Treatment of glioblastoma using multicomponent silica nanoparticles.

Glioblastomas (GBMs) remain highly lethal. This partially stems from the presence of brain tumor initiating cells (BTICs), a highly plastic cellular subpopulation that is resistant to current therapies. In addition to resistance, the blood-brain barrier limits the penetration of most drugs into GBMs. To effectively deliver a BTIC-specific inhibitor to brain tumors, we developed a multicomponent nanoparticle, termed Fe@MSN, which contains a mesoporous silica shell and an iron oxide core. Fibronectin-targeting ligands directed the nanoparticle to the near-perivascular areas of GBM. After Fe@MSN particles deposited in the tumor, an external low-power radiofrequency (RF) field triggered rapid drug release due to mechanical tumbling of the particle resulting in penetration of high amounts of drug across the blood-brain tumor interface and widespread drug delivery into the GBM. We loaded the nanoparticle with the drug 1400W, which is a potent inhibitor of the inducible nitric oxide synthase (iNOS). It has been shown that iNOS is preferentially expressed in BTICs and is required for their maintenance. Using the 1400W-loaded Fe@MSN and RF-triggered release, in vivo studies indicated that the treatment disrupted the BTIC population in hypoxic niches, suppressed tumor growth and significantly increased survival in BTIC-derived GBM xenografts.

Acquisition of meiotic DNA repair regulators maintain genome stability in glioblastoma.

Glioblastoma (GBM), the most prevalent type of primary intrinsic brain cancer in adults, remains universally fatal despite maximal therapy, including radiotherapy and chemotherapy. Cytotoxic therapy generates double-stranded DNA breaks (DSBs), most commonly repaired by homologous recombination (HR). We hypothesized that cancer cells coopt meiotic repair machinery as DSBs are generated during meiosis and repaired by molecular complexes distinct from genotoxic responses in somatic tissues. Indeed, we found that gliomas express meiotic repair genes and their expression informed poor prognosis. We interrogated the function of disrupted meiotic cDNA1 (DMC1), a homolog of RAD51, the primary recombinase used in mitotic cells to search and recombine with the homologous DNA template. DMC1, whose only known function is as an HR recombinase, was expressed by GBM cells and induced by radiation. Although targeting DMC1 in non-neoplastic cells minimally altered cell growth, DMC1 depletion in GBM cells decreased proliferation, induced activation of CHK1 and expression of p21CIP1/WAF1, and increased RPA foci, suggesting increased replication stress. Combining loss of DMC1 with ionizing radiation inhibited activation of DNA damage responses and increased radiosensitivity. Furthermore, loss of DMC1 reduced tumor growth and prolonged survival in vivo. Our results suggest that cancers coopt meiotic genes to augment survival under genotoxic stress, offering molecular targets with high therapeutic indices.

EphA2 promotes infiltrative invasion of glioma stem cells in vivo through cross-talk with Akt and regulates stem cell properties.

Diffuse infiltrative invasion is a major cause for the dismal prognosis of glioblastoma multiforme (GBM), but the underlying mechanisms remain incompletely understood. Using human glioma stem cells (GSCs) that recapitulate the invasive propensity of primary GBM, we find that EphA2 critically regulates GBM invasion in vivo. EphA2 was expressed in all seven GSC lines examined, and overexpression of EphA2 enhanced intracranial invasion. The effects required Akt-mediated phosphorylation of EphA2 on serine 897. In vitro the Akt-EphA2 signaling axis is maintained in the absence of ephrin-A ligands and is disrupted upon ligand stimulation. To test whether ephrin-As in tumor microenvironment can regulate GSC invasion, the newly established Efna1;Efna3;Efna4 triple knockout mice (TKO) were used in an ex vivo brain slice invasion assay. We observed significantly increased GSC invasion through the brain slices of TKO mice relative to wild-type (WT) littermates. Mechanistically EphA2 knockdown suppressed stem cell properties of GSCs, causing diminished self-renewal, reduced stem marker expression and decreased tumorigenicity. In a subset of GSCs, the reduced stem cell properties were associated with lower Sox2 expression. Overexpression of EphA2 promoted stem cell properties in a kinase-independent manner and increased Sox2 expression. Disruption of Akt-EphA2 cross-talk attenuated stem cell marker expression and neurosphere formation while having minimal effects on tumorigenesis. Taken together, the results show that EphA2 endows invasiveness of GSCs in vivo in cooperation with Akt and regulates glioma stem cell properties.

Phage display discovery of novel molecular targets in glioblastoma-initiating cells.

Glioblastoma is the most common primary intrinsic brain tumor and remains incurable despite maximal therapy. Glioblastomas display cellular hierarchies with self-renewing glioma-initiating cells (GICs) at the apex. To discover new GIC targets, we used in vivo delivery of phage display technology to screen for molecules selectively binding GICs that may be amenable for targeting. Phage display leverages large, diverse peptide libraries to identify interactions with molecules in their native conformation. We delivered a bacteriophage peptide library intravenously to a glioblastoma xenograft in vivo then derived GICs. Phage peptides bound to GICs were analyzed for their corresponding proteins and ranked based on prognostic value, identifying VAV3, a Rho guanine exchange factor involved tumor invasion, and CD97 (cluster of differentiation marker 97), an adhesion G-protein-coupled-receptor upstream of Rho, as potentially enriched in GICs. We confirmed that both VAV3 and CD97 were preferentially expressed by tumor cells expressing GIC markers. VAV3 expression correlated with increased activity of its downstream mediator, Rac1 (ras-related C3 botulinum toxin substrate 1), in GICs. Furthermore, targeting VAV3 by ribonucleic acid interference decreased GIC growth, migration, invasion and in vivo tumorigenesis. As CD97 is a cell surface protein, CD97 selection enriched for sphere formation, a surrogate of self-renewal. In silico analysis demonstrated VAV3 and CD97 are highly expressed in tumors and inform poor survival and tumor grade, and more common with epidermal growth factor receptor mutations. Finally, a VAV3 peptide sequence identified on phage display specifically internalized into GICs. These results show a novel screening method for identifying oncogenic pathways preferentially activated within the tumor hierarchy, offering a new strategy for developing glioblastoma therapies.

Therapeutic targeting of constitutive PARP activation compromises stem cell phenotype and survival of glioblastoma-initiating cells.

Glioblastoma-initiating cells (GICs) are self-renewing tumorigenic sub-populations, contributing to therapeutic resistance via decreased sensitivity to ionizing radiation (IR). GIC survival following IR is attributed to an augmented response to genotoxic stress. We now report that GICs are primed to handle additional stress due to basal activation of single-strand break repair (SSBR), the main DNA damage response pathway activated by reactive oxygen species (ROS), compared with non-GICs. ROS levels were higher in GICs and likely contributed to the oxidative base damage and single-strand DNA breaks found elevated in GICs. To tolerate constitutive DNA damage, GICs exhibited a reliance on the key SSBR mediator, poly-ADP-ribose polymerase (PARP), with decreased viability seen upon small molecule inhibition to PARP. PARP inhibition (PARPi) sensitized GICs to radiation and inhibited growth, self-renewal, and DNA damage repair. In vivo treatment with PARPi and radiotherapy attenuated radiation-induced enrichment of GICs and inhibited the central cancer stem cell phenotype of tumor initiation. These results indicate that elevated PARP activation within GICs permits exploitation of this dependence, potently augmenting therapeutic efficacy of IR against GICs. In addition, our results support further development of clinical trials with PARPi and radiation in glioblastoma.

Bevacizumab continuation beyond initial bevacizumab progression among recurrent glioblastoma patients.

BACKGROUND: Bevacizumab improves outcome for most recurrent glioblastoma patients, but the duration of benefit is limited and survival after initial bevacizumab progression is poor. We evaluated bevacizumab continuation beyond initial progression among recurrent glioblastoma patients as it is a common, yet unsupported practice in some countries. METHODS: We analysed outcome among all patients (n=99) who received subsequent therapy after progression on one of five consecutive, single-arm, phase II clinical trials evaluating bevacizumab regimens for recurrent glioblastoma. Of note, the five trials contained similar eligibility, treatment and assessment criteria, and achieved comparable outcome. RESULTS: The median overall survival (OS) and OS at 6 months for patients who continued bevacizumab therapy (n=55) were 5.9 months (95% confidence interval (CI): 4.4, 7.6) and 49.2% (95% CI: 35.2, 61.8), compared with 4.0 months (95% CI: 2.1, 5.4) and 29.5% (95% CI: 17.0, 43.2) for patients treated with a non-bevacizumab regimen (n=44; P=0.014). Bevacizumab continuation was an independent predictor of improved OS (hazard ratio=0.64; P=0.04). CONCLUSION: The results of our retrospective pooled analysis suggest that bevacizumab continuation beyond initial progression modestly improves survival compared with available non-bevacizumab therapy for recurrent glioblastoma patients require evaluation in an appropriately randomised, prospective trial.

Hypoxia-induced mixed-lineage leukemia 1 regulates glioma stem cell tumorigenic potential.

Normal stem cells reside in functional niches critical for self-renewal and maintenance. Neural and hematopoietic stem cell niches, in particular, are characterized by restricted availability of oxygen and the resulting regulation by hypoxia-inducible factors (HIFs). Glioblastoma multiforme (GBM) is the most common malignant brain tumor and also contains high degrees of hypoxia. Heterogeneity within the neoplastic compartment has been well characterized in GBM and may be derived from genetic and epigenetic sources that co-evolve during malignant progression. Recent experimental evidence has supported the importance of hypoxia in glioma stem cell (GSC) niches. We hypothesized that HIFs require epigenetic-modifying proteins to promote tumor malignancy in GBM. Here we demonstrate that in GBM the histone methyltransferase mixed-lineage leukemia 1 (MLL1) is induced by hypoxia and enhances hypoxic responses. Loss of MLL1 reduces the expression of HIF transcripts and HIF2α protein. Targeting MLL1 by RNA interference inhibited the expression of HIF2α and target genes, including vascular endothelial growth factor (VEGF). GSCs expressed higher levels of MLL1 than matched non-stem tumor cells and depletion of MLL1 reduced GSC self-renewal, growth, and tumorigenicity. These studies have uncovered a novel mechanism mediating tumor hypoxic responses linking microenvironmental regulation of epigenetic-modifying proteins to cellular heterogeneity and provide rationale for the design of more sophisticated clinical approaches targeting epigenetic regulation.

Distribution of CD133 reveals glioma stem cells self-renew through symmetric and asymmetric cell divisions.

Malignant gliomas contain a population of self-renewing tumorigenic stem-like cells; however, it remains unclear how these glioma stem cells (GSCs) self-renew or generate cellular diversity at the single-cell level. Asymmetric cell division is a proposed mechanism to maintain cancer stem cells, yet the modes of cell division that GSCs utilize remain undetermined. Here, we used single-cell analyses to evaluate the cell division behavior of GSCs. Lineage-tracing analysis revealed that the majority of GSCs were generated through expansive symmetric cell division and not through asymmetric cell division. The majority of differentiated progeny was generated through symmetric pro-commitment divisions under expansion conditions and in the absence of growth factors, occurred mainly through asymmetric cell divisions. Mitotic pair analysis detected asymmetric CD133 segregation and not any other GSC marker in a fraction of mitoses, some of which were associated with Numb asymmetry. Under growth factor withdrawal conditions, the proportion of asymmetric CD133 divisions increased, congruent with the increase in asymmetric cell divisions observed in the lineage-tracing studies. Using single-cell-based observation, we provide definitive evidence that GSCs are capable of different modes of cell division and that the generation of cellular diversity occurs mainly through symmetric cell division, not through asymmetric cell division.

Acidic stress promotes a glioma stem cell phenotype.

Malignant gliomas are lethal cancers that display cellular hierarchies with cancer stem cells at the apex. Glioma stem cells (GSCs) are not uniformly distributed, but rather located in specialized niches, suggesting that the cancer stem cell phenotype is regulated by the tumor microenvironment. Indeed, recent studies show that hypoxia and its molecular responses regulate cancer stem cell maintenance. We now demonstrate that acidic conditions, independent of restricted oxygen, promote the expression of GSC markers, self-renewal and tumor growth. GSCs exert paracrine effects on tumor growth through elaboration of angiogenic factors, and low pH conditions augment this expression associated with induction of hypoxia inducible factor 2α (HIF2α), a GSC-specific regulator. Induction of HIF2α and other GSC markers by acidic stress can be reverted by elevating pH in vitro, suggesting that raising intratumoral pH may be beneficial for targeting the GSC phenotype. Together, our results suggest that exposure to low pH promotes malignancy through the induction of a cancer stem cell phenotype, and that culturing cancer cells at lower pH reflective of endogenous tumor conditions may better retain the cellular heterogeneity found in tumors.

Hypoxia inducible factors in cancer stem cells.

Oxygen is an essential regulator of cellular metabolism, survival, and proliferation. Cellular responses to oxygen levels are monitored, in part, by the transcriptional activity of the hypoxia inducible factors (HIFs). Under hypoxia, HIFs regulate a variety of pro-angiogenic and pro-glycolysis pathways. In solid cancers, regions of hypoxia are commonly present throughout the tissue because of the chaotic vascular architecture and regions of necrosis. In these regions, the hypoxic state fluctuates in a spatial and temporal manner. Transient hypoxic cycling causes an increase in the activity of the HIF proteins above what is typical for non-pathologic tissue. The extent of hypoxia strongly correlates to poor patient survival, therapeutic resistance and an aggressive tumour phenotype, but the full contribution of hypoxia and the HIFs to tumour biology is an area of active investigation. Recent reports link resistance to conventional therapies and the metastatic potential to a stem-like tumour population, termed cancer stem cells (CSCs). We and others have shown that within brain tumours CSCs reside in two niches, a perivascular location and the surrounding necrotic tissue. Restricted oxygen conditions increase the CSC fraction and promote acquisition of a stem-like state. Cancer stem cells are critically dependant on the HIFs for survival, self-renewal, and tumour growth. These observations and those from normal stem cell biology provide a new mechanistic explanation for the contribution of hypoxia to malignancy. Further, the presence of hypoxia in tumours may present challenges for therapy because of the promotion of CSC phenotypes even upon successful killing of CSCs. The current experimental evidence suggests that CSCs are plastic cell states governed by microenvironmental conditions, such as hypoxia, that may be critical for the development of new therapies targeted to disrupt the microenvironment.

Cancer stem cells in brain tumor biology.

Tumors are aberrant organ systems containing a complex interplay between the neoplastic compartment and recruited vascular, inflammatory, and stromal elements. Furthermore, most cancers display a hierarchy of differentiation states within the tumor cell population. Molecular signals that drive tumor formation and maintenance commonly overlap with those involved in normal development and wound responses--two processes in which normal stem cells function. It is therefore not surprising that cancers invoke stem cell programs that promote tumor malignancy. Stem-cell-like cancer cells (or cancer stem cells) need not be derived from normal stem cells but may be subjected to evolutionary pressures that select for the capacity to self-renew extensively or differentiate depending on conditions. Current cancer model systems may not fully recapitulate the cellular complexity of cancers, perhaps partially explaining the lack of power of these models in predicting clinical outcomes. New methods are enabling researchers to identify and characterize cancer stem cells. Our laboratory focuses on the roles of brain tumor stem cells in clinically relevant tumor biology, including therapeutic resistance, angiogenesis, and invasion/metastasis. We hope that these studies will translate into improved diagnostic, prognostic, and therapeutic approaches for these lethal cancers.

Targeting SPARC expression decreases glioma cellular survival and invasion associated with reduced activities of FAK and ILK kinases.

Secreted protein acidic and rich in cysteine (SPARC) is an extracellular glycoprotein expressed in several solid cancers, including malignant gliomas, upon adoption of metastatic or invasive behaviors. SPARC expression in glioma cells promotes invasion and survival under stress, the latter process dependent on SPARC activation of AKT. Here we demonstrate that downregulation of SPARC expression with short interfering RNA (siRNA) in glioma cells decreased tumor cell survival and invasion. SPARC siRNA reduced the activating phosphorylation of AKT and two cytoplasmic kinases, focal adhesion kinase (FAK) and integrin-linked kinase (ILK). We determined the contributions of FAK and ILK to SPARC effects using SPARC protein and cell lines engineered to overexpress SPARC. SPARC activated FAK and ILK in glioma cells previously characterized as responsive to SPARC. Downregulation of either FAK or ILK expression inhibited SPARC-mediated AKT phosphorylation, and targeting both FAK and ILK attenuated AKT activation more potently than targeting either FAK or ILK alone. Decreased SPARC-mediated AKT activation correlated with a reduction in SPARC-dependent invasion and survival upon the downregulation of FAK and/or ILK expression. These data further demonstrate the role of SPARC in glioma tumor progression through the activation of intracellular kinases that may provide novel therapeutic targets for advanced cancers.

Phase I study of Gliadel wafers plus temozolomide in adults with recurrent supratentorial high-grade gliomas.

Both Gliadel wafers [1,3-bis(2-chloroethyl)-1-nitrosourea] and temozolomide (TEMO) have been shown in independent studies to prolong survival of patients with recurrent malignant glioma following surgery and radiotherapy. On the basis of preclinical evidence of synergism between Gliadel wafers and TEMO, a phase I study was designed to evaluate the toxicity of combining these 2 agents in the treatment of patients with recurrent supratentorial malignant glioma. All patients had surgical resection of the tumor at relapse, and up to 8 Gliadel (3.85%) wafers were placed in the surgical cavity following resection. Two weeks after surgery, TEMO was given orally daily for 5 days. Cohorts of 3 patients received TEMO at daily doses of 100 mg/m2, 150 mg/m2, and 200 mg/m2, respectively. Patients were assessed for toxicity 4 weeks after start of the first course of TEMO. Contrast-enhanced MRI of the brain was used to assesstumor response after the first cycle of TEMO. Patients with stable disease or response after the first cycle of TEMO were allowed to continue treatment at the same dose every 4 weeks for 12 cycles or until disease progression or unacceptable toxicity. Ten patients with a median age of 47 years (range, 22-66 years) were enrolled in this study. There were 7 patients with glioblastoma multiforme and 3 patients with anaplastic astrocytoma. Three patients were treated with TEMO at the first dose level of 100 mg/m2, 4 at the second dose level of 150 mg/m2, and 3 at the third dose level of 200 mg/m2. The 10 patients received a median of 3 cycles (range, 1-12 cycles) of TEMO following placement of Gliadel wafers. The treatment was well tolerated, with only 1 patient suffering grade III thrombocytopenia at the highest dose level. Two patients at each dose level had no evidence of disease progression after treatment. Four patients suffered progressive disease on therapy. Our study demonstrates that TEMO can be given safely after placement of Gliadel (3.85%) wafers. The recommended dosage for TEMO for a phase II study of this combination is 200 mg/m2 per day for 5 days.

A genetically tractable model of human glioma formation.

Gliomas remain one of the deadliest forms of cancer. Improved therapeutics will require a better understanding of the molecular nature of these tumors. We, therefore, mimicked the most common genetic changes found in grade III-IV gliomas, disruption of the p53 and RB pathways and activation of telomere maintenance and independence from growth factors, through the ectopic expression of the SV40 T/t-Ag oncogene, an oncogenic form of H-ras (H-ras(V12G)), and the human telomerase catalytic subunit hTERT in normal human astrocytes. The resulting cells displayed many of the hallmarks of grade III-IV gliomas, including greatly expanded life span and growth in soft agar and, most importantly, were tumorigenic with pathology consistent with grade III-IV neuroectodermal tumors in mice. This model system will, for the first time, allow the biological significance of selected genetic alterations to be studied in human gliomas.

Phase I trial of carmustine plus O6-benzylguanine for patients with recurrent or progressive malignant glioma.

PURPOSE: The major mechanism of resistance to alkylnitrosourea therapy involves the DNA repair protein O(6)-alkylguanine-DNA alkyltransferase (AGT), which removes chloroethylation or methylation damage from the O(6) position of guanine. O(6)-benzylguanine (O(6)-BG) is an AGT substrate that inhibits AGT by suicide inactivation. We conducted a phase I trial of carmustine (BCNU) plus O(6)-BG to define the toxicity and maximum-tolerated dose (MTD) of BCNU in conjunction with the preadministration of O(6)-BG with recurrent or progressive malignant glioma. PATIENTS AND METHODS: Patients were treated with O(6)-BG at a dose of 100 mg/m(2) followed 1 hour later by BCNU. Cohorts of three to six patients were treated with escalating doses of BCNU, and patients were observed for at least 6 weeks before being considered assessable for toxicity. Plasma samples were collected and analyzed for O(6)-BG, 8-oxo-O(6)-BG, and 8-oxoguanine concentration. RESULTS: Twenty-three patients were treated (22 with glioblastoma multiforme and one with anaplastic astrocytoma). Four dose levels of BCNU (13.5, 27, 40, and 55 mg/m(2)) were evaluated, with the highest dose level being complicated by grade 3 or 4 thrombocytopenia and neutropenia. O(6)-BG rapidly disappeared from plasma (elimination half-life = 0. 54 +/- 0.14 hours) and was converted to a longer-lived metabolite, 8-oxo-O(6)-BG (elimination half-life = 5.6 +/- 2.7 hours) and further to 8-oxoguanine. There was no detectable O(6)-BG 5 hours after the start of the O(6)-BG infusion; however, 8-oxo-O(6)-BG and 8-oxoguanine concentrations were detected 25 hours after O(6)-BG infusion. The mean area under the concentration-time curve (AUC) of 8-oxo-O(6)-BG was 17.5 times greater than the mean AUC for O(6)-BG. CONCLUSION: These results indicate that the MTD of BCNU when given in combination with O(6)-BG at a dose of 100 mg/m(2) is 40 mg/m(2) administered at 6-week intervals. This study provides the foundation for a phase II trial of O(6)-BG plus BCNU in nitrosourea-resistant malignant glioma.

Transforming growth factor-beta-mediated p15(INK4B) induction and growth inhibition in astrocytes is SMAD3-dependent and a pathway prominently altered in human glioma cell lines.

We sought to characterize the pathway by which the multifunctional cytokine transforming growth factor-beta (TGF-beta) inhibits the proliferation of normal astrocytes, and we analyzed the alterations in the TGF-beta pathway in human glioma cell lines. Upon TGF-beta treatment, primary rat astrocytes showed a significant decrease in DNA synthesis upon thymidine incorporation with a cell cycle arrest in the G(1) phase. Western analysis of the astrocytes revealed that the expression of the cyclin-dependent kinase inhibitor (CdkI) p15(INK4B) was significantly up-regulated upon TGF-beta treatment without a change in other CdkI levels. The retinoblastoma protein (Rb) became hypophosphorylated, and Cdk2 activity decreased. Analysis of Smad3 null mouse astrocytes showed a significant loss of both TGF-beta-mediated growth inhibition and p15(INK4B) induction compared with wild-type mouse astrocytes. Infection of rat astrocytes by SMAD3 and SMAD4 adenoviruses failed to induce increased expression of p15(INK4B), implying indirect transcriptional regulation of p15(INK4B) by SMAD3. High-grade human gliomas secrete TGF-beta, yet are resistant to its growth inhibitory effects. Analysis of the effects of TGF-beta on 12 human glioma cell lines showed that TGF-beta mildly inhibited the growth of six lines, had no effect on four lines, and stimulated the growth of two lines. The majority of glioma lines had homozygous deletions of the p15(INK4B) gene, except for two lines that expressed p15(INK4B) protein, which was induced further upon TGF-beta treatment. Three lines mildly induced CdkI p21(WAF1) expression in response to TGF-beta. Most tumor lines retained other TGF-beta-mediated responses, including extracellular matrix protein and angiogenic factor secretion, which may contribute to increased malignant behavior. This suggests that the loss of p15(INK4B) may explain, in part, the selective loss of growth inhibition by TGF-beta in gliomas to form a more aggressive tumor phenotype.

The effect of L-amino acid oxidase on activity of melphalan against an intracranial xenograft.

We have previously shown that diet restriction-induced depletion of large neutral amino acids (LNAAs) in murine plasma to 46% of control significantly enhances intracranial delivery of melphalan without enhancing delivery to other organs. Studies have now been conducted to determine whether more substantial LNAA depletion could further enhance intracranial delivery of melphalan. Treatment with L-amino acid oxidase (LOX) significantly depleted murine plasma LNAAs: phenylalanine, leucine, and tyrosine (> 95%); methionine (83%); isoleucine (70%); and valine (46%). Experiments evaluating the intracellular uptake of melphalan and high-pressure liquid chromatography quantitation of melphalan metabolites revealed, however, that melphalan is rapidly degraded in the presence of LOX, and that the timing of the administration of melphalan following the use of LOX to deplete LNAAs is crucial. Conditions were found under which LOX-mediated degradation of melphalan was minimized and LNAA depletion was maximized, resulting in a potentiation of the antitumor effect of melphalan on human glioma xenografts in nude mice. Such potentiation could not be obtained using diet restriction alone.