consequently results in cisplatin resistance. Moreover, we provide a target to overcome resistance, which may form the basis for the development of novel therapeutic approaches to alleviate cisplatin resistance in TCs or other cisplatin-resistant types of solid tumors.

Methods

Cell lines and reagents.

A well-defined panel of cisplatin-sensitive and cisplatin-resistant human EC cell lines, all expressing wild-type p53, were used (7, 20, 22, 23, 53). The 2102EP cell line was obtained from L. Looijenga (Department of Pathology, Erasmus Medical Center, Rotterdam, the Netherlands). Tera, Tera-CP, 2102EP, Scha and 833KE, and the human breast carcinoma cell line MCF-7 (used as a control) were cultured in RPMI 1640 medium (Gibco; Invitrogen) supplemented with 10% FCS (SanBio) at 37°C in a humidified atmosphere with 5% CO2. Cisplatin was purchased from Bristol-Myers Co., LY294002 from Cell Signaling, and Triciribine from Biomol.

Drug sensitivity assay and apoptosis.

Drug sensitivity testing was performed as described previously (20, 53). Acridine orange fluorescent staining of nuclei was used to distinguish apoptotic from vital cells as described previously (7, 20, 54).

WB and preparation of fractionated proteins.

Lysates were examined by WB as described previously (7, 20). Nuclear and cytosolic protein fractions were prepared as described previously (7). Antibodies used are listed in Supplemental Methods.

Immunofluorescence.

Cells were fixed with 4% paraformaldehyde and blocked with 0.1% Triton X-100, 1% BSA, and 1% normal goat serum (NGS) in PBS. This was followed by immunostaining with the corresponding antibodies and counterstaining with Alexa Fluor goat secondary antibodies (Molecular Probes; Invitrogen). Finally, cells were stained with Hoechst 33258 (Molecular Probes; Invitrogen) and mounted with Vectashield (Vector Laboratories). The Quantimet 600S system (Leica Microsystems) was used for analysis. The data were exported as bmp files and processed using Corel Photo-Paint X4 (Corel Corp.).

Caspase 3 activity.

Caspase 3 was assayed according to the manufacturer’s instructions using DEVD-MCA (Zebra Bioscience BV). Fluorescence was monitored in an FL600 Fluorimeter Bio-Tek plate reader (Beun de Ronde).

Tera-p21-Δ-NLS stably transduced cell line.

pMSCV-p21-ΔNLS, pMSCV-IGFP, and pCL-Ampho were provided by J.J. Schuringa and H. Schepers (Hematology Research, University Medical Center Groningen). For viral production, 293T cells were transfected with pCL-Ampho and MSCV-IGFP either with or without p21-ΔNLS. GFP-positive Tera cells were sorted on a fluorescent-activated cell sorter (MoFlow; Cytomation).

RNA interference, miRNA antisense, and transfections.

Specific siRNA and negative control (scrambled) were purchased from Eurogentec. Synthetic anti-miRNA was purchased from IDT and pre–miR-17-5p from Ambion. All sequences and methods are listed in Supplemental Methods. For the luciferase reporter assay, Tera cells were transfected with psiCHECK2-p21-3′ UTR using Fugene, lysates were made, and assay was performed as described previously (55).

IP.

Cells (107) were harvested, washed with ice-cold PBS, and lysed in 500 μl lysis buffer (20 mM Tris HCl, pH 7.6, 150 mM NaCl, 0.2% NP-40, protease inhibitor COMPLETE, 1 mM PMSF, 1 mM NaF, and 1 mM DTT). Lysates were clarified, and protein concentration was equalized with Bradford and incubated for 16 hours with a mixture of agarose-conjugated anti-p21 (F5 and C19; Santa Cruz Biotechnology Inc.). Immunocomplexes were washed 5 times and eluted with 0.5 M glycine/HCl pH 2.4, mixed 1:1 with standard 2× WB sample buffer, and examined by WB analysis as described above.

Patients, IHC, and miRNA ISH.

General patient characteristics and histopathology are summarized in Supplemental Table 1. The TC patients with chemosensitive disease have been reported before (56). From the patients with TC diagnosed and treated at our institution between 1985 and 2007 with refractory disease (defined as patients not achieving a response on initial treatment or renewed elevation of tumor marker levels within 4 weeks after completion of chemotherapy), patients with sufficient EC component in their tumor samples (n = 7) were selected. TC specimens were used to represent all histological subtypes of the primary tumor (YS, yolk sac tumor; ChC, choriocarcinoma; T, mature teratoma/immature teratoma; S, seminoma) and each of the 3 different prognosis groups according to the IGCCCG classification (57). The studies were approved by the medical ethical committee of University Medical Center Groningen, and all patients gave informed consent prior to study entry. For each patient, representative paraffin-embedded tumor material was collected and serial 3-μm sections were cut.

For IHC, sections were deparaffinized in xylene and rehydrated in alcohol; antigen retrieval was performed followed by blocking of endogenous peroxidase (30 min, 3% H2O2). Subsequently, slides were incubated for 1 hour with the primary antibodies Oct4 (C20; Santa Cruz Biotechnology Inc.), p21 (EA10, Oncogene), p-p21 (Thr145; Santa Cruz Biotechnology Inc.), and Ki-67 (MIB-1; Dako) and counterstained with HRP-conjugated secondary antibodies (Dako). DAB was used as chromagen to visualize peroxidase activity. Counterstaining was performed with hematoxylin. Immunoglobulin class-matched control sera were used as negative controls. Normal colon and normal skin served as a positive control for p21.

miRNA ISH was performed as reported previously (58). For both ISH and IHC, slides were scanned using ScanScope CS System (Aperio) and pictures were taken with the ImageScope software package (Aperio).

RNA isolation and TaqMan miRNA quantitative PCR.

Total RNA was isolated and miRNA-106b family expression was measured as previously described (55). The miRNA expression was normalized to RNU48 expression, resulting in a ΔCt from which the 2–ΔCt value was derived and depicted.

Statistics.

Results of at least 3 experiments are expressed as mean ± SD. Student’s unpaired t test was used to compare values of test and control samples. Differences were considered significant at P < 0.05. Fisher’s exact χ2 test was used for categorical patient variables. Differences were considered significant with 2-sided test at P < 0.05.

Supplementary Material

Supplemental data:

Acknowledgments

The authors wish to thank Wytske Boersma-van Eck for help with IHC staining; Lu Ping Tan for ISH; Theo Plantinga and Marieke Smit for technical support; and Hein Schepers for his technical advice with respect to transducing Tera cells with pMSCV-p21-ΔNLS; Esther de Haas for help with collecting phenotypic data of testicular cancer patients from the UMCG database; and Jan Jacob Schuringa and Marcel van Vugt for critical reading of the manuscript.

Footnotes

Conflict of interest: The authors have declared that no conflict of interest exists.

Citation for this article: J Clin Invest. 2010;120(10):3594–3605. doi:10.1172/JCI41939.

References

1. Kondagunta GV, et al. Relapse-free and overall survival in patients with pathologic stage II nonseminomatous germ cell cancer treated with etoposide and cisplatin adjuvant chemotherapy. J Clin Oncol. 2004;22(3):464–467. doi: 10.1200/JCO.2004.07.178. [PubMed] [Cross Ref]
2. Einhorn LH. Role of the urologist in metastatic testicular cancer. J Clin Oncol. 2007;25(9):1024–1025. doi: 10.1200/JCO.2006.07.6075. [PubMed] [Cross Ref]
3. Horwich A, Shipley J, Huddart R. Testicular germ-cell cancer. Lancet. 2006;367(9512):754–765. doi: 10.1016/S0140-6736(06)68305-0. [PubMed] [Cross Ref]
4. Johnstone RW, Ruefli AA, Lowe SW. Apoptosis: a link between cancer genetics and chemotherapy. Cell. 2002;108(2):153–164. doi: 10.1016/S0092-8674(02)00625-6. [PubMed] [Cross Ref]
5. Houldsworth J, et al. Human male germ cell tumor resistance to cisplatin is linked to TP53 gene mutation. Oncogene. 1998;16(18):2345–2349. doi: 10.1038/sj.onc.1201770. [PubMed] [Cross Ref]
6. Heidenreich A, et al. Immunohistochemical and mutational analysis of the p53 tumour suppressor gene and the bcl–2 oncogene in primary testicular germ cell tumours. APMIS. 1998;106(1):90–99. doi: 10.1111/j.1699-0463.1998.tb01324.x. [PubMed] [Cross Ref]
7. Spierings DC, de Vries EG, Stel AJ, te Rietstap N, Vellenga E, de Jong S. Low p21Waf1/Cip1 protein level sensitizes testicular germ cell tumor cells to Fas-mediated apoptosis. Oncogene. 2004;23(28):4862–4872. doi: 10.1038/sj.onc.1207617. [PubMed] [Cross Ref]
8. Houldsworth J, Korkola JE, Bosl GJ, Chaganti RS. Biology and genetics of adult male germ cell tumors. J Clin Oncol. 2006;24(35):5512–5518. doi: 10.1200/JCO.2006.08.4285. [PubMed] [Cross Ref]
9. Guillou L, et al. Germ cell tumors of the testis overexpress wild-type p53. Am J Pathol. 1996;149(4):1221–1228. [PubMed]
10. Asada M, et al. Apoptosis inhibitory activity of cytoplasmic p21(Cip1/WAF1) in monocytic differentiation. EMBO J. 1999;18(5):1223–1234. doi: 10.1093/emboj/18.5.1223. [PubMed] [Cross Ref]
11. Abbas T, Dutta A. p21 in cancer: intricate networks and multiple activities. Nat Rev Cancer. 2009;9(6):400–414. doi: 10.1038/nrc2657. [PMC free article] [PubMed] [Cross Ref]
12. Schepers H, Geugien M, Eggen BJ, Vellenga E. Constitutive cytoplasmic localization of p21(Waf1/Cip1) affects the apoptotic process in monocytic leukaemia. Leukemia. 2003;17(11):2113–2121. doi: 10.1038/sj.leu.2403106. [PubMed] [Cross Ref]
13. Zhan J, et al. Negative regulation of ASK1 by p21Cip1 involves a small domain that includes Serine 98 that is phosphorylated by ASK1 in vivo. Mol Cell Biol. 2007;27(9):3530–3541. doi: 10.1128/MCB.00086-06. [PMC free article] [PubMed] [Cross Ref]
14. Bartkova J, Thullberg M, Rajpert-De Meyts E, Skakkebaek NE, Bartek J. Cell cycle regulators in testicular cancer: loss of p18INK4C marks progression from carcinoma in situ to invasive germ cell tumours. Int J Cancer. 2000;85(3):370–375. doi: 10.1002/(SICI)1097-0215(20000201)85:3<370::AID-IJC13>3.0.CO;2-A. [PubMed] [Cross Ref]
15. Juric D, et al. Gene expression profiling differentiates germ cell tumors from other cancers and defines subtype-specific signatures. Proc Natl Acad Sci U S A. 2005;102(49):17763–17768. doi: 10.1073/pnas.0509082102. [PubMed] [Cross Ref]
16. Oosterhuis JW, Looijenga LH. Testicular germ-cell tumours in a broader perspective. Nat Rev Cancer. 2005;5(3):210–222. doi: 10.1038/nrc1568. [PubMed] [Cross Ref]
17. Vaughn DJ, et al. Treatment of growing teratoma syndrome. N Engl J Med. 2009;360(4):423–424. doi: 10.1056/NEJMc0808558. [PubMed] [Cross Ref]
18. Datta MW, Macri E, Signoretti S, Renshaw AA, Loda M. Transition from in situ to invasive testicular germ cell neoplasia is associated with the loss of p21 and gain of mdm-2 expression. Mod Pathol. 2001;14(5):437–442. doi: 10.1038/modpathol.3880331. [PubMed] [Cross Ref]
19. Huddart RA, Titley J, Robertson D, Williams GT, Horwich A, Cooper CS. Programmed cell death in response to chemotherapeutic agents in human germ cell tumour lines. Eur J Cancer. 1995;31A(5):739–746. [PubMed]
20. Spierings DC, de Vries EG, Vellenga E, de Jong S. Loss of drug-induced activation of the CD95 apoptotic pathway in a cisplatin-resistant testicular germ cell tumor cell line. Cell Death Differ. 2003;10(7):808–822. doi: 10.1038/sj.cdd.4401248. [PubMed] [Cross Ref]
21. di Pietro A, Vries EG, Gietema JA, Spierings DC, de Jong S. Testicular germ cell tumours: the paradigm of chemo-sensitive solid tumours. Int J Biochem Cell Biol. 2005;37(12):2437–2456. doi: 10.1016/j.biocel.2005.06.014. [PubMed] [Cross Ref]
22. Burger H, et al. Distinct p53-independent apoptotic cell death signalling pathways in testicular germ cell tumour cell lines. Int J Cancer. 1999;81(4):620–628. doi: 10.1002/(SICI)1097-0215(19990517)81:4<620::AID-IJC19>3.0.CO;2-S. [PubMed] [Cross Ref]
23. Timmer-Bosscha H, et al. cis-diamminedichloroplatinum(ii) resistance in vitro and in vivo in human embryonal carcinoma cells. Cancer Res. 1993;53(23):5707–5713. [PubMed]
24. Levkau B, et al. Cleavage of p21Cip1/Waf1 and p27Kip1 mediates apoptosis in endothelial cells through activation of Cdk2: role of a caspase cascade. Mol Cell. 1998;1(4):553–563. doi: 10.1016/S1097-2765(00)80055-6. [PubMed] [Cross Ref]
25. Jin YH, Yoo KJ, Lee YH, Lee SK. Caspase 3-mediated cleavage of p21WAF1/CIP1 associated with the cyclin A-cyclin-dependent kinase 2 complex is a prerequisite for apoptosis in SK-HEP-1 cells. J Biol Chem. 2000;275(39):30256–30263. doi: 10.1074/jbc.M001902200. [PubMed] [Cross Ref]
26. Adachi S, et al. Cyclin A/cdk2 activation is involved in hypoxia-induced apoptosis in cardiomyocytes. Circ Res. 2001;88(4):408–414. [PubMed]
27. Zhou BP, Liao Y, Xia W, Spohn B, Lee MH, Hung MC. Cytoplasmic localization of p21Cip1/WAF1 by Akt-induced phosphorylation in HER-2/neu-overexpressing cells. Nat Cell Biol. 2001;3(3):245–252. doi: 10.1038/35060032. [PubMed] [Cross Ref]
28. Harms C, et al. Phosphatidylinositol 3-Akt-kinase-dependent phosphorylation of p21(Waf1/Cip1) as a novel mechanism of neuroprotection by glucocorticoids. J Neurosci. 2007;27(17):4562–4571. doi: 10.1523/JNEUROSCI.5110-06.2007. [PubMed] [Cross Ref]
29. Di Vizio D, et al. Loss of the tumor suppressor gene PTEN marks the transition from intratubular germ cell neoplasias (ITGCN) to invasive germ cell tumors. Oncogene. 2005;24(11):1882–1894. doi: 10.1038/sj.onc.1208368. [PubMed] [Cross Ref]
30. Ivanovska I, et al. MicroRNAs in the miR-106b family regulate p21/CDKN1A and promote cell cycle progression. Mol Cell Biol. 2008;28(7):2167–2174. doi: 10.1128/MCB.01977-07. [PMC free article] [PubMed] [Cross Ref]
31. Petrocca F, et al. E2F1-regulated microRNAs impair TGFbeta-dependent cell-cycle arrest and apoptosis in gastric cancer. Cancer Cell. 2008;13(3):272–286. doi: 10.1016/j.ccr.2008.02.013. [PubMed] [Cross Ref]
32. Josephson R, et al. Qualification of embryonal carcinoma 2102Ep as a reference for human embryonic stem cell research. Stem Cells. 2007;25(2):437–446. doi: 10.1634/stemcells.2006-0236. [PubMed] [Cross Ref]
33. Marson A, et al. Connecting microRNA genes to the core transcriptional regulatory circuitry of embryonic stem cells. Cell. 2008;134(3):521–533. doi: 10.1016/j.cell.2008.07.020. [PMC free article] [PubMed] [Cross Ref]
34. Suzuki A, Tsutomi Y, Yamamoto N, Shibutani T, Akahane K. Mitochondrial regulation of cell death: mitochondria are essential for procaspase 3-p21 complex formation to resist Fas-mediated cell death. Mol Cell Biol. 1999;19(5):3842–3847. [PMC free article] [PubMed]
35. Zhou BB, Li H, Yuan J, Kirschner MW. Caspase-dependent activation of cyclin-dependent kinases during Fas-induced apoptosis in Jurkat cells. Proc Natl Acad Sci U S A. 1998;95(12):6785–6790. doi: 10.1073/pnas.95.12.6785. [PubMed] [Cross Ref]
36. Jin YH, Yim H, Park JH, Lee SK. Cdk2 activity is associated with depolarization of mitochondrial membrane potential during apoptosis. Biochem Biophys Res Commun. 2003;305(4):974–980. doi: 10.1016/S0006-291X(03)00870-2. [PubMed] [Cross Ref]
37. Choi JS, et al. Cyclin-dependent protein kinase 2 activity is required for mitochondrial translocation of Bax and disruption of mitochondrial transmembrane potential during etoposide-induced apoptosis. Apoptosis. 2007;12(7):1229–1241. doi: 10.1007/s10495-006-0047-3. [PubMed] [Cross Ref]
38. Huang H, Regan KM, Lou Z, Chen J, Tindall DJ. CDK2-dependent phosphorylation of FOXO1 as an apoptotic response to DNA damage. Science. 2006;314(5797):294–297. doi: 10.1126/science.1130512. [PubMed] [Cross Ref]
39. Bartkova J, et al. Deregulation of the RB pathway in human testicular germ cell tumours. J Pathol. 2003;200(2):149–156. doi: 10.1002/path.1353. [PubMed] [Cross Ref]
40. Kimura T, et al. AKT signaling promotes derivation of embryonic germ cells from primordial germ cells. Development. 2008;135(5):869–879. doi: 10.1242/dev.013474. [PubMed] [Cross Ref]
41. Palumbo C, et al. Expression of the PDGF alpha-receptor 1.5 kb transcript, OCT-4, and c-KIT in human normal and malignant tissues. Implications for the early diagnosis of testicular germ cell tumours and for our understanding of regulatory mechanisms. J Pathol. 2002;196(4):467–477. doi: 10.1002/path.1064. [PubMed] [Cross Ref]
42. Rapley EA, et al. Somatic mutations of KIT in familial testicular germ cell tumours. Br J Cancer. 2004;90(12):2397–2401. [PMC free article] [PubMed]
43. McIntyre A, Summersgill B, Spendlove HE, Huddart R, Houlston R, Shipley J. Activating mutations and/or expression levels of tyrosine kinase receptors GRB7, RAS, and BRAF in testicular germ cell tumors. Neoplasia. 2005;7(12):1047–1052. doi: 10.1593/neo.05514. [PMC free article] [PubMed] [Cross Ref]
44. Goddard NC, McIntyre A, Summersgill B, Gilbert D, Kitazawa S, Shipley J. KIT and RAS signalling pathways in testicular germ cell tumours: new data and a review of the literature. Int J Androl. 2007;30(4):337–348. doi: 10.1111/j.1365-2605.2007.00769.x. [PubMed] [Cross Ref]
45. Rapley EA, et al. A genome-wide association study of testicular germ cell tumor. Nat Genet. 2009;41(7):807–810. doi: 10.1038/ng.394. [PMC free article] [PubMed] [Cross Ref]
46. Kanetsky PA, et al. Common variation in KITLG and at 5q31.3 predisposes to testicular germ cell cancer. Nat Genet. 2009;41(7):811–815. doi: 10.1038/ng.393. [PMC free article] [PubMed] [Cross Ref]
47. Ihle NT, Powis G. The biological effects of isoform-specific PI3-kinase inhibition. Curr Opin Drug Discov Devel. 2010;13(1):41–49. [PubMed]
48. Voorhoeve PM, et al. A genetic screen implicates miRNA-372 and miRNA-373 as oncogenes in testicular germ cell tumors. Cell. 2006;124(6):1169–1181. doi: 10.1016/j.cell.2006.02.037. [PubMed] [Cross Ref]
49. West JA, et al. A role for Lin28 in primordial germ-cell development and germ-cell malignancy. Nature. 2009;460(7257):909–913. [PMC free article] [PubMed]
50. He L, et al. A microRNA polycistron as a potential human oncogene. Nature. 2005;435(7043):828–833. doi: 10.1038/nature03552. [PubMed] [Cross Ref]
51. Foshay KM, Gallicano GI. miR-17 family miRNAs are expressed during early mammalian development and regulate stem cell differentiation. Dev Biol. 2009;326(2):431–443. doi: 10.1016/j.ydbio.2008.11.016. [PubMed] [Cross Ref]
52. Giuliano CJ, et al. Retinoic acid represses a cassette of candidate pluripotency chromosome 12p genes during induced loss of human embryonal carcinoma tumorigenicity. Biochim Biophys Acta. 2005;1731(1):48–56. [PubMed]
53. Sark MW, et al. Cellular basis for differential sensitivity to cisplatin in human germ cell tumour and colon carcinoma cell lines. Br J Cancer. 1995;71(4):684–690. [PMC free article] [PubMed]
54. Timmer-Bosscha H, de Vries EG, Meijer C, Oosterhuis JW, Mulder NH. Differential effects of all-trans-retinoic acid, docosahexaenoic acid, and hexadecylphosphocholine on cisplatin-induced cytotoxicity and apoptosis in a cisplantin-sensitive and resistant human embryonal carcinoma cell line. Cancer Chemother Pharmacol. 1998;41(6):469–476. doi: 10.1007/s002800050769. [PubMed] [Cross Ref]
55. Gibcus JH, et al. Hodgkin lymphoma cell lines are characterized by a specific miRNA expression profile. Neoplasia. 2009;11(2):167–176. [PMC free article] [PubMed]
56. de Haas EC, et al. Clinical evaluation of M30 and M65 ELISA cell death assays as circulating biomarkers in a drug-sensitive tumor, testicular cancer. Neoplasia. 2008;10(10):1041–1048. [PMC free article] [PubMed]
57. [No authors listed] International Germ Cell Consensus Classification: a prognostic factor-based staging system for metastatic germ cell cancers. International Germ Cell Cancer Collaborative Group. J Clin Oncol. 1997;15(2):594–603. [PubMed]
58. Tan LP, et al. miRNA profiling of B-cell subsets: specific miRNA profile for germinal center B cells with variation between centroblasts and centrocytes. Lab Invest. 2009;89(6):708–716. doi: 10.1038/labinvest.2009.26. [PubMed] [Cross Ref]

Articles from The Journal of Clinical Investigation are provided here courtesy of American Society for Clinical Investigation