Analysis of cell-free fetal DNA for non-invasive prenatal diagnosis in a family with neonatal diabetes.

AIMS: An early genetic diagnosis of neonatal diabetes guides clinical management and results in improved treatment in ~ 40% of patients. In the offspring of individuals with neonatal diabetes, a prenatal diagnosis allows accurate estimation of the risk of developing diabetes and, eventually, the most appropriate treatment for the baby. In this study, we performed non-invasive prenatal genetic testing for a fetus at risk of inheriting a paternal KCNJ11 p.R201C mutation causing permanent neonatal diabetes. METHODS: A droplet digital polymerase chain reaction assay was used to detect the presence of the mutation in cell-free circulating DNA (cfDNA) extracted from maternal plasma at 12 and 16 weeks' gestation. RESULTS: The mutation was not detected in the cfDNA samples, suggesting that the fetus had not inherited the KCNJ11 mutation. The fetal DNA fraction was estimated at 6.2% and 10.7%, which is above the detection limit of the assay. The result was confirmed by Sanger sequencing after the baby's birth, confirming that the baby's risk of developing neonatal diabetes was reduced to that of the general population. CONCLUSIONS: We report the first case of non-invasive prenatal testing in a family with neonatal diabetes. A prenatal diagnosis in families at high risk of monogenic diabetes informs both prenatal and postnatal management. Although the clinical impact of this novel technology still needs to be assessed, its implementation in clinical practice (including cases at risk of inheriting mutations from the mother) will likely have a positive impact upon the clinical management of families affected by monogenic diabetes.

Improved genetic testing for monogenic diabetes using targeted next-generation sequencing.

AIMS/HYPOTHESIS: Current genetic tests for diagnosing monogenic diabetes rely on selection of the appropriate gene for analysis according to the patient's phenotype. Next-generation sequencing enables the simultaneous analysis of multiple genes in a single test. Our aim was to develop a targeted next-generation sequencing assay to detect mutations in all known MODY and neonatal diabetes genes. METHODS: We selected 29 genes in which mutations have been reported to cause neonatal diabetes, MODY, maternally inherited diabetes and deafness (MIDD) or familial partial lipodystrophy (FPLD). An exon-capture assay was designed to include coding regions and splice sites. A total of 114 patient samples were tested--32 with known mutations and 82 previously tested for MODY (n = 33) or neonatal diabetes (n = 49) but in whom a mutation had not been found. Sequence data were analysed for the presence of base substitutions, small insertions or deletions (indels) and exonic deletions or duplications. RESULTS: In the 32 positive controls we detected all previously identified variants (34 mutations and 36 polymorphisms), including 55 base substitutions, ten small insertions or deletions and five partial/whole gene deletions/duplications. Previously unidentified mutations were found in five patients with MODY (15%) and nine with neonatal diabetes (18%). Most of these patients (12/14) had mutations in genes that had not previously been tested. CONCLUSIONS/INTERPRETATION: Our novel targeted next-generation sequencing assay provides a highly sensitive method for simultaneous analysis of all monogenic diabetes genes. This single test can detect mutations previously identified by Sanger sequencing or multiplex ligation-dependent probe amplification dosage analysis. The increased number of genes tested led to a higher mutation detection rate.

SLC2A2 mutations can cause neonatal diabetes, suggesting GLUT2 may have a role in human insulin secretion.

AIMS: The gene SLC2A2 encodes GLUT2, which is found predominantly in pancreas, liver, kidney and intestine. In mice, GLUT2 is the major glucose transporter into pancreatic beta cells, and biallelic Slc2a2 inactivation causes lethal neonatal diabetes. The role of GLUT2 in human beta cells is controversial, and biallelic SLC2A2 mutations cause Fanconi-Bickel syndrome (FBS), with diabetes rarely reported. We investigated the potential role of GLUT2 in the neonatal period by testing whether SLC2A2 mutations can present with neonatal diabetes before the clinical features of FBS appear. METHODS: We studied SLC2A2 in patients with transient neonatal diabetes mellitus (TNDM; n = 25) or permanent neonatal diabetes mellitus (PNDM; n = 79) in whom we had excluded the common genetic causes of neonatal diabetes, using a combined approach of sequencing and homozygosity mapping. RESULTS: Of 104 patients, five (5%) were found to have homozygous SLC2A2 mutations, including four novel mutations (S203R, M376R, c.963+1G>A, F114LfsX16). Four out of five patients with SLC2A2 mutations presented with isolated diabetes and later developed features of FBS. Four out of five patients had TNDM (16% of our TNDM cohort of unknown aetiology). One patient with PNDM remains on insulin at 28 months. CONCLUSIONS: SLC2A2 mutations are an autosomal recessive cause of neonatal diabetes that should be considered in consanguineous families or those with TNDM, after excluding common causes, even in the absence of features of FBS. The finding that patients with homozygous SLC2A2 mutations can have neonatal diabetes supports a role for GLUT2 in the human beta cell.

Ceramide-induced G2 arrest in rhabdomyosarcoma (RMS) cells requires p21Cip1/Waf1 induction and is prevented by MDM2 overexpression.

The sphingoplipid ceramide is responsible for a diverse range of biochemical and cellular responses including a putative role in modulating cell cycle progression. Herein, we describe that an accumulation of ceramide, achieved through the exogenous application of C(6)-ceramide or exposure to sphingomyelinase, induces a G(2) arrest in Rhabdomyosarcoma (RMS) cell lines. Utilizing the RMS cell line RD, we show that this G(2) arrest required the rapid induction of p21(Cip1/Waf1) independent of DNA damage. This was followed at later time points (48 h) by the commitment to apoptosis. Apoptosis was prevented by Bcl-2 overexpression, but permitted the maintenance of elevated p21(Cip1/Waf1) protein expression and the stabilization of the G(2) arrest response. Inhibition of p21(Cip1/Waf1) protein synthesis with cyclohexamide (CHX) or silencing of p21(Cip1/Waf1) with siRNA, prevented ceramide-mediated G(2) arrest and the late induction of apoptosis. Further, adopting the recent discovery that murine double minute 2 (MDM2) controls p21(Cip1/Waf1) expression by presenting this CDK inhibitor to the proteasome for degradation, RD cells overexpressing MDM2 abrogated ceramide-mediated p21(Cip1/Waf1) induction, G(2) arrest and the late ensuing apoptosis. Collectively, these data further support the notion that ceramide accumulation can modulate cell cycle progression. Additionally, these observations highlight MDM2 expression and proteasomal activity as key determinants of the cellular response to ceramide accumulation.

Growth and characterization of childhood rhabdomyosarcomas as xenografts.

Eleven surgical specimens of childhood rhabdomyosarcoma (RMS), two bone marrow samples, and cells from ascitic fluid from 1 patient were implanted sc into immune-deprived inbred CBA/CaJ mice. From these, seven lines of RMS were obtained as xenografts, each retaining the histologic characteristics of the tumor of origin and human lactate dehydrogenase isozymes; these represented 6 of the 11 surgical specimens, whereas one line originated from the transplanted cells from ascitic fluid. It was concluded that children RMS has a fairly high frequency of heterotransplantability in this system and that such a laboratory model may prove useful in the development of new therapeutic regimens in this disease.

Relationship between topotecan systemic exposure and tumor response in human neuroblastoma xenografts.

BACKGROUND: Topotecan is a topoisomerase I inhibitor with activity against xenografts of childhood solid tumors and established clinical activity against neuroblastoma and rhabdomyosarcoma. We have studied the relationship between systemic exposure to and the antitumor activity of topotecan lactone (the active form of the drug) in the xenograft models. Furthermore, we determined whether the responses seen in these models occur at systemic exposure levels that are tolerable in children. METHODS: Neuroblastoma xenografts derived from the tumors of six different patients were established subcutaneously in immune-deprived mice. Topotecan was administered by intravenous bolus injection 5 days a week for 2 consecutive weeks, repeated every 21 days for three cycles. The minimum daily doses that induced complete responses (CRs) and partial responses (PRs) were determined. Topotecan lactone pharmacokinetic studies were performed in both tumor-bearing and nontumor-bearing mice. RESULTS: The minimum doses associated with CRs and PRs in four of the six neuroblastoma xenografts were 0.61 and 0.36 mg/kg body weight, respectively. The topotecan lactone single-day systemic exposures associated with these doses were 88 and 52 ng x hr/mL, respectively. There was an approximately sixfold difference in topotecan lactone systemic exposure (290 ng x hr/mL versus 52 ng x hr/mL) associated with achieving CRs in the least-sensitive and most-sensitive tumors, respectively. CONCLUSIONS: Neuroblastoma xenografts are highly sensitive to topotecan therapy, and responses in mice are achieved at systemic exposures similar to those that are clinically effective and tolerable in children. These results support the concept of deriving preclinical data relating systemic exposure to antitumor activity in xenograft models. Such data may be valuable in making informed decisions regarding the clinical development of new agents.

Chemotherapy of childhood rhabdomyosarcomas growing as xenografts in immune-deprived mice.

Xenografts derived from the neoplastic tissues of children with rhabdomyosarcoma have been used in immune-deprived mice to examine the efficacy of agents known to be active against this disease, and in others that received either limited or no clinical evaluation. Two models were derived; xenografts were established from tumors obtained from either (a) untreated patients or (b) from patients who had become refractory to conventional therapy. Model a identified as being effective each of these clinically used agents: vincristine, dactinomycin, cyclophosphamide, and doxorubicin; mitomycin C and 5-(3,3-dimethyl-1-triazeno)-2-methylimidazole-4-carboxamide also showed activity, as did busulfan in one tumor line. Tumors derived from refractory patients were significantly less responsive to all agents examined.

Reciprocal cross-resistance in human rhabdomyosarcomas selected in vivo for primary resistance to vincristine and L-phenylalanine mustard.

Primary resistance to vincristine (VCR) has been selected in rhabdomyosarcoma xenograft HxRh12 by sequential administration of VCR at 1.5 and subsequently 3 mg/kg/passage. The resistant tumor (HxRh12/VCR-3) was approximately 4-fold resistant to VCR and resistance was stable in the absence of selecting pressure (greater than 2 yr). HxRh12/VCR-3 was 2- to 3-fold cross-resistant to L-phenylalanine mustard (L-PAM) but only slightly cross-resistant to ifosfamide. To determine whether selection for primary resistance to L-PAM conferred cross-resistance to VCR we selected an L-PAM-resistant subline of rhabdomyosarcoma xenograft HxRh28 (HxRh28/L-PAM-13). This tumor was 2- to 3-fold resistant to L-PAM and 3-(p-fluorophenyl)-L-alanyl-3-[m-bis-(2-chloroethyl)-aminophenyl]-L- alanyl-L-methionine ethoxyhydrochloride, cross-resistant to cyclophosphamide and ifosfamide, and completely resistant to VCR under in vivo conditions. Pharmacokinetic studies in HxRh12/VCR-3 showed decreased retention of [G-3H]VCR but not alteration in metabolism. Expression of mdr1, a gene that encodes P-glycoprotein, associated with the multiple drug resistance phenotype, was examined. Expression of mdr1 was detected in both HxRh12 and HxRh28 tumors, sensitive to VCR, but there was no increase in expression in tumors selected for primary resistance to VCR or L-PAM. Data suggest that mechanisms other than those associated with "classical" multiple drug resistance confer resistance in these tumors. In clinical evaluation against childhood rhabdomyosarcoma, L-PAM has demonstrated only slight activity in patients relapsing on conventional therapy (including VCR) but demonstrated marked activity in patients with advanced previously untreated disease. It appears likely, therefore, that cross-resistance between VCR and L-PAM as demonstrated in this model may have clinical significance.

Cell cycle control processes determine cytostasis or cytotoxicity in thymineless death of colon cancer cells.

Thymidylate synthase (TS) is a target of critical importance to the survival of human colon cancer cells since, upon inhibition, cells subsequently undergo thymineless death induced by dTTP deficiency. Using genetically marked mutants deficient in TS (TS-) and a derived population (Thy4) that is resistant to commitment to thymineless death, resistance was conferred by the ability of cells to arrest at a point either in late G1 or at the onset of S induced by dThd deprivation. Thus, Thy4 cells initially synchronized in G0 by leucine deprivation and released in the absence of dThd remained viable at 5 days, demonstrated delayed onset of nucleosomal ladder formation, and retained clonogenic potential (cytostatic response). In contrast, TS- and asynchronous Thy4 cells lost 50% clonogenic potential in 65 h and > 90% in 5 days (cytotoxic response). [3H]DNA precursor studies indicated failure of synchronized Thy4 but not TS- cells to progress through S, with arrest of Thy4 close to the G1/S boundary. Cell cycle control processes including: (a) the locus of dThd deprivation in G1; and (b) a potential checkpoint close to the G1/S border, may dictate whether consequences of dThd or dTTP restriction become cytostatic or cytotoxic.

Stability of vincristine complexes in cytosols derived from xenografts of human rhabdomyosarcoma and normal tissues of the mouse.

The selective action of vincristine (VCR) has been correlated with longer retention of the drug in neoplastic tissue compared with normal tissues of the mouse (J. A. Houghton, L. G. Williams, P. M. Torrance, and P. J. Houghton, Cancer Res., 44: 582-590, 1984). In order to examine the basis for this differential, the stability of drug-protein complexes was examined further. The stability of drug-protein complexes formed in cytosols derived from HxRh18 tumors, ileum, liver, kidney, skeletal muscle, blood, brain, spleen, lung, and bone marrow was examined. Protein-bound [3H]VCR was isolated by gel filtration of [3H]VCR-cytosol mixtures from each tissue except for ileum and blood. Complexes formed in brain and HxRh18 cytosols were stable at 37 degrees for at least 2 h; all other complexes were unstable. For liver, kidney, and muscle, half-times of complexes were in a similar order to the initial rates of elimination of [3H]VCR from these tissues in vivo but were of shorter duration. The HxRh18-[3H]-VCR complex was unstable at 37 degrees in the presence of cytosols prepared from ileum, kidney, liver, and lung. Drug metabolism by these tissues was not detected in vitro. In the presence of heat-treated extracts from ileum or kidney, [3H]VCR complex was stable, suggesting that the destabilizing factor may be enzymic. Degradation of 125iodinated tubulin, analyzed by polyacrylamide-sodium dodecyl sulfate gel electrophoresis, occurred in the presence of ileum but not skeletal muscle or brain cytosols. This correlated with the stability of HxRh18-[3H]VCR complexes. In the presence of kidney cytosol, however, the molecular weight of 125I-tubulin remained unchanged, suggesting a different mechanism. Based upon data obtained, cytosols from normal tissues may be categorized into three classes: (a) those that formed stable complexes (brain); (b) those that formed unstable complexes but also destabilized preformed complex (ileum, kidney, liver, lung); and (c) tissues that formed unstable complexes but did not destabilized preformed complex (skeletal muscle, spleen, bone marrow, blood). The degree of instability of complexes formed in cytosols prepared from normal tissues appears to correlate with rapid loss of VCR from these tissues in vivo and hence may represent mechanism(s) for the selective action of this antineoplastic agent.

Determinants of intrinsic sensitivity to Vinca alkaloids in xenografts of pediatric rhabdomyosarcomas.

The determinants of intrinsic sensitivity to Vinca alkaloids in vivo were examined in 3 pediatric rhabdomyosarcoma xenografts maintained s.c. in immune-deprived mice. The three lines differed in their sensitivity to VCR and VLB: two lines (Rh12 and Rh28) were extremely sensitive to VCR, whereas Rh18 tumors were less sensitive. Rh28 tumors were also very responsive to VLB, which demonstrated only marginal activity in the other two lines. After administration of equimolar doses (3 mg/kg) of [3H]-VCR and [3H]VLB to tumor-bearing mice, [3H]VCR reached concentrations approaching 1.5 microM in cell water of each tumor line within 4 hr, at which time greater than 93% of the drug was cell-associated. The drug was subsequently retained at this level for at least 72 hr studied. [3H]VLB accumulated to lower maximal concentrations (approximately equal to 1 microM) within 8 hr, but was not retained and, by 72 hr, reached concentrations that were 3- to 4-fold lower than those of [3H]VCR. The extent of drug retention correlated with the antitumor activity except in Rh28 tumors, which were sensitive to VLB, but did not retain the drug. The threshold level for achieving cytotoxicity may, thus, be very low in this line. In normal tissues, maximal concentrations of both [3H]VCR and [3H]VLB were achieved within 1 hr of administration i.p. to tumor-bearing mice. In ileum, liver, and kidney, these were approximately 10-fold higher than the peak levels achieved within tumors or plasma, but declined rapidly to parallel the decrease in plasma reaching concentrations greater than 5-fold lower than the concentration of [3H]VCR in tumors at 72 hr after treatment. Drug concentrations in skeletal muscle also declined rapidly, whereas neither [3H] VCR nor [3H]VLB accumulated to any great extent in brain. The blood volumes of ileum, kidney, and liver were greater than for tumor tissues. Hence, the extent of drug delivery did not necessarily influence therapeutic selectivity. In the case of [3H]VLB, concentrations in tumors approached those of normal tissues at 72 hr after injection. At 24 hr after treatment, 86 to 99% of [3H] VCR and 78 to 90% of [3H]VLB were present in tumors as the parent compound, which also predominated in normal tissues. Metabolites or in vivo degradation products were also identified. Selective retention in tumors appears to be the mechanism by which therapeutic selectivity is achieved with VCR in rhabdomyosarcoma xenografts.(ABSTRACT TRUNCATED AT 400 WORDS)

Characterization of the pools of 5,10-methylenetetrahydrofolates and tetrahydrofolates in xenografts of human colon adenocarcinoma.

The method for measuring polyglutamate forms of CH2-H4PteGlu and H4PteGlu, by entrapment in ternary complexes with [6-3H]5-fluoro-2'-deoxyuridylate and Lactobacillus casei thymidylate synthase has been characterized. Results demonstrated that (a) the relationship between concentration of CH2-H4PteGlu and complex isolated on nondenaturing gels was dependent upon the number of glutamyl residues, and an alternative method for data analysis has been presented, (b) the relationship was linear over a 100-fold change in concentration, (c) formation of isolatable complex was time dependent, (d) noncovalent complexes formed with PteGlu2-5 could be isolated only at concentrations considerably higher than those required for CH2-H4PteGlu1-6, and (e) endogenous deoxyuridylate would be unlikely to interfere significantly with the assay. The distribution of polyglutamates of CH2-H4PteGlu and the combined pools of CH2-H4PteGlu plus H4PteGlu were subsequently examined in three human colon adenocarcinoma xenografts. In each tumor, the pentaglutamate of CH2-H4PteGlu and H4PteGlu was the most prominent species, followed by the hexaglutamate, constituting 68 to 92% of the CH2-H4PteGlu pool, and greater than 93% of the combined pools. A small percentage of di-, tri-, and tetraglutamates were also detected. Using a catalytic assay, the combined pool of CH2-H4PteGlu and H4PteGlu was estimated in the range of 0.5 to 2.7 microM in cell water, and for CH2-H4PteGlu, from 185 nM to 1.7 microM. Using thymidylate synthase purified from colon adenocarcinoma HxVRC5, CH2-H4PteGlu5 (where the subscript digit attached to the glutamate portion equals the number of glutamate residues) stabilized the covalent ternary complex at greater than 200-fold lower concentration in comparison to CH2-H4PteGlu1. Data indicated that in each colon tumor, the concentrations of CH2-H4PteGlun or CH2-H4PteGlun plus H4PteGlun were suboptimal for the interaction of 5-fluoro-2'-deoxyuridylate with thymidylate synthase, and would predict for relatively transient inhibition of thymidylate synthase after treatment with 5-fluorouracil. These data support therapeutic modulation to increase the concentration of CH2-H4PteGlun in the treatment of colon adenocarcinomas with 5-fluorouracil.

Hypoxanthine:guanine phosphoribosyltransferase activity in xenografts of human osteosarcoma.

It has recently been reported that human osteosarcomas may lack the purine salvage pathway enzyme, hypoxanthine:guanine phosphoribosyltransferase (EC 2.4.2.8). We have established a quantitative assay for measurement of this enzyme in human osteosarcoma xenografts with analysis of products by thin-layer chromatography. Nucleotidase or phosphatase activity was readily detected and could be abolished by preheating cytosol at 60 degrees C for 10 min and performing the assay at pH 10. Alternatively, the use of 25 mM NaF at pH 7.4 also inhibited this activity. The pH optimum for this enzyme in red blood cell sonicates and tumor cytosols was pH 10. All six human osteosarcoma xenografts contained hypoxanthine:guanine phosphoribosyltransferase activity ranging from 0.97 to 4.06 nmol/min/mg of protein at pH 7.4. Control human red blood cell sonicates demonstrated activity of 0.83 nmol/min/mg of protein. These data demonstrate that human osteosarcoma xenografts contain substantial activities of this purine salvage pathway enzyme.

In situ selection of a human rhabdomyosarcoma resistant to vincristine with altered beta-tubulins.

In order to simulate more closely conditions in which resistance to vincristine (VCR) is selected in human solid tumors, a human rhabdomyosarcoma grown as a xenograft in immune-deprived mice has been selected for resistance in situ. Karyotype analysis showed the resistant line, HxRh18/VCR-3, to have a diploid modal number, with no apparent translocations, whereas the predominant population in the parental, sensitive HxRh18 xenograft demonstrated a modal number near-tetraploid with many marker chromosomes. From the rapid rate at which resistance was selected and from karyotypic evidence, data strongly suggest that HxRh18/VCR-3 was a subpopulation within the parent tumor. When grown in the same host, HxRh18/VCR-3 tumors accumulated less drug, and the rate of [3H]VCR loss was 5-fold greater than in HxRh18 tumors. Thus, accumulation and retention of [3H]VCR in HxRh18/VCR-3 resistant tumors was identical to that of [3H]vinblastine (VLB) in HxRh18 xenografts. HxRh18 xenografts are intrinsically resistant to VLB. Analysis by high-performance liquid chromatography of [3H]VCR:protein complexes in HxRh18 cytosols indicated one binding species (Mr 95,000 to 116,000), probably the tubulin heterodimer. Of interest was the observation that beta-tubulin species, identified on Western blots by monoclonal antibody, differed in these tumors. In HxRh18/VCR-3, less acidic beta-tubulins of HxRh18 were decreased or absent, with three additional more acidic isoforms present in the resistant line. As vincristine may bind to the beta-subunit of tubulin, this may have importance to vincristine resistance in vivo.

Development and characterization of a human colon adenocarcinoma xenograft deficient in thymidine salvage.

In order to determine the contribution of thymidine (dThd) salvage to intrinsic resistance to antimetabolites (5-fluoropyrimidines, antifolates) in the human colon adenocarcinoma xenograft, H X GC3, a subline deficient in thymidine kinase has been developed. A cell line (GC3/M) was established in continuous culture that demonstrated a karyotype identical to that of the stem line of H X GC3 (46,X, - Y + 12). After inoculation of GC3/M cells into immune-deprived CBA/CaJ mice, the H X GC3/M xenografts retained histological, histochemical, and growth characteristics of the H X GC3 xenograft. To develop a line deficient in dThd salvage, GC3/M cells were selected with BrdUrd (100 micrograms/ml). Three clones characterized were unable to proliferate in HAT medium, and were deficient (less than 10% control) in the cytosolic form of thymidine kinase. Activities of dThd phosphorylase and dTMP synthase were unchanged from parental GC3/M cells. Of the three clones inoculated into mice, GC3/M TK- 100 C3 was tumorigenic, the xenografts demonstrating histological and growth characteristics similar to H X GC3. The in vivo activity of the cytosolic form of dThd kinase was 3.5% of that in H X GC3 xenografts. Incorporation in vivo of [methyl-3H]dThd into acid insoluble material was 14% of that in H X GC3 tumors. Autoradiographs prepared from these tumors demonstrated that incorporation of radiolabel into nuclei occurred only in stromal cells derived from the host. It is anticipated that H X GC3/M TK- 100 C3 will be a line valuable for determining the role of dThd salvage in intrinsic resistance to 5-fluoropyrimidines or antifolates in human colon adenocarcinomas growing as xenografts and also the relevance of dTMp synthase as a target for antimetabolites in this histiotype.

Locus of the interaction among 5-fluorouracil, leucovorin, and interferon-alpha 2a in colon carcinoma cells.

Prior studies from these laboratories demonstrated 3.2-fold potentiation of 5-fluorouracil (FUra) cytotoxicity by recombinant human interferon-alpha 2a (rIFN-alpha 2a) in GC3/cl colon adenocarcinoma cells that was significantly enhanced to 14-fold when FUra was combined with rIFN-alpha 2a + a mixture of the diasteroisomers of the biologically active (6S) and inactive (6R) leucovorin or 5-formyl-H4PteGlu (LV), events that were reversible by thymidine (dThd). In GC3/clTS-c3/c3 cells, deficient in thymidylate synthase, rIFN-alpha 2a cytotoxicity was not influenced by the concentration of dThd, indicating no direct effect at the level of dThd-less stress. Direct assays of thymidylate synthase indicated no significant difference between FUra-induced accumulation of total thymidylate synthase or free or unbound thymidylate synthase in cells receiving FUra + modulators. In addition, the cytotoxic activity of CB3717, a specific quinazoline-based inhibitor of thymidylate synthase, was not potentiated by rIFN-alpha 2a. These studies suggested that thymidylate synthase was not the primary target site for rIFN-alpha 2a activity. Since data indicated that a 5-fluoropyrimidine was required in the interaction among FUra, LV, and rIFN-alpha 2a, attention was focused at the level of DNA. Both DNA single-strand breaks (SSBs) and DNA double-strand breaks (DSBs) induced by FUra were significantly elevated by rIFN-alpha 2a and LV administered as single modulators and were influenced by the concentrations of both FUra and rIFN-alpha 2a. However; when FUra was combined with LV, rIFN-alpha 2a further potentiated the frequency of DNA SSBs, and data correlated with the relative cytotoxic activity of FUra-LV-rIFN-alpha 2a combinations. No effect on CB3717-induced DNA SSBs or DSBs by rIFN-alpha 2a was demonstrated. Drug exposure for 48 h was required to detect measurable differences in DNA SSB frequency among FUra-LV-rIFN-alpha 2a treatment groups and correlated with decreased clonogenic survival under these conditions. Continuous exposure to FUra (72 h) allowed shorter exposures to LV and/or rIFN-alpha 2a (48 h) to maintain maximal cytotoxicity. Shorter exposure times for FUra during continuous exposure to the modulators were less cytotoxic. Data suggest that the primary locus of the interaction among FUra, LV, and rIFN-alpha 2a lies at the level of DNA. rIFN-alpha 2a may exert its effects via enhancement of FUra base excision or incorporation into DNA, events that subsequently become influenced by thymidylate synthase inhibition and dThd-less stress and are further potentiated by LV.(ABSTRACT TRUNCATED AT 400 WORDS)

Biochemical determinants of responsiveness to 5-fluorouracil and its derivatives in xenografts of human colorectal adenocarcinomas in mice.

The level of thymidylate synthetase (EC 2.1.1.45) and its activity have been measured in a series of human colorectal adenocarcinomas growing as xenografts in immune-deprived mice. Enzyme activity varied between 8.4 and 124 pmol per mg protein per hr; within each tumor line, this activity correlated with the capacity to bind [6-3H]5-fluoro-2'-deoxyuridine 5'-monophosphate ([6-3H]FdUMP), which varied between 0.16 and 1.68 pmol [6-3H]FdUMP bound per g tissue. Highest and lowest activities were measured in tumor lines that were insensitive to 5-fluorouracil, 5-fluorouridine, and 5-fluoro-2'-deoxyuridine. The ratio of the maximum free FdUMP concentration to thymidine 5'-monophosphate synthetase-binding activity did not differentiate fluorinated pyrimidine-responsive lines from those innately insensitive. Maximum potential binding of [6-3H]FdUMP in vitro was measured without addition of dl-L-5,10-methylenetetrahydrofolate (CH2FH4) in cytosol from two tumor lines, both of which demonstrated some sensitivity to fluorinated pyrimidine therapy. The other four insensitive tumor lines required CH2FH4 to be added in order to attain maximum [6-3H]FdUMP binding. Similar data were obtained using nitrocellulose membrane filtration to isolate both covalent and noncovalent complexes. Direct measurement of thymidine 5'-monophosphate synthetase activity after incubation of tumor cytosols with FdUMP, with or without added CH2FH4 showed that, in nonresponsive tumors, maximum enzyme inhibition was achieved only in the presence of exogenous cofactor. It is suggested that the availability of cofactor may prove important in the formation of the ternary complex CH2FH4:thymidine 5'-monophosphate synthetase:FdUMP when high concentrations of FdUMP are present for only short periods of time.

Radiosensitivity of thymidylate synthase-deficient human tumor cells is affected by progression through the G1 restriction point into S-phase: implications for fluoropyrimidine radiosensitization.

Recent studies of fluoropyrimidine (FP)-mediated radiosensitization (RS) have focused on the molecular mechanisms underlying regulation of the cell cycle, particularly at the G1-S transition. Although thymidylate synthase (TS) inhibition by FP is necessary, we hypothesize that FP-RS is temporally dependent on progression of cells into S-phase under conditions of altered deoxynucleotide triphosphate pools, particularly an increased dATP:dTTP ratio, which subsequently results in enhanced DNA fragmentation and cell death. To better understand the mechanism of FP-RS, we characterized the cellular and biochemical responses to ionizing radiation (IR) alone, using different synchronization techniques in two isogenic, TS-deficient mutant cell lines, JH-1 (TS-) and JH-2 (Thy4), derived previously from a human colon cancer cell line. After G0 synchronization by leucine deprivation, these clones differ under subsequent growth conditions and dThd withdrawal: JH-2 cells have an intact G1 arrest (>72 h) and delayed cell death (>96 h), whereas JH-1 cells progress rapidly into early S-phase and undergo acute cell death (<24 h). No difference in the late S-phase and G2-M cell populations were noted between these growth-stimulated, G0-synchronized TS-deficient cell lines with dThd withdrawal. Biochemically, the intracellular ratio of dATP:dTTP increased substantially in JH-1 cells as cells progressed into early S-phase compared with JH-2 cells, which remained in G1 phase. Synchronized JH-1 cells showed significantly decreased clonogenic survival and an increase in DNA fragmentation after IR when compared with JH-2 cells. RS was demonstrated by an increase in alpha and decrease in beta, using linear quadratic analyses. An alternative synchronization technique used mimosine to induce a block in late G1, close to G1-S border. Both JH-1 and JH-2 cells, synchronized in late G1 and following growth stimulation, now progressed into S-phase identically (<24 h), with similarly increased dATP:dTTP ratios under dThd withdrawal conditions. These late G1-synchronized JH-1 and JH-2 cells also showed a comparable reduction in clonogenic survival and similar patterns of increased DNA fragmentation following IR. We suggest, based on the cellular and biochemical differences in response to IR between G0- and late G1-synchronized cells, that S-phase progression through the G1 restriction point under an altered (increased) dATP:dTTP ratio is a major determinant of FP-RS.

Growth and differentiation of a human T-cell leukemia cell line, CCRF-CEM, grafted in mice.

The growth of human CCRF-CEM T-cell lymphoblastic leukemia was studied in mice immune deprived by different techniques, and in CD-nu/nu athymic mice. Female CBA/CaJ mice were immune deprived by infant thymectomy, priming with 1-beta-D-arabinofuranosylcytosine (200 mg/kg) 48 h prior to total body irradiation (925 cGy) designated theta ara-C gamma; or after thymectomy the mice received 925 cGy total body irradiation with marrow reconstitution (4 x 10(6) nucleated cells), designated theta gamma BM. Only in mice immune deprived by theta gamma BM, subsequently given a single dose of cyclophosphamide (100 mg/kg) 18-24 h before transplantation of CCRF-CEM, was there progressive reproducible engraftment and tumor growth. For mice immune deprived in this manner the tumor engraftment rate was 100 and 80% of tumors achieved greater than or equal to 1 cm3 within 46 days. In immune-deprived CBA/CaJ mice, but not CD-nu/nu athymic mice, tumor transplanted to the s.c. site metastasized to paraaortic and axillary nodes. Metastatic spread to lymph nodes was confirmed by immunophenotyping and by karyotyping. In contrast to the CCRF-CEM cells in culture, which expressed cytoplasmic CD3 (T3) but not surface CD3, both s.c. and metastatic CCRF-CEM line was exposed to phorbol-12-myristate 13-acetate in vitro to mimic the apparent differentiation which occurred in the xenografted cells, and a similar expression of surface CD3 after treatment was seen. This surface expression of CD3 was accompanied by production of mRNA for the T-cell receptor alpha chain and surface expression of the T-cell receptor. Identical T-cell receptor beta and gamma chain gene rearrangements were found for the CCRF-CEM line in vitro and the xenografted cells in vivo, demonstrating that only one clone was present and that differences in immunophenotyping were not the result of clonal selection. These results suggest that host (mouse) hematopoietic factors could affect human leukemic cell differentiation.

Mutation(s) of the thymidylate synthase gene of human adenocarcinoma cells causes a thymidylate synthase-negative phenotype that can be attenuated by exogenous folates.

Biological characterization of a human colon adenocarcinoma cell line deficient in thymidylate synthase (TS-) is described. The clone, designated TS-C1/C1, was derived from the parental line GC3/C1 by selection in medium containing aminopterin, thymidine (dThd), and low concentrations of 5-formyltetrahydrofolate (5-CHO-H4PteGlu), and was subsequently reselected by single-step cloning in 500 microM methotrexate in the presence of dThd. This clone retained its TS- phenotype, was highly resistant to methotrexate (greater than 100,000-fold), and remained tumorigenic in mice (P.J. Houghton, et al., Proc. Natl. Acad. Sci. USA, 86: 1377-1381, 1989). In studies reported, it is shown that high levels of exogenous folate can support the growth of the TS- C1/C1 clone in the absence of dThd. Activation of dTMP biosynthesis de novo was demonstrated within 6 h of exposing cells to 20 microM [6R,S]5-CHO-H4PteGlu, and greater than or equal to 80% of activity was lost within 24 h of removing this folate from the medium. The labeling index was determined by autoradiographic techniques using [6-3H]2'-deoxyuridine. None of the greater than 6,000 cells radiolabeled in the absence of [6R,S]5-CHO-H4PteGlu, whereas 33.5% labeled in the presence of 20 microM exogenous folate. Relative to the parental (TS+) clone, there was a greater than 87,500-, an 8,182-, and a 425-fold higher requirement for 5-methyltetrahydrofolate ([6R,S]5-CH3-H4PteGlu), PteGlu, and [6R,S]5-CH3-H4PteGlu to support 50% maximal colony formation in the absence of dThd. Quantitative analysis of the combined pools of 5,10-methylenetetrahydrofolate (CH2-H4PteGlun) and H4PteGlun showed that parental GC3/C1 cells had higher endogenous folate pools compared to TS-C1/C1 cells [168 +/- 40 (SD) and 10.9 +/- 0.3 fmol/10(6) cells, respectively]. Qualitatively the distribution of polyglutamate species and their redistribution in cells exposed to 20 microM [6R,S]5-CHO-H4PteGlu were similar in the two lines. Analysis of pools in a second, independently derived, TS- clone (TS-C3/C3, a transcription-negative mutant) demonstrated undetectable levels of CH2-H4PteGlun and H4PteGlun. This line cannot be rescued by exogenous folate. The data thus suggest that deletion of dTMP synthase activity may cause redistribution of reduced folate pools. In cytosolic extracts from parental GC3/C1 (TS+) cells, [6R]CH2-H4PteGlu1 acted as a cofactor in the release of 3H2O from [5-3H]dUMP, whereas no activity was detected in cytosols from TS-C1/C1. In contrast dTMP synthase activity was detected in cytosols from TS- C1/C1 cells in the presence of [6R]CH2-H4PteGlu5.(ABSTRACT TRUNCATED AT 400 WORDS)