|Home | About | Journals | Submit | Contact Us | Français|
Tumor cell resistance to ionizing radiation and chemotherapy is a major obstacle in cancer therapy. One factor contributing to this is integrin-mediated adhesion to ECM. The adapter protein particularly interesting new cysteine-histidine-rich 1 (PINCH1) is recruited to integrin adhesion sites and promotes cell survival, but the mechanisms underlying this effect are not well understood. Here we have shown that PINCH1 is expressed at elevated levels in human tumors of diverse origins relative to normal tissue. Furthermore, PINCH1 promoted cell survival upon treatment with ionizing radiation in vitro and in vivo by perpetuating Akt1 phosphorylation and activity. Mechanistically, PINCH1 was found to directly bind to protein phosphatase 1α (PP1α) — an Akt1-regulating protein — and inhibit PP1α activity, resulting in increased Akt1 phosphorylation and enhanced radioresistance. Thus, our data suggest that targeting signaling molecules such as PINCH1 that function downstream of focal adhesions (the complexes that mediate tumor cell adhesion to ECM) may overcome radio- and chemoresistance, providing new therapeutic approaches for cancer.
Mutations in proto-oncogenes or tumor suppressors like Ras and p53, alteration in apoptosis signaling, and changes in the tumor microenvironment are common traits of tumor cell resistance to therapy (1–6), resulting in cancer stem cell survival and tumor recurrence (7). In combination with surgery and chemotherapy, more than 50% of cancer patients receive radiotherapy with high curative potential resulting from permanent local tumor control. In recent years, radiotherapy achieved additional success in local tumor control and cancer cure by the identification and targeting of molecules that critically regulate the radiation response of tumor cells (7–9). As molecules of focal adhesions (FAs), which mediate cell-ECM interactions, essentially contribute to radio- and chemoresistance of tumor cells (1, 10, 11), the present study aimed to identify further potent therapeutic cancer targets associated with FAs.
The major components of FAs are integrins, growth factor receptors, cytoplasmic signaling, and adapter molecules, which coalesce to a large membrane-associated multiprotein structure and signaling hub to control critical cell functions including survival, proliferation, migration, and cancer therapy response (1, 9, 12, 13). Among the essential regulators of FA function are the 5 Lin-1, Isl-1, Mec-3 (LIM) domain–containing particularly interesting new cysteine-histidine-rich 1 (PINCH1) and PINCH2 proteins (14). In avian and mouse models, PINCH1 has been shown to be critical for cell survival, spreading, adhesion, and migration by forming a ternary protein complex with integrin-linked kinase (ILK) and parvin (13, 15). Furthermore, the PINCH proteins have also been shown to bind Nck adapter protein 2 (Nck2; ref. 16), Ras suppressor protein 1 (Rsu1; ref. 17), and thymosin β4 (18). In concert with transmembrane integrin and growth factor receptors, the PINCH/ILK/parvin complex enables cooperative signaling for the regulation of cell survival via the PI3K/PKB/Akt1 and Ras/MAPK cascades (17, 19).
Akt1 plays a central role in apoptosis, tumor progression, resistance to cancer therapy, and radiation survival response (20–22). PI3K-dependent Akt1 activation by integrins and growth factor receptors through phosphorylation at amino acid residues S473 and T308 precedes Akt1-dependent phosphorylation of downstream targets such as forkhead box O (FoxO) transcription factors, glycogen synthase kinase 3 (GSK3), and Bad (20). Upon recruitment of Akt1 to the cell membrane, initiated by interactions of its pleckstrin homology (PH) domain with phosphoinositides, the 3-phosphoinositide–dependent kinase 1–dependent (PDK1-dependent) T308 phosphorylation in the activation T-loop is critical for Akt1 activation (23, 24). Subsequently, phosphorylation of S473 within a C-terminal hydrophobic motif of Akt1 occurs by mammalian target of rapamycin (mTOR), thereby determining substrate specificity of Akt1 (21, 25, 26). Fine tuning of Akt1 kinase activity also requires protein tyrosine phosphatases such as phosphatase and tensin homolog deleted on chromosome 10 (PTEN), which dephosphorylates phosphatidylinositol-(3,4,5)-trisphosphate (27), and protein phosphatase 1 (PP1) and PP2A (28, 29), which directly dephosphorylate Akt1. PTEN, PP1, and PP2A are ubiquitously expressed and can bind and regulate a large number of different proteins. For example, PP1 contributes to β1 integrin/β3 integrin interactions (30), and its activity is induced by ionizing radiation in a manner dependent on ataxia telangiectasia mutated (ATM; ref. 31). Many of these proteins recruit PP1 through a hydrophobic consensus motif, termed PP1-binding site, which consists of a RVXF binding groove (29, 32).
In the present study, we evaluated whether there is a link between the prosurvival factors PINCH1 and Akt1 for regulating radioresistance. We found that PINCH1 was overexpressed in human tumors and was essentially contributing to radiation cell survival in vitro and in vivo. Mechanistically, this PINCH1 function was mediated by direct interaction of PINCH1 and PP1α through a KFVEF binding motif at the LIM5 domain of PINCH1. The binding of PP1α by PINCH1 inhibited PP1α activity, causing Akt1 phosphorylation and radioresistance.
To examine how PINCH1 affects cell survival upon irradiation, we exposed mouse embryonic fibroblasts (MEFs) lacking PINCH1 expression (PINCH1–/– MEFs; Figure Figure1A)1A) to ionizing radiation and found in colony formation assays that they showed significantly enhanced sensitivity to irradiation compared with PINCH1fl/fl MEFs (Figure (Figure1,1, A–C). Similar results were obtained when cells were grown on 2D or 3D laminin-rich, basement membrane–like ECM (lrECM; Figure Figure1,1, B and C). Furthermore, enhanced sensitivity to irradiation or to the chemotherapeutic cisplatin was seen when PINCH1–/– cells were grown on fibronectin (Supplemental Figure 1, A and B; supplemental material available online with this article; doi: 10.1172/JCI41078DS1). Importantly, reexpression of EGFP-PINCH1 by retroviral transduction abolished the radiosensitization (Figure (Figure1,1, A and B, and Supplemental Figure 1A).
To verify the PINCH1 knockout–dependent increase in radiosensitivity in vivo, allograft tumors of immortalized PINCH1fl/fl and PINCH1–/– cell lines were established in immunocompromised mice (Figure (Figure1,1, D–F, and Supplemental Figure 2, A and B). When tumors reached a diameter of 6 mm, they were irradiated with increasing single doses of X-rays (26–62 Gy) and monitored up to 210 days. Consistent with the in vitro results, PINCH1–/– allografts demonstrated significantly higher radiosensitivity than did PINCH1fl/fl allografts in terms of both tumor growth delay and tumor recurrence–free survival (Figure (Figure1,1, E and F, Supplemental Figure 3, A–C, and Supplemental Tables 1 and 2). Moreover, PINCH1fl/fl tumors exhibited a different morphology and grew faster, as indicated by volume measurement and Ki-67 labeling index, than did PINCH1–/– tumors (Supplemental Table 2 and Supplemental Figure 4, A–C).
The sensitivity of tumors to radiation can be substantially reduced by microenvironmental factors such as hypoxia (33). Therefore, we examined hypoxia, necrosis, and vascularization in PINCH1fl/fl and PINCH1–/– tumors by injecting the hypoxia marker pimonidazole and Hoechst 33342 and then staining for the endothelial cell marker CD31. Interestingly, no significant differences for hypoxia, necrosis, and vascularization were found between PINCH1fl/fl and PINCH1–/– tumors (Supplemental Figure 4, A, D, and E).
To define the signaling cascades responsible for PINCH1-dependent regulation of radiation survival, we examined the prosurvival protein kinase Akt1 (21). Consistent with compromised survival upon irradiation, PINCH1–/– cells and PINCH1–/– allografts showed reduced levels of phosphorylated S473 and T308 of Akt1 compared with PINCH1fl/fl cells, PINCH1fl/fl allografts, and PINCH1 knockout MEFs reconstituted with EGFP-PINCH1 (Figure (Figure2,2, A–H, and Supplemental Figure 5A). The reduced Akt1 activity was also reflected by significantly decreased phosphorylation of FoxO1 at S256, whereas other Akt1 interacting proteins, such as FoxO4 and GSK3β (Supplemental Figure 5B and ref. 21), remained unchanged.
To verify changes in Akt1 activity, Akt1 kinase activity was measured in Akt1 immunoprecipitates from 2D and 3D lrECM cell cultures and found to be significantly reduced both in monolayer and in 3D cell cultures upon PINCH1 deletion (Figure (Figure2I).2I). To test the robustness of the PINCH1/Akt1 interrelation, we next examined changes in Akt1 phosphorylation upon siRNA-mediated PINCH1 knockdown and measured clonogenic survival of PINCH1fl/fl and EGFP-PINCH1 cells after Akt1 siRNA knockdown and irradiation. Strikingly, PINCH1 knockdown caused a strong reduction of Akt1 S473 and T308 phosphorylation (Figure (Figure2J),2J), and Akt1 knockdown significantly enhanced the radiosensitivity of PINCH1-expressing MEFs (Figure (Figure2K).2K). Taken together, these data suggest that PINCH1 is critical for radiation survival and Akt1 phosphorylation and activity.
To determine whether PINCH1 also plays a role in human tumors, we investigated its mRNA expression in selected tumor entities and corresponding normal tissues using Oncomine data sets as well as its protein expression in biopsies from patients with colorectal carcinoma. Importantly, comparison of PINCH1 transcriptomic profiles from tumor versus normal tissues from the most frequent tumor entities worldwide, such as lung, colon, breast, and prostate, showed a highly significant increase in mRNA expression in the tumors (Figure (Figure3A3A and Supplemental Table 3). Moreover, biopsies of colorectal carcinomas (Figure (Figure3B)3B) also revealed a highly significant elevation of PINCH1 protein expression compared with normal colon.
To confirm the role of PINCH1 in regulating resistance to radiation by interacting with Akt1, we next depleted PINCH1 in human cancer cell lines originating from colon (DLD1, HCT15, and HCT116), lung (A549 and H1299), cervix (Hela), skin (A431), and pancreas (PaTu and MiaPaCa2). In perfect agreement with our findings using MEFs, all carcinoma cell lines showed significant radiosensitization after siRNA-mediated PINCH1 depletion compared with cells treated with nonspecific siRNA controls (Figure (Figure3C3C and Supplemental Figure 6A). Similarly, PINCH1-depleted cancer cell lines exhibited significant chemosensitization to cisplatin, gemcitabine, and 5-fluorouracil (5-FU; Figure Figure3D3D and Supplemental Figure 6, B and C). Furthermore, PINCH1 depletion resulted in a striking reduction of the phosphorylation of Akt1 (S473 and T308), FoxO1 (S256), and FoxO4 (S197), but not GSK3β, relative to siRNA controls (Figure (Figure3E3E and Supplemental Figure 7, A–C). Thus, PINCH1 crucially determines Akt1 phosphorylation and radiation survival in mammalian cells with varying genetic backgrounds.
To further define whether the effects of PINCH1 are linked with Akt1, we evaluated the impact of single Akt1 versus combined Akt1/PINCH1 knockdown in DLD1 cells. Single Akt1 knockdown significantly reduced the radiation survival of DLD1 cells (Figure (Figure4A).4A). Combined Akt1/PINCH1 knockdown resulted in survival that was superimposable to that obtained with single Akt1 knockdown (Figure (Figure4A).4A). These data indicate that PINCH1 and Akt1 act in the same signaling pathway.
How does PINCH1 regulate Akt1 phosphorylation and kinase activity? We hypothesized that PINCH1 regulates either the phosphorylation of specific sites of Akt1, i.e., S473 or T308, or membrane translocation of Akt1, i.e., through its PH domain (21, 34). To distinguish between the 2 possibilities, we transiently transfected PINCH1-depleted human DLD1 colorectal carcinoma cells with different RFP-Akt1 plasmids (35) and assayed clonogenic survival and Akt1 phosphorylation. Strikingly, on a PINCH1 knockdown background, neither Akt1 mutants expressing single S→A kinase function–disrupting substitution at residue 473 (S473A) nor mutants expressing double kinase function–disrupting substitution of S→A plus T→A at residue 473 and 308 (S473A/T308A) were able to revert clonogenic radiation survival levels to that of Akt1 WT (Figure (Figure4,4, B and C). Additionally, both the constitutive active Akt1 mutant (S→D plus T→D at residues 473 and 308 [S473D/T308D]) as well as the deletion mutant of the aminoterminal PH domain of Akt1 (Akt1ΔPH) mediated radiation survival in the absence of PINCH1 in a manner similar to that observed for Akt1 WT (Figure (Figure4,4, B and C). The coincidence of similarity in radiation survival mediated by Akt1 WT, S473D/T308D, and Akt1ΔPH on a PINCH1 knockdown background and presence of phosphorylated Akt1 S473 (Figure (Figure4,4, B and C) indicate that the S473 phosphorylation site is the essential site for Akt1 regulation of radiation survival by PINCH1. Moreover, the retained ability of the Akt1ΔPH construct to rescue radiation survival of PINCH1-depleted DLD1 cells suggests that translocation of Akt1 to the cell membrane is not required for its prosurvival signaling. In the presence of PINCH1, DLD1 radiation survival remained unaltered by all tested Akt1 plasmids (Figure (Figure4C).4C). Taken together, these findings demonstrate that Akt1 lies downstream of PINCH1 in survival signaling.
To investigate whether PINCH1 and Akt1 interact physically, we performed mass spectrometry (MS) on EGFP and Akt1 immunoprecipitates from EGFP-PINCH1–reconstituted MEFs. The Akt1-regulating PP1α was detected in both EGFP-PINCH1 and Akt1 immunoprecipitates, identifying PP1α as a potential PINCH1-to-Akt1 linker protein (Supplemental Table 4). It has previously been shown that PP1α can control Akt1 activity (28), although the underlying regulatory mechanism is still unclear.
Next, we examined whether PINCH1, Akt1, and PP1α colocalize in cells and how loss of PINCH1 affects PP1α phosphatase activity. Interestingly, both Akt1 and PP1α were found together with EGFP-PINCH1 in FAs (Supplemental Figure 8). Moreover, loss of PINCH1 expression caused significantly elevated PP1α phosphatase activity without affecting the PP1α protein levels (Figure (Figure5A5A and Supplemental Figure 9, A and B). Similarly, DLD1 cells also exhibited significantly increased phosphatase activity upon PINCH1 knockdown (Figure (Figure5B5B and Supplemental Figure 9C). Finally, 2 nonspecific phosphatase inhibitors, okadaic acid (36) and tautomycetin (37), efficiently reduced the activity of PP1α in PINCH1fl/fl MEFs (Figure (Figure5A).5A). These data suggest that PINCH1 binds and regulates the phosphatase activity of PP1α.
To identify the PP1α binding motif of PINCH1, we analyzed the primary sequence of PINCH1 and found a highly conserved putative PP1α binding motif (32), KFVEF, at the aminoterminal PINCH1-LIM5 domain (Figure (Figure5C).5C). EGFP-PINCH1 without the LIM5 domain or point mutations in the KFVEF site (double V→A plus F→A substitution at residues 304 and 306 [V304A/F306A]; triple K→A plus V→A plus F→A substitution at residues 302, 304, and 306 [K302A/V304A/F306A]) were expressed in PINCH1–/– MEFs (Figure (Figure5D)5D) and failed to coprecipitate PP1α (Figure (Figure5E).5E). This indicates that the LIM5 domain with an intact KFVEF motif is required for an efficient PINCH1/PP1α interaction. As a consequence of the severely reduced PINCH1/PP1α association by deletion or mutation of the KFVEF motif, PP1 phosphatase activity remained elevated (Figure (Figure5F),5F), and Akt1 phosphorylation decreased (Figure (Figure6A),6A), compared with PINCH1–/– MEFs expressing EGFP-PINCH1 WT.
Although it has recently been shown that ILK is a pseudokinase (38), we excluded the PINCH1 binding partner ILK as a potential S473 kinase (19). In whole-cell lysates, we found reduced ILK expression levels in PINCH1–/– EGFP controls, but normal ILK expression in PINCH1–/– MEFs expressing EGFP-PINCH1 WT, EGFP-PINCH1ΔLIM5, or mutated KFVEF motifs (Figure (Figure6A).6A). These findings confirmed the interdependence of ILK expression and PINCH1 (39). Furthermore, ILK colocalized with PINCH1 WT and PINCH1 mutants in FAs (Figure (Figure6B),6B), whereas ILK was not in FAs of PINCH1–/– MEFs (data not shown). Taken together, these results suggest that PINCH1 regulates Akt1 through PP1α in an ILK-independent manner. As a consequence of an impeded PINCH1/PP1α interaction, the cellular radiosensitivity of PINCH1–/– MEFs expressing EGFP, EGFP-PINCH1ΔLIM5, or mutated KFVEF motifs was significantly enhanced at the clinically relevant radiation dose per fraction of 2 Gy compared with EGFP-PINCH1 WT–expressing MEFs (Figure (Figure6C). 6C).
Here, we showed that mRNA and protein expression of the FA adapter protein PINCH1 were significantly increased in human tumors of diverse origin, particularly colorectal carcinoma; that PINCH1 was critical for the regulation of cellular radio- and chemosensitivity; and that this function was mediated through the binding of PP1α to the KFVEF motif within the LIM5 domain, which caused inhibition of PP1α, activation of its downstream target Akt1, and, finally, enhanced radiation cell survival.
The combination of surgery and chemo- and radiotherapy leads to cancer cure rates greater than 50% among all cancer types. Limitations for further increases of these cure rates predominantly arise from normal tissue radiosensitivity and from dose-limiting, chemotherapy-related side effects. Therefore, the identification of key molecules acting in the survival pathways of tumor cells may allow the development of potent targeted strategies reducing tumor cell resistance to radio- and chemotherapy. Although it is well accepted that integrin-mediated adhesion promotes tumor cell resistance, it is less well understood how integrins execute this prosurvival effect in cancer cells. We report here that loss of PINCH1, which acts downstream of integrins, dramatically diminished clonogenic radiation cell survival and reduced cellular resistance to cisplatin, gemcitabine, and 5-FU. These profound effects were observed in PINCH1–/– MEFs and PINCH1-depleted human carcinoma cell lines originating from colon, lung, cervix, skin, and pancreas and confirmed in the MEF tumor xenograft model by measurement of tumor recurrence–free survival. Considering the administration of high, clinically irrelevant radiation doses in this study, examination of tumor recurrence under clinically more relevant fractionated radiation regimes in PINCH1-depleted human tumor models is warranted to support our notion of PINCH1 as putative cancer target. It is well known that cellular radiosensitivity can be substantially modulated by the tumor micromilieu, for example through pathological blood vessels or hypoxia (33, 40, 41). Since the tumor micromilieu was not affected by the loss of PINCH1, we conclude that the differences in tumor recurrence–free survival are tumor cell autonomous and predominantly PINCH1 dependent.
We and others have identified FA proteins that control tumor cell resistance. They include β1 and β3 integrins (4, 42, 43) as well as the EGFR (8, 9, 44), which acts together with integrins. These proteins were found to be overexpressed in diverse human malignancies. Interestingly, this was also the case for PINCH1, whose mRNA was significantly elevated compared with corresponding normal tissues. Furthermore, we found significantly increased levels of PINCH1 protein in colorectal carcinomas. Thus, the elevated PINCH1 levels in different cancers support our notion of PINCH1 serving as a potential cancer target.
LIM-containing proteins such as PINCH1 serve as a platform for recruiting and regulating the activity of downstream signaling molecules, either by direct activation or inhibition of the enzymatic activity or by spatial localization (45). PINCH1 was shown to recruit a number of proteins, such as ILK and Nck2, that link integrin and growth factor receptor signaling. While the interaction with Nck2 at the PINCH1 LIM4 domain seems extremely weak (46), ILK binding to the LIM1 domain of PINCH1 is of high affinity and required for the stability and localization of the PINCH1/ILK/parvin complex to FAs (15, 39, 47). Furthermore, the PINCH1/ILK interaction triggers signaling cascades required for cell survival and proliferation (39, 47, 48). The underlying mechanism for these important tasks was initially ascribed to the catalytic activity of ILK (49). It has been suggested that ILK controls cell proliferation by altering the activity of GSK3β or of specific prosurvival signal transduction pathways, such as the PI3K/Akt1 axis, which also play essential roles in radioresistance (22, 50). As a large number of genetic studies in flies, worms, and mice clearly showed that ILK is a pseudokinase (13, 24, 38), we sought for a mechanistic explanation of how PINCH1 regulates the activity of Akt1 and, consequently, radioresistance of tumor cells.
Since simultaneous PINCH1/Akt1 depletion caused radiation cell survival similar to that of single depletion of each of these proteins alone, we hypothesized that these 2 proteins act in the same signaling cascade. Given that overexpression of WT Akt1, constitutively active Akt1, and Akt1ΔPH rescued radiation survival of PINCH1-depleted tumor cells, and kinase function–disrupting mutations at S473 and S473/T308 failed to do so, we concluded that clonogenic survival of irradiated cells critically depends on the phosphorylation of Akt1. The PH domain of Akt1 seems not to be required to abolish PINCH1 knockdown–mediated radiosensitization. This finding could be explained by a possible prosurvival communication between PINCH1 and Akt1ΔPH distant from the cell membrane.
To unravel the cooperation between PINCH1 and Akt1, we precipitated PINCH1 and Akt1, respectively, and determined their interactomes by MS. In both PINCH1 and Akt1 precipitates, we found the Akt1-regulating PP1α. A sequence homology search revealed a single highly conserved PP1α binding motif, KFVEF, in the LIM5 domain of PINCH1. In immunoprecipitates from PINCH1–/– MEFs expressing LIM5 domain–deleted PINCH1 or PINCH1 with point mutations in the KFVEF motif (i.e., V304A/F306A and K302A/V304A/F306A), PP1α was absent, thus confirming the high relevance of this particular binding sequence for PP1α binding to PINCH1. When PP1α was properly bound to PINCH1 through the KFVEF motif at the PINCH1 LIM5 domain, PP1α phosphatase activity was inhibited. Intriguingly, this PP1α inhibition effectively prevented dephosphorylation of Akt1 at S473 and T308 and enhanced cellular resistance to ionizing radiation. With regard to the pseudokinase ILK and its requirement for proper signaling together with PINCH1, we examined ILK protein expression and FA localization in PINCH1–/– MEFs expressing WT PINCH1, PINCH1ΔLIM5, or KFVEF mutants. Although the expression of PINCH1 mutants rescued total ILK expression and subcellular localization to FAs, the PINCH1 mutants were unable to restore Akt1 S473 phosphorylation and PP1α functionality. Despite these molecular observations, the expression of PINCH1 mutants caused higher radiation survival compared with EGFP controls, indicating additional, yet to be defined prosurvival signals. Therefore, it remains to be determined which function the structural interaction between PINCH1 and ILK plays for PP1α binding. Our findings demonstrate a mode of action for how PINCH1 and PP1α interact to regulate Akt1 phosphorylation and cellular radioresistance, which we believe to be novel.
In summary, our data showed PINCH1 overexpression in various human tumors and defined what we believe to be a novel function of the LIM-only, FA-associated adapter protein PINCH1 in the cellular sensitivity to ionizing radiation in vitro and in vivo. The identification of stimulatory signals from PINCH1 to the prosurvival protein kinase Akt1 via PP1α delineates a mechanistic connection by which FA-associated processes essentially contribute to the regulation of the radiation survival response (Figure (Figure6D).6D). On the basis of PINCH1 overexpression in human malignancies, targeting PINCH1 represents a promising new therapeutic concept to overcome radio- and chemoresistance of tumor cells and to improve cure rates of cancer patients.
PINCH1fl/fl and PINCH1–/– MEFs and EGFP and EGFP-tagged full-length PINCH1 (EGFP-PINCH1) expressing PINCH1-deficient fibroblasts were generated as previously described (47). Human DLD1, HCT15, and HCT116 (colorectal adenocarcinoma); A549 and H1299 (lung carcinoma); Hela (cervical carcinoma); A431 (skin); and PaTu and MiaPaCa2 (pancreatic adenocarcinoma) cell lines were from ATCC. Single doses of 200 kV X-rays were applied. Treatment with cisplatin (0–50 μM; Neocorp), gemcitabine (Lilly), or 5-FU (0–250 μM; Medac) was performed for 1 hour followed by washing 3 times using PBS. Clonogenic survival was determined in 2D and 3D cell cultures as described in Supplemental Methods and as previously published (51).
Cell lysis, SDS-PAGE, Western blotting, and protein detection were performed as previously described (52).
Akt1 plasmids were generated as described previously (35). The mouse PINCH1 WT and PINCH1 LIM5 deletion mutant (PINCH1ΔLIM5) were generated by PCR-based amplification with specific primers. Constructs were flanked with KpnI and BamHI restriction sites and inserted into the KpnI and BamHI sites of pEGFP-C1 (Clontech). Mutation of the KFVEF motif was performed using appropriate primers (see Supplemental Methods) and the QuikChange II Site-Directed Mutagenesis Kit (Stratagene) according to the manufacturer’s instructions and confirmed by sequencing.
In-gel digestion, LC-MS/MS analysis on the LTQ ion trap mass spectrometer (Thermo Fisher Scientific), and spectra processing procedure were performed on anti-GFP and anti-Akt1 antibody pulldowns from EGFP-PINCH1 MEFs as described previously (53). Proteins were identified in the mouse International Protein Index database ( http://www.ebi.ac.uk/IPI; accessed April 2008) by MASCOT version 2.1 software (Matrix Science Ltd.). Protein hits were considered confident when at least 2 MS/MS spectra matched the corresponding database sequences with peptide ion scores exceeding the confidence threshold suggested by MASCOT (P < 0.05).
To measure PP1 activity, the ProFluor Ser/Thr Phosphatase Assay kit (Promega) was used as described previously (54). Okadaic acid and tautomycetin were from Applichem and Tocris, respectively.
Animal facilities and experiments were approved by the Animal Care and Use Committee of the Medical Faculty Carl Gustav Carus, Dresden University of Technology, and the Regierungspräsidium Dresden according to the German animal welfare regulations. Immunocompromised 7- to 14-week-old male and female NMRI (nu/nu) mice were further immunosuppressed by whole-body irradiation 1–5 days before tumor transplantation. We serially transplanted cryopreserved chunks of PINCH1fl/fl and PINCH1–/– tumors onto the back of mice. For the experiments, pieces of approximately 1 mm size of second- and third-passage tumors with median growth rate were subcutaneously transplanted onto the right hind leg. PCR genotyping and PINCH1 protein expression of tumors was performed as described in Supplemental Methods. After reaching a tumor volume of 0.10–0.32 cm3, animals were randomly allocated to the different radiation dose levels in groups of 4 animals. Single doses of 26, 32, 38, 44, 50, 56, or 62 Gy (200 kV X-rays) were given under homogeneous hypoxia using a heavy clamp placed over the proximal thigh of the tumor-bearing leg of anesthetized mice (120 mg/kg body weight ketamine i.p. and 16 mg/kg xylazine i.p.) 2 minutes before and during irradiation. PINCH1fl/fl and PINCH1–/– tumors were equally distributed over the dose groups (Supplemental Table 1).
The tumor diameter was measured twice per week using a caliper. Tumor volumes were determined by the formula of a rotational ellipsoid: π/6 × a × b2, where a is the longest and b is the perpendicular shorter tumor axis. Tumor growth time (TGT) of unirradiated and irradiated PINCH1fl/fl and PINCH1–/– tumors was evaluated from growth curves of individual tumors as the time needed after start of the experiment to reach 2 and 5 times the starting volume (TGTV2 and TGTV5, respectively). Tumor volume was measured until tumors reached a volume of about 1.5 cm3. Local tumor recurrences after irradiation were scored when the tumor volume increased for 3 consecutive measurements after passing a nadir. A total of 85 PINCH1fl/fl and 99 PINCH1–/– tumors were evaluated for local tumor control. In PINCH1fl/fl and PINCH1–/– tumor-bearing animals, 95% of all recurrences were scored before days 51 and 102 after irradiation, respectively. The last recurrences occurred at days 128 and 147 for PINCH1fl/fl and PINCH1–/–, respectively. Animals without recurrent tumors within the radiation field (i.e., with locally controlled tumors) were followed up until death. For analysis of local tumor control data, censored animals were taken into account as described in Supplemental Methods. For histology, a total of 12 PINCH1fl/fl and 9 PINCH1–/– tumors were subjected to histological analysis. Methods to assess vasculature, hypoxia, and perfusion on frozen sections and index for Ki-67 staining on paraffin-embedded tumors were performed as described in Supplemental Methods. For analysis of Akt1 expression and phosphorylation, pieces of PINCH1fl/fl and PINCH1–/– tumors were homogenized and lysed (25 mM Tris-HCl, 1.5 mM EGTA, 0.5 mM EDTA, 1% Triton X-100, 1 mM Na3VO4, and 1 mM sodium pyruvate) on ice and subjected to SDS-PAGE, Western blotting, and protein detection.
For the analysis of PINCH expression data, Oncomine ( http://www.oncomine.org) was searched for studies comparing tumor with normal tissue containing probes for PINCH1 (selected studies are listed in Supplemental References). Within a given data set, the probe was selected that possessed the higher expression values. The normalized values of the individual data sets were combined and further analyzed. Since single data sets are split by tissue type at Oncomine, the sample identifier was used to eliminate duplicates. Expression values representing chronically inflamed tissues like chronic pancreatitis were purged from the data. For the highly diverse class of brain tumors, only glioblastomas were included. Gene expression differences were analyzed by 2-sided Student’s t test using GraphPad Prism software version 4.03.
Paraffin blocks of 24 human tissue specimens each from tumor and normal tissues were taken from the archive of the Institute of Pathology under approval of the local ethics committee of the Dresden University of Technology and with the patients’ consent. Paraffin sections were stained with a monoclonal mouse anti-PINCH1 antibody (P-9371; Sigma-Aldrich) at a dilution of 1:300 in an autostainer 480 (Labvision) according to the manufacturer’s instructions. Counterstaining was performed with hematoxylin. Staining intensities of the samples were evaluated by a pathologist and were semiquantitatively scored using a 4-tiered score (staining intensities of 0, 1, 2, and 3) as described previously (55). Statistical analysis tested PINCH1 expression in tumor versus normal tissues using a c2 test. Negative control reactions yielded no staining confirming the specificity of the antibody recognition.
Data are presented as mean ± SD or mean ± SEM. The level of significance was determined by unpaired, 2-sided Student’s t test, Mann-Whitney U test (median values for TGT), or log-rank test (actuarial estimates for time to local tumor recurrence were obtained using the Kaplan-Meier method; GraphPad Prism software version 4.03).
We thank Eric J. Bernhard, Erik H.J. Danen, and Kyle Legate for critical discussions and reading of the manuscript. We further thank Inga Lange and Daniela Tschuck for technical assistance and James Raleigh for the pimonidazole antibody. This work was supported in part by the Bundesministerium für Bildung und Forschung (to M. Baumann and N. Cordes), Deutsche Forschungsgemeinschaft grant Ba1433/5 (to M. Baumann and D. Zips), Austrian Science Foundation grant SFB021 and the Max Planck Society (to R. Fassler), and Cancer Research UK Core funding to Lincolns Inn Fields Laboratories (to B. Larijani).
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J Clin Invest. 2010;120(7):2516–2527. doi:10.1172/JCI41078.