A blood-based epigenetic test for early detection of nasopharyngeal carcinoma (NPC)

6063

Background: NPC is highly curable in early stages but 70% of NPC patients are diagnosed with advanced disease due to lack of effective screening. Genetic and epigenetic alterations involved in the pathogenesis of NPC are known. The higher order chromosomal structures reflecting aberrant transcriptional states of these genes can be measured via techniques such as chromosome conformation capture. Detection of these changes in peripheral blood may provide an accurate test for the early cancer detection. Methods: Blood samples have been collected from 84 patients with histologically confirmed NPC and 100 matched controls. Samples from 45 NPC patients and 68 controls have been analyzed. Fourteen genes known to be dysregulated in NPC were investigated. Potential higher order juxtaposition sites in the candidate genes were predicted using pattern recognition software. PCR primer sets were designed around the chosen sites to screen potential markers. Twenty-two markers showing predictability between NPC and control samples were analysed for optimal reproducibility using alternative primer sets. The optimal sets of markers were then tested amongst the complete set of samples. The dataset was processed by re-sampling using the synthetic minority oversampling technique. The overall sample was split into two groups (66% training set and 34% test set) in the classification. Results: Sixteen markers from 7 candidate genes were found to be optimal in differentiating between NPC and control samples in the first 103 samples. Using the multilayer perceptron (MLP) classification, the following results were obtained: Sensitivity 88.9%, 95% CI (79.2% - 98.6%); Specificity 72.7%, 95% CI (58.9% - 86.5%); PPV 72.7%, 95% CI (58.9% - 86.5%); NPV 88.9%, 95% CI (79.2% - 98.6%). The accuracy of the test was similar in detection of stage I and II NPC versus that of stage III or IV NPC. Conclusions: Using a PCR-based method to detect alterations in the cancer epigenome, the feasibility of developing a blood test of potential utility in early diagnosis of NPC was demonstrated. Analysis of larger numbers of patient samples and optimization of markers are ongoing. The performance characteristics of the test in the total population of 184 samples will be presented.

Pleuropulmonary paragonimiasis: radiologic findings in 71 patients

Pleuropulmonary paragonimiasis is a disease caused by lung flukes characterized by migration of a juvenile worm in the early stage and by formation of cysts around the worm later on. The purpose of this study was to describe the radiologic manifestations of pleuropulmonary paragonimiasis, with special emphasis on the worm cyst and worm migration track. We retrospectively studied 71 patients who had evidence of pleuropulmonary paragonimiasis on chest radiographs (n = 71) and CT scans (n = 17). The diagnosis was based on the detection of eggs or on positive antibody tests. On chest radiographs, 59 patients (83%) had pulmonary lesions and 43 patients (61%) had pleural lesions. Pulmonary findings included patchy air-space consolidation (n = 37) with or without cystic changes; ring shadows (n = 16); and peripheral linear opacities (n = 29), which were more prominent in patients with pleural effusion. Twelve patients (17%) had bilateral pleural effusions or pneumothoraces. On CT scans, round low-attenuation cystic lesions (5-15 mm), filled either with fluid (n = 5) or gas (n = 5), were characteristically seen within the consolidation. Peripheral linear opacities seen on radiographs were suggestive of worm migration tracks on CT scans. CT may reveal an intracystic worm. Although the findings vary depending on the stage of the disease, findings on chest radiographs are usually typical of paragonimiasis. CT provides more specific information about the worm cyst and the worm migration track.

Enterobacterial common antigen in rfb deletion mutants of Salmonella typhimurium

The his-rfb deletion series of Salmonella typhimurium mutants characterized previously by Nikaido et al. was examined for the presence of the enterobacterial common antigen (ECA). All deletions not extending further to the left than the genes for cytidine phosphoabequose synthesis were ECA positive, whereas longer deletions (extending to the genes for thymidine diphosphorhamnose synthesis or further) were ECA negative. When these long-his-rfb deletion strains were studied further, it became clear that they (four out of four studied) had accumulated a second mutation, called rff, close to ilv, which prevented the synthesis of ECA. When rff- was replaced by rff+, the recombinants, now having the his-rfb deletion only, produced traces of ECA, showed reduced viability, increased sensitivity to sodium dodecyl sulfate (SDS) and to a lesser extent, to other anionic detergents, and accumulated secondary "suppressor" mutations upon storage. Such suppressor-containing mutants could be isolated by selecting for resistance to 1% SDS. Thirty of 46 SDS-resistant mutants studied had a second mutation, which alone prevented the synthesis of ECA, close to ilv. This ilv-linked mutation was similar to the rff mutation of the strains studied originally. The new rff mutation was similar to previously described rfe mutations in its close linkage to ilv and association with an ECA-negative phenotype. It differed from rfe, however, by not affecting the synthesis of the O antigens (O-6,7) of group C1. In Salmonella group C1, all ECA genes identified thus far are linked to ilv (rfe and/or rff) and none is linked to rfb.

Role of a lipopolysaccharide gene for immunogenicity of the enterobacterial common antigen

It is known that only certain strains of the family of Enterobacteriaceae, notably rough (R) mutants with the type R1 or R4 core, evoked antibodies in high titers against the common enterobacterial antigen (CA) after immunization of rabbits with heated cell suspensions. The present investigation deals with genetic and immunochemical aspects of certain R1 and R4 mutants isolated from Escherichia coli 08 and various Shigella serotypes which, unexpectedly, do not induce CA antibody formation. Immunochemical and genetical (transduction and conjugation) experiments revealed that the rough phenotype of these special mutants was evoked by a mutation of pyrE-linked rfa gene, called rfaL, which is involved in translocation of O-specific polysaccharides onto the lipopolysaccharide core. The transduction of the defective rfaL, allele into appropriate rough recipients results in transductants which have simultaneously lost the ability to evoke CA antibodies. This finding suggests that a close connection exists between the function of the rfaL gene and the expression of CA immunogenicity in R1 and R4 mutants. One of the strains synthesized neither O-hapten nor CA, suggesting a mutation in a region equivalent to the rfe genes of Salmonella.

Strain-related differences in immunosuppressive effects of Enterobacteriaceae and their lipopolysaccharides on production in rabbits of antibody to enterobacterial common antigen

Certain polysaccharides have been shown to inhibit the antibody response of rabbits to the common enterobacterial antigen (CA). The present investigation revealed that striking differences exist in the immunosuppressive effects of enteric bacteria and their lipolysaccharides (lps), depending upon CA production by the strains. Mixtures of immunogenic strains (Escherichia coli F2378 [R4], E. coli F470 [R1], or Shigella boydii F3140 [R]) and non-immunogenic CA-producing strains, such as E. coli O1, E. coli O113, Salmonella montevideo, and S. minnesota, as well as the R mutants E. coli F614 (R1), E. coli F757 (R1), and S. typhimurium his 642 (Ra), failed to elicit CA antibodies. In contrast, mixtures of the immunogen and CA-negative strains S. typhimurium his 386 (Ra) and S. minnesota P595 (Re) or R555 (Ra) yielded antibodies in titers similar to those elicited by the immunogen alone. Further, LPS of CA-positive but not of CA-negative strains exerted this immunosuppressive effect. Quantitative studies revealed that LPS of S. minnesota in amounts of 100 mug/ml was strongly immunosuppressive, in amounts of 20 mug/ml slightly effective, and in amounts of 4 mug/ml ineffective. It is postulated that hitherto unknown differences exist, either in composition or in configuration, between LPS obtained from different microorganisms to account for the strain-related differences in immunosuppressive effects and, further, that the immunosuppressive LPS interacts with immunogenic CA.

Participation of lipopolysaccharide genes in the determination of the enterobacterial common antigen: analysis of R mutants of Salmonella minnesota

A series of R (rough) Salmonella minnesota mutants with rfb, rfe, and rfa mutations leading to various defects in the biosynthesis of cell wall lipopolysaccharide was analyzed as to their enterobacterial common antigen (CA) content. All mutants that had functional rfe genes were CA(+) as is the wild-type parent. This includes mutants with the most defective lipopolysaccharide core types, demonstrating that core structures are not a necessary part of CA. All rfe(-) mutants (complete lipopolysaccharide core, defective synthesis of O side chains) were defective in the synthesis of CA. A smooth strain was accidentally found to be CA(-); the mutation responsible for this defect was also located, like rfe, very close to ilv.

Gel precipitation of common enterobacterial antigen by its antibody

Reinvestigation of the precipitation of the common enterobacterial antigen (CA) by its antibody revealed the following. (1) The precipitation in gel of CA by CA antiserum is characterized by a distinct arc-shape precipitation line close to the antiserum well. (2) Only CA antisera of high titer, as determined by hemagglutination, caused precipitation of CA in gel. However, not all antisera of similar hemagglutinin titer produced gel precipitation. (3) Concentration of CA antisera through collodion membrane increased the precipitating capacity of some CA antisera. (4) Various preparations containing the CA determinant were suitable antigens in this reaction, including the ethanolsoluble fraction of various <i>Enterobacteriaceae</i> and certain lipopolysaccharides. (5) The specificity of the reaction was documented by observing lines of identity with various CA-positive preparations and failure of precipitation of CA-negative preparations, including the ethanol-soluble fraction of <i>S. minnesota</i> Re 595, supernates of agar-grown cultures of <i>Staphylococcus</i> and <i>Pseudomonas</i>, certain lipopolysaccharides, and Vi antigen. (6) The precipitin line of CA by CA antibodies is distinctly different from those formed by·and Re-core antigen-antibody reactions.

Biochemical basis of the immunogenicity of the common enterobacterial antigen

Of the numerous members of the family Enterobacteriaceae only a few strains, notably Escherichia coli O14 and R mutants of the E. coli R1-core type, engender antibodies against the common enterobacterial antigen (CA) following immunization of rabbits with heated suspensions or culture supernatants; other members produce nonimmunogenic CA of identical serological specificity. The biochemical basis of the immunogenic properties of CA of the former strains was investigated by determining the relationship between the CA determinant and the lipopolysaccharide molecule. Lipopolysaccharides extracted from R mutants of the E. coli R1-core type or of E. coli O14 by the phenol-chloroform-petrol ether method contain the CA determinant, in contrast to extracts of other CA-producing R mutants. This is evident from the observation that only the former absorb CA antibodies, utilizing erythrocytes coated with alkali-treated lipopolysaccharide preparations. Based on the findings that CA of R mutants of E. coli R1-core type follows lipopolysaccharide during all purification steps and that alkali treatment increases its affinity for erythrocytes parallel to that of the lipopolysaccharide, it is concluded that the CA determinant either is part of the lipopolysaccharide molecule or is strongly complexed with it. It is suggested that this association between CA and the lipopolysaccharide of E. coli R1-core type and E. coli O14 accounts for the heat stability of the immunogenicity of CA of these unusual strains.

Immunogenicity of the common enterobacterial antigen produced by smooth and rough strains

Enterobacteriaceae share a common antigen (CA). This antigen exists as a powerful immunogen, when produced by Escherichia coli O14, and as a minimally effective immunogen, when present in cultures of most other smooth strains. The present study was directed toward the assessment of immunogenicity of this antigen produced by various rough mutants and their parent strains as well as that of viable and nonviable bacterial suspensions. Rabbits were immunized intravenously, and the passive hemagglutination test was used for the quantitation of CA antibodies. The following results were obtained. (i) Immunization of rabbits with viable or heat-killed (1 hr, 100 C) suspensions of smooth parent strains resulted in only slight or insignificant antibody production. Of seven additional smooth strains of Enterobacteriaceae tested, none induced CA antibodies upon immunization with killed (1 hr, 100 C) suspensions, and only two engendered antibodies in moderate titers when viable suspensions were used, possibly due to the presence of undetected R mutant cells. (ii) In contrast, administration of viable suspensions of the corresponding mutant strains, regardless of core types, including E. coli R1, R2, R3, and Salmonella Ra types, caused a significant antibody response. (iii) The immunogenicity of R mutants differed strikingly, inasmuch as mutants with the R1 core [E. coli O8(-):K27(-)(R) and Shigella boydii type 3(-)(R)] engendered CA antibodies when heated suspensions were administered, in contrast to heated suspensions of mutants with the R2, R3, and Ra cores. (iv) Viability per se is not an abolute requirement for immunogenicity of R2 mutants, since CA antibodies were engendered by suspensions killed at 60 C or by merthiolate or formaldehyde. (v) Ethanol fractionation revealed that immunogenic CA of the R1 mutant is ethanol-insoluble and that of the parental smooth strain is ethanol-soluble; thus, CA of R1 mutants resembles that of E. coli O14 in both immunogenicity and ethanol insolubility.

Production by Aeromonas of common enterobacterial antigen and its possible taxonomic significance

In a study of production of the common enterobacterial antigen (CA) by members of the Aeromonas group, ten strains of A. shigelloides, nine strains of A. hydrophila, and nine strains of A. salmonicida were used. Passive hemagglutination and hemolysis tests as well as the hemagglutination-inhibition procedure revealed that all strains of A. shigelloides, in contrast to the strains of the other two species, produce this antigenic determinant. The antigen of A. shigelloides was demonstrated even when the supernatant fluids of agar-grown cultures were used in a dilution of 1:1,000, whereas 10-times concentrated supernatant fluids obtained from the other two species were negative. Supernatant liquids of cultures of A. shigelloides failed to induce a significant CA immune response in rabbits; nonetheless, the animals were primed immunologically and responded with prompt production of CA antibodies in significant titers to a booster injection of a subeffective dose of CA obtained from Salmonella typhimurium. Strains of A. hydrophila and A. salmonicida neither induced CA antibody formation nor primed the animals. It is concluded that of the three species of the Aeromonas group only A. shigelloides, which may produce O antigen cross-reacting with Shigella sonnei and which has been isolated from patients with dysentery or gastroenteritis, produces CA. Production of this antigen, therefore, may help to characterize microorganisms belonging or related to the family Enterobacteriaceae.

Chlorphenesin: an Antigen-Associated Immunosuppressant

Chlorphenesin (3- p -chlorophenoxy-1,2-propanediol), when injected intravenously together with either of two common bacterial antigens, inhibits the antibody response of the rabbit. The antigens studied are those common to Enterobacteriaceae and to gram-positive bacteria. The immunosuppression is contingent upon incubation of chlorphenesin and antigen in vitro prior to administration, since separate injection of antigen and inhibitor or of mixtures without prior incubation yields undiminished antibody response. Chlorphenesin, as shown by hemagglutination-inhibition tests, does not alter the antigenic determinants, because antibody neutralization occurs in the presence or absence of the drug. The immunosuppressive effect is reversible, since precipitation of chlorphenesin at 4 C substantially restores immunogenicity. Animals immunized with antigen-drug mixtures, which fail to respond with significant antibody production, nonetheless are immunologically primed. It is concluded that chlorphenesin represents another example of antigen-associated immunosuppressants.

Antigen-associated immunosuppressant: effect of serum on immune response

Serum from various animal species, including the test animals themselves, inhibits the antibody response of the rabbit to two bacterial antigens, provided that antigen and immunosuppressant interact prior to injection. The degree of immunosuppression is related to the length of incubation in vitro of antigen and serum. Serum does not hinder or destroy the antigenic determinant. Bacterial antibodies do not account for inhibition of the antibody response. Antigen-associated serum components, as yet unidentified, may affect the early events of the immune response.