Microsatellite instability in cancer of the proximal colon.

Colorectal tumor DNA was examined for somatic instability at (CA)n repeats on human chromosomes 5q, 15q, 17p, and 18q. Differences between tumor and normal DNA were detected in 25 of the 90 (28 percent) tumors examined. This instability appeared as either a substantial change in repeat length (often heterogeneous in nature) or a minor change (typically two base pairs). Microsatellite instability was significantly correlated with the tumor's location in the proximal colon (P = 0.003), with increased patient survival (P = 0.02), and, inversely, with loss of heterozygosity for chromosomes 5q, 17p, and 18q. These data suggest that some colorectal cancers may arise through a mechanism that does not necessarily involve loss of heterozygosity.

Flying in the face of resistance: antiviral-independent benefit of HIV protease inhibitors on T-cell survival.

Human immunodeficiency virus (HIV) infection results in excessive apoptosis of infected and uninfected cells, mediated by host and viral factors present in plasma. As HIV protease inhibitors (PIs) have intrinsic antiapoptotic properties, we questioned whether HIV PIs could block HIV-induced CD4+ T-cell death independent of their effects on HIV replication. We demonstrate that HIV PIs block the death of CD4+ T cells induced by HIV glycoprotein 120 (gp120), Vpr, and Tat, as well as host signals Fas ligand, tumor necrosis factor, and tumor necrosis factor-related apoptosis-inducing ligand. Using gp120/CXCR4 as a model, we show that the HIV PIs specifically block mitochondrial apoptosis signaling. Furthermore, HIV PIs inhibit CD4+ T-cell death induced by viruses with high-level resistance to PIs (P<0.01) and apoptosis induced by serum of HIV patients with known resistance to HIV PIs (P=0.01). Together, these results show that HIV PIs block CD4+ T-cell death and have a beneficial effect on CD4+ T-cell survival despite PI resistance.

IkappaB kinase-dependent chronic activation of NF-kappaB is necessary for p21(WAF1/Cip1) inhibition of differentiation-induced apoptosis of monocytes.

The molecular mechanisms regulating monocyte differentiation to macrophages remain unknown. Although the transcription factor NF-kappaB participates in multiple cell functions, its role in cell differentiation is ill defined. Since differentiated macrophages, in contrast to cycling monocytes, contain significant levels of NF-kappaB in the nuclei, we questioned whether this transcription factor is involved in macrophage differentiation. Phorbol 12-myristate 13-acetate (PMA)-induced differentiation of the promonocytic cell line U937 leads to persistent NF-kappaB nuclear translocation. We demonstrate here that an increased and persistent IKK activity correlates with monocyte differentiation leading to persistent NF-kappaB activation secondary to increased IkappaBalpha degradation via the IkappaB signal response domain (SRD). Promonocytic cells stably overexpressing an IkappaBalpha transgene containing SRD mutations fail to activate NF-kappaB and subsequently fail to survive the PMA-induced macrophage differentiation program. The differentiation-induced apoptosis was found to be dependent on tumor necrosis factor alpha. The protective effect of NF-kappaB is mediated through p21(WAF1/Cip1), since this protein was found to be regulated in an NF-kappaB-dependent manner and to confer survival features during macrophage differentiation. Therefore, NF-kappaB plays a key role in cell differentiation by conferring cell survival that in the case of macrophages is mediated through p21(WAF1/Cip1).

Casein kinase II phosphorylates I kappa B alpha at S-283, S-289, S-293, and T-291 and is required for its degradation.

The phosphoprotein I kappa B alpha exists in the cytoplasm of resting cells bound to the ubiquitous transcription factor NF-kappa B (p50-p65). In response to specific cellular stimulation, I kappa B alpha is further phosphorylated and subsequently degraded, allowing NF-kappa B to translocate to the nucleus and transactivate target genes. To identify the kinase(s) involved in I kappa B alpha phosphorylation, we first performed an I kappa B alpha in-gel kinase assay. Two kinase activities of 35 and 42 kDa were identified in cellular extracts from Jurkat T and U937 promonocytic cell lines. Specific inhibitors and immunodepletion studies identified the I kappa B alpha kinase activities as those of the alpha and alpha' subunits of casein kinase II (CKII). Immunoprecipitation studies demonstrated that CKII and I kappa B alpha physically associate in vivo. Moreover, phosphopeptide maps of I kappa B alpha phosphorylated in vitro by cellular extracts and in vivo in resting Jurkat T cells contained the same pattern of phosphopeptides as observed in maps of I kappa B alpha phosphorylated in vitro by purified CKII. Sequence analysis revealed that purified CKII and the kinase activity within cell extracts phosphorylated I kappa B alpha at its C terminus at S-283, S-288, S-293, and T-291. The functional role of CKII was tested in an in vitro I kappa B alpha degradation assay with extracts from uninfected and human immunodeficiency virus (HIV)-infected U937 cells. Immunodepletion of CKII from these extracts abrogated both the basal and enhanced HIV-induced degradation of I kappa B alpha. These studies provide new evidence that the protein kinase CKII physically associates with I kappa B alpha in vivo, induces multisite (serine/threonine) phosphorylation, and is required for the basal and HIV-induced degradation of I kappa B alpha in vitro.

Activation of IkappaB kinase beta by protein kinase C isoforms.

The atypical protein kinase C (PKC) isotypes (lambda/iotaPKC and zetaPKC) have been shown to be critically involved in important cell functions such as proliferation and survival. Previous studies have demonstrated that the atypical PKCs are stimulated by tumor necrosis factor alpha (TNF-alpha) and are required for the activation of NF-kappaB by this cytokine through a mechanism that most probably involves the phosphorylation of IkappaB. The inability of these PKC isotypes to directly phosphorylate IkappaB led to the hypothesis that zetaPKC may use a putative IkappaB kinase to functionally inactivate IkappaB. Recently several groups have molecularly characterized and cloned two IkappaB kinases (IKKalpha and IKKbeta) which phosphorylate the residues in the IkappaB molecule that serve to target it for ubiquitination and degradation. In this study we have addressed the possibility that different PKCs may control NF-kappaB through the activation of the IKKs. We report here that alphaPKC as well as the atypical PKCs bind to the IKKs in vitro and in vivo. In addition, overexpression of zetaPKC positively modulates IKKbeta activity but not that of IKKalpha, whereas the transfection of a zetaPKC dominant negative mutant severely impairs the activation of IKKbeta but not IKKalpha in TNF-alpha-stimulated cells. We also show that cell stimulation with phorbol 12-myristate 13-acetate activates IKKbeta, which is entirely dependent on the activity of alphaPKC but not that of the atypical isoforms. In contrast, the inhibition of alphaPKC does not affect the activation of IKKbeta by TNF-alpha. Interestingly, recombinant active zetaPKC and alphaPKC are able to stimulate in vitro the activity of IKKbeta but not that of IKKalpha. In addition, evidence is presented here that recombinant zetaPKC directly phosphorylates IKKbeta in vitro, involving Ser177 and Ser181. Collectively, these results demonstrate a critical role for the PKC isoforms in the NF-kappaB pathway at the level of IKKbeta activation and IkappaB degradation.

Expression of p53 and 17p allelic loss in colorectal carcinoma.

Mutations in the p53 gene are the most common genetic changes in cancer thus far. Many p53 mutations result in a protein product having a prolonged half-life compared to wild-type p53. The mutant protein is frequently detectable immunohistochemically, whereas the wild-type p53 present in normal cells is not. We examined 90 colorectal carcinomas for increased expression of p53 using 3 p53 specific monoclonal antibodies, PAb1801, PAb421, and PAb240. Overall, 70% of the colorectal carcinomas stained for p53. Each tumor's DNA was also assessed for loss of heterozygosity on chromosome 17p, the location of the p53 gene. Of those tumors that reacted with the anti-p53 antibodies, 76% showed loss on chromosome 17p. Tumors with loss of heterozygosity on 17p generally stained with all 3 antibodies, whereas those without loss tended to stain with just one antibody, typically PAb240. Fifteen tumors were examined for the presence of specific p53 mutations. A total of 10 mutations were found, 6 were missense and 2 were deletions, and all but one of the tumors with missense mutations stained for p53.

Transcription of the RelB gene is regulated by NF-kappaB.

RelA and RelB are two members of the NF-kappaB family that differ structurally and functionally. While RelA is regulated through its cytosolic localization by inhibitor proteins or IkappaB and not through transcriptional mechanisms, the regulation of RelB is poorly understood. In this study we demonstrate that stimuli (TNF or LPS) lead within minutes to the nuclear translocation of RelA, but require hours to result in the nuclear translocation of RelB. The delayed nuclear translocation of RelB correlates with increases in its protein synthesis which are secondary to increases in RelB gene transcription. RelA is alone sufficient to induce RelB gene transcription and to mediate the stimuli-driven increase in RelB transcription. Cloning and characterization of the RelB 5' untranslated gene region indicates that RelB transcription is dependent on a TATA-less promoter containing two NF-kappaB binding sites. One of the NF-kappaB sites is primarily involved in the binding of p50 while the other one in the binding and transactivation by RelA and also RelB. Lastly, it is observed that p21, a protein involved in cell cycle control and oncogenesis known to be regulated by NF-kappaB, is upregulated at the transcriptional level by RelB. Thus, RelB is regulated at least at the level of transcription in a RelA and RelB dependent manner and may exert an important role in p21 regulation.

Efficacy of oral diltiazem to control ventricular response in chronic atrial fibrillation at rest and during exercise.

Although digoxin is often the first choice for control of ventricular response in chronic atrial fibrillation, it fails to slow exercise rates. Diltiazem, a calcium channel antagonist that slows atrioventricular conduction, was administered to 16 patients who failed to achieve adequate rate control on low level exercise testing despite digoxin therapy. Therapeutic response to diltiazem was assessed with submaximal and maximal exercise tests and 24 hour ambulatory electrocardiographic monitoring. During the diltiazem treatment phase, ventricular response at rest diminished (96 +/- 17 versus 69 +/- 10 beats/min, p less than 0.001) as did rate during submaximal exercise (155 +/- 28 versus 116 +/- 26, p less than 0.001), maximal exercise (163 +/- 14 versus 133 +/- 26, p less than 0.001) and average ventricular response during 24 hour monitoring (87 +/- 13 versus 69 +/- 10, p less than 0.001). Rate at rest decreased 26 +/- 15% and submaximal exercise rate diminished 24 +/- 12%. Thirteen (81%) of the 16 patients exhibited at least 15% slowing of rate at rest and during submaximal exercise. Eleven patients (69%) reported alleviation of symptoms. There was no change in serum digoxin levels during diltiazem treatment (1.3 +/- 0.5 versus 1.3 +/- 0.6 ng/ml, p = NS). On withdrawal of diltiazem, ventricular response returned to baseline values. Diltiazem is an effective agent for control of ventricular response, both at rest and during exercise, in digoxin-treated patients with chronic atrial fibrillation.

Independent and interactive effects of digoxin and quinidine on the atrial fibrillation threshold in dogs.

To assess the effects of digoxin as single therapy and in combination with quinidine in the treatment of atrial fibrillation, the atrial fibrillation threshold was determined from the right atrial appendage and Bachmann's bundle in 11 open chest dogs. In group 1 (six dogs), the atrial fibrillation threshold was determined at baseline, post-quinidine (10 mg/kg intravenously) and then post-digoxin (50 micrograms/kg intravenously). In group 2 (five dogs), the order of drug administration was reversed. The results of this study were: 1) Digoxin had no significant effect on the atrial fibrillation threshold when given alone. 2) Quinidine significantly increased the atrial fibrillation threshold (p less than 0.002) and the addition of digoxin resulted in a further increase in threshold (p less than 0.002). 3) Quinidine produced greater suppression of atrial fibrillation induction at the right atrial site than at the Bachmann's bundle site, suggesting differential effects of quinidine on atrial fibers.

Intravenous digital left ventriculography at rest and with atrial pacing as a screening procedure for coronary artery disease.

Digital subtraction left ventriculography using intravenous contrast injection was evaluated as a screening diagnostic method for coronary heart disease. Intravenous ventriculography was performed in 61 patients with 35 cc of contrast medium injected into a central vein (usually the inferior vena cava). Recognition of regional wall motion abnormalities by this technique was shown to be comparable with direct left ventriculography in 40 patients who underwent both imaging modalities at rest. If the rest digital ventriculogram was normal, it was repeated after incremental atrial pacing to the onset of chest pain or to a maximal heart rate of 150 beats/min. Forty-four of the 61 patients had significant coronary artery disease, of whom 10 had a wall motion abnormality at rest on intravenous ventriculography. With pacing, 28 of the 34 remaining patients developed a new wall motion abnormality. Thus, 38 (86%) of 44 patients with coronary heart disease were identified by wall motion abnormalities. One of the 17 patients without coronary artery disease had an abnormal rest study and was incorrectly assigned a diagnosis of coronary disease. Intravenous digital ventriculograms approximate those obtained by direct ventriculography. When combined with atrial pacing they are a sensitive and specific means of detecting coronary artery disease.

Temporal evolution of the human coronary collateral circulation after myocardial infarction.

An analysis of the coronary collateral circulation in a consecutive series of 116 postinfarction angiograms from patients with persistent 100% occlusion of their infarct artery is reported. Patients were classified into four groups according to the interval between acute infarction and angiography. Of 42 patients studied within 6 hours of infarction (Group I), 52% had no evidence of any coronary collateral development as compared with only 8% (1 of 16 patients) studied 1 day to 2 weeks after infarction (Group II). Virtually all patients studied beyond 2 weeks after myocardial infarction (14 to 45 days, Group III) and later than 45 days (Group IV) had visible collateral flow. Angiographically "well developed" collateral channels were seen in only 16% of Group I patients compared with 62, 75 and 84% of patients in Groups II to IV, respectively. Of six patients studied twice, on the day of the infarction and 2 weeks later, only one patient had collateral vessels on the day of infarction, whereas all six patients did at follow-up study. Group I patients were studied as part of a randomized acute myocardial infarction reperfusion trial, whereas the other patients were referred for angiography primarily because of post-infarction ischemia. Within the limitations imposed by the patient selection process, it is concluded that well developed coronary collateral vessels are rarely present at the time of infarction. After infarction, they develop rapidly and are generally demonstrable within 2 weeks. It may also be inferred that the preservation of ischemic myocardium by well developed coronary collateral vessels at the time of myocardial infarction may be an uncommon occurrence.

PKC-zeta-associated CK2 participates in the turnover of free IkappaBalpha.

The atypical PKC isoenzymes, zeta and iota, activate NF-kappaB, a mechanism thought to mediate the anti-apoptotic and proliferative features of these kinases. PKC-zeta has been shown to be associated with an IkappaBalpha kinase in resting cells. In this study, we have sought to identify the PKC-zeta associated kinase and understand how PKC-zeta mediates basal IkappaBalpha turnover in vivo. We demonstrate that the PKC-zeta-associated IkappaBalpha kinase is CK2. This kinase, previously shown to phosphorylate the PEST domain of IkappaB molecules, co-precipitates with PKC-zeta in resting cells. In vitro, PKC-zeta interacts with CK2-beta. The in vivo PKC-zeta-associated CK2 preferentially phosphorylates S293 of IkappaBalpha as compared to non-associated CK2. The functional relevance of this observation is supported by the fact that the turnover of free IkappaBalpha in resting cells is S293-dependent. Moreover, overexpressing PKC-zeta results in lower steady-state protein levels of free IkappaBalpha, which is dependent on S293. Lastly, it is shown that PKC-zeta wt but not kinase dead leads to the in vitro phosphorylation of both CK2-alpha and beta. These studies demonstrate that the association between CK2 and PKC-zeta may play a major role in the control of the basal turnover of free IkappaBalpha, in the absence of extracellular stimuli.

Serine 32 and serine 36 of IkappaBalpha are directly phosphorylated by protein kinase CKII in vitro.

IkappaBalpha is an inherently unstable protein which binds to and retains the ubiquitous transcription factor NFkappaB in the cytoplasm of resting cells. A continuous low level translocation of NFkappaB to the nucleus, secondary to the basal turnover of IkappaBalpha, is hypothesized to be necessary for cellular maturation, survival and, potentially, transformation. In response to cellular stimulation by inflammatory cytokines or mitogens, IkappaBalpha is rapidly degraded allowing larger pools of NFkappaB to translocate to the nucleus. Phosphorylation of IkappaBalpha at serine 32 (S32) and serine 36 (S36) is necessary for this stimuli-induced degradation. IKKalpha/beta kinases and p90(rsk1)are involved in stimuli-induced targeting of one or both of these IkappaBalpha sites. Whether other kinases phosphorylate S32 and S36 directly, and if so, what function they serve in NFkappaB activation remains unknown. Here we present evidence of a direct phosphorylation of IkappaBalpha at both S32 and S36 by purified or immunoprecipitated protein kinase CKII (PK-CKII) and a specific in vivo association between IkappaBalpha and PK-CKII. This PK-CKII-specific kinase activity is not found within the IKKalpha/beta-containing signalsome complex and is biochemically distinct from that of the IKKalpha/beta kinases. The identification of an additional N-terminal IkappaBalpha kinase which is constitutively active and not significantly inducible raises numerous possibilities as to its role in cellular function.

The location of the Philadelphia chromosomal breakpoint site and prognosis in chronic granulocytic leukemia.

Chronic granulocytic leukemia (CGL) is associated with a reciprocal translocation between chromosomes 9 and 22. The breakpoint sites on chromosome 22 are clustered in a limited region known as the major breakpoint cluster region (Mbcr). This region is approximately 5.8 Kb long and can be arbitrarily subdivided into five zones (1 through 5 from the 5' towards the 3' end) as defined by the particular sites of three restriction endonucleases. Using Southern blot analysis with two DNA probes, one spanning both the 5' and 3' regions of the Mbcr while the other only the 3' region, we mapped the precise location of the chromosomal breakpoints within the Mbcr in 62 patients with CGL and examined possible clinical correlations. There were 39 patients with 5' breakpoints (zones 1-3) and 23 patients with 3' breakpoints (zones 4 and 5). We found no correlation between the clinical phase of the disease at last followup and breakpoint distributions. The distributions of chronic phase duration (CPD) and survival were similar between patients with 5' breakpoints (median CPD = 4.0 years) and those with 3' breakpoints (median CPD = 5.2 years). Presenting clinical features and the rates of lymphoblastic transformation were also similar among the subgroups. Our data suggest that the precise location of the breakpoint within the Mbcr in CGL may not have clinical relevance.

Determinants of transstenotic gradients observed during angioplasty: an experimental model.

Pressure gradient measurement across a stenosis is used during angioplasty to aid catheter positioning and estimate dilatation efficacy. The angioplasty catheter itself, however, further reduces lumen size, and therefore augments the transstenotic gradient. To more precisely define the catheter influence on gradient, we derived a theoretical expression relating the measured gradient with the angioplasty catheter in situ to the "true" gradient; that is, the gradient in the absence of the angioplasty catheter. We then tested this theoretical construct in a canine femoral artery angioplasty model. Fifty-four measurements were performed using 23 separate, 3-mm-long, 40 to 70% stenoses. As predicted by the theoretic model, "true" gradient is compounded by the angioplasty catheter principally as a function of the angioplasty catheter diameter (Dc) and the stenosis diameter (Ds). The best-fit curve of data points relating "true" and compounded gradients to various Dc and Ds combinations can be expressed as: Measured gradient = K X true gradient, where K = 0.25 (e)4.47 (Dc divided by Ds) and e = 2.718. Thus, the transstenotic gradient measured at angioplasty overestimates "true" resting gradient in a predictable manner, which is dependent on the ratio of Dc to Ds.

Recurrent early ischemic events after thrombolysis for acute myocardial infarction.

Myocardium salvaged by early thrombolysis and then perfused through a residual stenosis may be at risk for ischemic events. To investigate this possibility, the short-term (2-week) clinical course of 81 consecutive patients managed within a randomized intracoronary thrombolysis trial was reviewed. All patients underwent coronary angiography within 5 hours of symptoms of acute myocardial infarction and were stratified into the following 3 outcome groups: patients with initially subtotal occlusion (subtotal group, n = 17), those with initial total occlusion and infarct artery reperfusion (reperfused group, n = 24) and those with continued infarct artery occlusion (occluded group, n = 40). Recurrent ischemic events were defined as spontaneous typical angina, provokable angina on predischarge exercise testing, and reinfarction. Eleven of 17 patients (65%) in the subtotal and 11 of 23 patients (48%) in the reperfused groups had an ischemic event (difference not significant). In contrast, 4 of 37 patients (11%) with occlusion had an ischemic event (p less than 0.01 compared with patients in the subtotal or reperfused groups). Four patients were excluded because of early (within 72 hours) elective coronary bypass surgery or death from pump failure. To eliminate the impact of multivessel coronary artery disease (CAD), 39 patients with 1-vessel CAD were analyzed separately. Five of 9 patients (56%) in the subtotal group, 3 of 10 (30%) in the reperfused group and only 2 of 20 (10%) in the occluded group had an ischemic event. These observations suggest the need for a more definitive revascularization strategy for acute myocardial infarction.

A randomized, angiographically controlled trial of intracoronary streptokinase in acute myocardial infarction.

Fifty-five patients with acute myocardial infarction evaluated within 4 hours of the onset of symptoms were entered into an angiographically controlled trial of intracoronary streptokinase (IC STK). Forty-three patients with total occlusion of their infarct artery were randomized to either IC STK or intracoronary nitroglycerin (IC NTG), and 12 patients with less-than-complete occlusion received only IC NTG. Reperfusion of a totally occluded vessel was achieved in 69% of STK patients and 17% of IC NTG patients. Time from onset of symptoms to peak CK activity was significantly shorter in reperfused patients and patients with subtotal occlusion on initial angiography than in patients with total occlusion who were not reperfused (p less than 0.0001). Comparison of radionuclide ejection fractions (EF) determined acutely and 10 to 14 days after infarction failed to show improvement in either the STK or IC NTG group (mean decrease of 2.8% and 0.4%, respectively). In contrast, patients with subtotal occlusion on baseline angiography demonstrated a significant (p = 0.05) spontaneous improvement in EF over 2 weeks (7.3% increase).

Detection of left ventricular wall motion abnormalities for the diagnosis of coronary artery disease: a comparison of exercise radionuclide and pacing intravenous digital ventriculography.

Intravenous digital ventriculography before and after pacing was compared with equilibrium gated nuclear ventriculography at rest and after exercise. Specifically, the relative abilities of the 2 techniques to detect resting and stress-related wall motion abnormalities were tested. Twelve normal patients and 28 patients with coronary artery disease (CAD) were tested. Neither technique produced a new wall motion abnormality in a patient with normal coronary arteries. Six patients with CAD had a history of a myocardial infarction (MI); an abnormality at rest was present in all 6 by both techniques. Of the 22 patients with CAD and a normal baseline ventriculogram, a wall motion abnormality developed in 18 during digital ventriculography with pacing; a wall motion abnormality developed in 15 with exercise nuclear ventriculography. Wall motion abnormalities by nuclear ventriculography (performed in the left anterior oblique projection) tended to be apical; digital ventriculography (performed in the right anterior oblique projection) more often produced an abnormality of the anterior or inferior wall, which could be predictive of coronary anatomy. Thus, the 2 techniques are substantially equivalent for the detection of wall motion abnormalities in CAD.

DNA damage measured using the alkaline elution assay depends on time between subculturing of cells and irradiation.

In experiments utilizing the alkaline filter elution assay for radiation-induced DNA damage we observed an unexpected dependence of hypoxic dose-response curves on the length of time V79 cells were in exponential growth between subculturing and irradiation. Dose-response curves for DNA from cells irradiated in air were identical regardless of whether the exponential-phase cells had been subcultured 24 or 48 h prior to irradiation, but cells irradiated in hypoxia 24 h after subculture displayed a dose-response curve for DNA damage which was two times steeper than that obtained for cells irradiated in hypoxia 48 h after subculture. Possible mechanisms for this effect are discussed.

Interactions of radioprotectors and oxygen in cultured mammalian cells. II. Effects of dithiothreitol on radiation-induced DNA damage and comparison with cell survival.

The effects of the sulfhydryl-containing compound dithiothreitol (DTT) on radiation-induced DNA damage have been studied using two different assays: DNA unwinding hydroxyapatite chromatography and alkaline filter elution. DNA damage as measured by both assays for cells irradiated in air shows drug concentration-dependent radioprotection reaching high levels (dose reduction factor, DRF = 3) at high DTT concentrations. The pattern and degree of protection against DNA damage are the same as shown previously for cell survival. However, when cells are irradiated in hypoxia, DNA damage as measured by the unwinding technique is decreased less by low DTT concentrations than is survival, but DNA damage is decreased to a much greater extent (DRF = 3) at high concentrations of DTT (compared to DRF = 1.5 for cell survival). DNA damage as measured by the alkaline elution assay after hypoxic irradiation is decreased to a much greater extent at all concentrations of DTT with DRF = 1.6 at 1 mM and increasing to DRF = 4.5 at high levels of DTT. These results are discussed in terms of the different types of DNA damage produced in cells irradiated in air versus hypoxia and the differences in types of damage measured by the two different DNA assays and cell survival.