Independent and coordinate effects of ADP-ribosyltransferase and GTPase-activating activities of exoenzyme S on HT-29 epithelial cell function.

Type III-mediated translocation of exoenzyme S (ExoS) into HT-29 epithelial cells by Pseudomonas aeruginosa causes complex alterations in cell function, including inhibition of DNA synthesis, altered cytoskeletal structure, loss of readherence, microvillus effacement, and interruption of signal transduction. ExoS is a bifunctional protein having both GTPase-activating (GAP) and ADP-ribosyltransferase (ADPRT) functional domains. Comparisons of alterations in HT-29 cell function caused by P. aeruginosa strains that translocate ExoS having GAP or ADPRT mutations allowed the independent and coordinate functions of the two activities to be assessed. An E381A ADPRT mutation revealed that ExoS ADPRT activity was required for effects of ExoS on DNA synthesis and long-term cell rounding. Conversely, the R146A GAP mutation appeared to have little impact on the cellular effects of ExoS. While transient cell rounding was detected following exposure to the E381A mutant, this rounding was eliminated by an E379A-E381A ADPRT double mutation, implying that residual ADPRT activity, rather than GAP activity, was effecting transient cell rounding by the E381A mutant. To explore this possibility, E381A and R146A-E381A mutants were examined for their ability to ADP-ribosylate Ras in vitro or in vivo. While no ADP-ribosylation of Ras was detected by either mutant in vitro, both mutants were able to modify Ras when translocated by the bacteria, with the R146A-E381A mutant causing more efficient modification than the E381A mutant, in association with increased inhibition of DNA synthesis. Comparisons of Ras ADP-ribosylation by wild-type and E381A mutant ExoS by two-dimensional electrophoresis found the former to ADP-ribosylate Ras at two sites, while the latter modified Ras only once. These studies draw attention to the key role of ExoS ADPRT activity in causing the effects of bacterially translocated ExoS on DNA synthesis and cell rounding. In addition, the studies provide insight into the enhancement of ExoS ADPRT activity within the eukaryotic cell microenvironment and into possible modulatory roles that the GAP and ADPRT domains might have on the function of each other.

Sex of first child and breast cancer survival in young women.

Two studies have reported that young women with breast cancer face increased risk of early mortality if their first child was male rather than female. An immunological mechanism has been suggested. We sought to confirm these results in a larger, historical cohort study of 223 parous women who were aged <45 years at breast cancer diagnosis during 1983-1987. Subjects were identified through the Maine Cancer Registry. Follow-up data were obtained from hospitals, physicians, and death certificates. Reproductive history data were obtained from the next of kin of the deceased women, birth certificates, physicians, hospitals, and lastly, subjects. With a 7-year follow-up, multivariate modeling found a lower mortality risk in women with a male first child (hazard ratio [HR] 0.51, 95% confidence interval [CI] 0.32-0.81, log-rank comparison). The survival advantage remained for at least 13 years in women with a male firstborn. Thus, previous studies were not confirmed. Mortality risk in young women with breast cancer is not increased by having borne a male first child rather than a female first child.

Comparison of the exoS gene and protein expression in soil and clinical isolates of Pseudomonas aeruginosa.

Exoenzyme S (ExoS) is translocated into eukaryotic cells by the type III secretory process and has been hypothesized to function in conjunction with other virulence factors in the pathogenesis of Pseudomonas aeruginosa. To gain further understanding of how ExoS might contribute to P. aeruginosa survival and virulence, ExoS expression and the structural gene sequence were determined in P. aeruginosa soil isolates and compared with ExoS of clinical isolates. Significantly higher levels of ExoS ADP-ribosyltransferase (ADPRT) activity were detected in culture supernatants of soil isolates compared to those of clinical isolates. The higher levels of ADPRT activity of soil isolates reflected both the increased production of ExoS and the production of ExoS having a higher specific activity. ExoS structural gene sequence comparisons found the gene to be highly conserved among soil and clinical isolates, with the greatest number of nonsynonymous substitutions occurring within the region of ExoS encoding GAP function. The lack of amino acid changes in the ADPRT region in association with a higher specific activity implies that other factors produced by P. aeruginosa or residues outside the ADPRT region are affecting ExoS ADPRT activity. The data are consistent with ExoS being integral to P. aeruginosa survival in the soil and suggest that, in the transition of P. aeruginosa from the soil to certain clinical settings, the loss of ExoS expression is favored.

Coronary calcium in adults with type 1 diabetes: a stronger correlate of clinical coronary artery disease in men than in women.

We studied the relationship of coronary artery calcification (CAC), a marker of coronary atherosclerosis, with prevalent clinical coronary artery disease (CAD) and established cardiovascular disease (CVD) risk factors in a type 1 diabetic population. At the 10-year follow-up examination of the Pittsburgh Epidemiology of Diabetes Complications (EDC) Study cohort, 302 adults (mean age 38.1 +/- 7.8 years) received electron beam tomography (EBT) scanning of the heart and a clinical examination. Clinical CAD was defined as a confirmed history of myocardial infarction (MI), angiographic stenosis > or =50%, Pittsburgh EDC Study physician-diagnosed angina, or ischemic electrocardiogram (ECG). CAC correlated with most CVD risk factors. CAC had 84 and 71% sensitivity for clinical CAD in men and women, respectively, and 100% sensitivity for MI or obstructive CAD. A CACS cut point of 400 was the most efficient coronary calcium correlate of CAD. In subjects with angina only, CAC sensitivity was 83% in men and 46% in women. In logistic regression, CAC, ECG R-R variation, peripheral vascular disease, and Beck Depression Inventory independently correlated with prevalent CAD in men and overall. Except for CAC, the same variables independently correlated with CAD in women, and age also entered the model. CAC was an independent correlate of MI or obstructive CAD in both sexes and was the strongest independent correlate in men, but CAC was not independently associated with angina and ischemic ECG in either sex. It is concluded that EBT-detected CAC is strongly correlated with CAD in type 1 diabetes-particularly in men.

Functional comparison of the NAD binding cleft of ADP-ribosylating toxins.

Although a common core structure forms the active site of ADP-ribosylating (ADPRT) toxins, the limited-sequence homology within this region suggests that different mechanisms are being used by toxins to perform their shared function. To explain differences in their mechanisms of NAD binding and hydrolysis, the functional interrelationship of residues predicted to perform similar functions in the beta3-strand of the NAD binding cleft of different ADPRT toxins was compared. Replacing Tyr54 in the A-subunit of diphtheria toxin (DTA) with a serine, its functional homologue in cholera toxin (CT), resulted in the loss of catalytic function but not NAD binding. The catalytic role of the aromatic portion of Tyr54 in the ADPRT reaction was confirmed by the ability of a Tyr54-to-phenylalanine DTA mutant to retain ADPRT activity. In reciprocal studies, positioning a tyrosine in the beta3-strand of the A1-subunit of CT (CTA1) caused both loss of function and altered structure. The restricted flexibility of the CTA1 active site relative to function became evident upon the loss of ADPRT activity when a conservative Val60-to-leucine mutation was performed. We conclude from our studies that DT and CT maintain a similar mechanism of NAD binding but differ in their mechanisms of NAD hydrolysis. The aromatic moiety at position 54 in DT is integral to NAD hydrolysis, while NAD hydrolysis in CT appears highly dependent on the precise positioning of specific residues within the beta3-strand of the active-site cleft.

Selective killing of cancer cells based on loss of heterozygosity and normal variation in the human genome: a new paradigm for anticancer drug therapy.

Most drugs for cancer therapy are targeted to relative differences in the biological characteristics of cancer cells and normal cells. The therapeutic index of such drugs is theoretically limited by the magnitude of such differences, and most anticancer drugs have considerable toxicity to normal cells. Here we describe a new approach for developing anticancer drugs. This approach, termed variagenic targeting, exploits the absolute difference in the genotype of normal cells and cancer cells arising from normal gene sequence variation in essential genes and loss of heterozygosity (LOH) occurring during oncogenesis. The technology involves identifying genes that are: 1) essential for cell survival; 2) are expressed as multiple alleles in the normal population because of the presence of one or more nucleotide polymorphisms; and 3) are frequently subject to LOH in several common cancers. An allele-specific drug inhibiting the essential gene remaining in cancer cells would be lethal to the malignant cell and would have minimal toxicity to the normal heterozygous cell that retains the drug-insensitive allele. With antisense oligonucleotides designed to target two alternative alleles of replication protein A, 70-kDa subunit (RPA70) we demonstrate in vitro selective killing of cancer cells that contain only the sensitive allele of the target gene without killing cells expressing the alternative RPA70 allele. Additionally, we identify several other candidate genes for variagenic targeting. This technology represents a new approach for the discovery of agents with high therapeutics indices for treating cancer and other proliferative disorders.

Pseudomonas aeruginosa exoenzyme S disrupts Ras-mediated signal transduction by inhibiting guanine nucleotide exchange factor-catalyzed nucleotide exchange.

Pseudomonas aeruginosa exoenzyme S double ADP-ribosylates Ras at Arg(41) and Arg(128). Since Arg(41) is adjacent to the switch 1 region of Ras, ADP-ribosylation could interfere with Ras-mediated signal transduction via several mechanisms, including interaction with Raf, or guanine nucleotide exchange factor-stimulated or intrinsic nucleotide exchange. Initial experiments showed that ADP-ribosylated Ras (ADP-r-Ras) and unmodified Ras (Ras) interacted with Raf with equal efficiencies, indicating that ADP-ribosylation did not interfere with Ras-Raf interactions. While ADP-r-Ras and Ras possessed equivalent intrinsic nucleotide exchange rates, guanine nucleotide exchange factor (Cdc25) stimulated the nucleotide exchange of ADP-r-Ras at a 3-fold slower rate than Ras. ADP-r-Ras did not affect the nucleotide exchange of Ras, indicating that the ADP-ribosylation of Ras was not a dominant negative phenotype. Ras-R41K and ADP-r-Ras R41K possessed similar exchange rates as Ras, indicating that ADP-ribosylation at Arg(128) did not inhibit Cdc25-stimulated nucleotide exchange. Consistent with the slower nucleotide exchange rate of ADP-r-Ras as compared with Ras, ADP-r-Ras bound its guanine nucleotide exchange factor (Cdc25) less efficiently than Ras in direct binding experiments. Together, these data indicate that ADP-ribosylation of Ras at Arg(41) disrupts Ras-Cdc25 interactions, which inhibits the rate-limiting step in Ras signal transduction, the activation of Ras by its guanine nucleotide exchange factor.

Differential sensitivity of human epithelial cells to Pseudomonas aeruginosa exoenzyme S.

Exoenzyme S (ExoS) is an ADP-ribosyltransferase produced and directly translocated into eukaryotic cells by the opportunistic pathogen Pseudomonas aeruginosa. Model systems that allow bacterial translocation of ExoS have found ExoS to have multiple effects on eukaryotic cell function, affecting DNA synthesis, actin cytoskeletal structure, and cell matrix adherence. To understand mechanisms underlying differences observed in cell sensitivities to ExoS, we examined the effects of bacterially translocated ExoS on multiple human epithelial cell lines. Of the cell lines examined, confluent normal kidney (NK) epithelial cells were most resistant to ExoS, while tumor-derived cell lines were highly sensitive to ExoS. Analysis of the mechanisms of resistance indicated that cell association as well as an intrinsic resistance to morphological alterations were associated with increased resistance to ExoS. Conversely, increased sensitivity to ExoS appeared to be linked to epithelial cell growth, with tumor cells capable of undergoing non-contact-inhibited, anchorage-independent growth all being sensitive to ExoS, and NK cells becoming sensitive to ExoS when subconfluent and growing. Consistent with the possibility that growth-related, actin-based structures are involved in sensitivity to ExoS, scanning electron microscopy revealed cellular extensions from sensitive, growing cells to bacteria, which were not readily evident in resistant cells. In all studies, the severity of effects of ExoS on cell function directly correlated with the degree of Ras modification, indicating that sensitivity to ExoS in some manner related to the efficiency of ExoS translocation and its ADP-ribosylation of Ras. Our results suggest that factors expressed by growing epithelial cells are required for the bacterial contact-dependent translocation of ExoS; as normal epithelial cells differentiate into polarized confluent monolayers, expression of these factors is altered, and cells in turn become more resistant to the effects of ExoS.

ADP-ribosylation of oncogenic Ras proteins by pseudomonas aeruginosa exoenzyme S in vivo.

The exoenzyme S (ExoS)-producing Pseudomonas aeruginosa strain, 388, and corresponding ExoS knock-out strain, 388deltaexoS, were used in a bacterial and mammalian co-culture system as a model for the contact-dependent delivery of ExoS into host cells. Examination of DNA synthesis and Ras ADP-ribosylation in tumour cell lines expressing normal and mutant Ras revealed a decrease in DNA synthesis concomitant with ADP-ribosylation of Ras proteins after exposure to ExoS-producing bacteria, but not after exposure to non-ExoS-producing bacteria. Examination of normal H-Ras, K-Ras and N-Ras by two-dimensional electrophoresis after exposure to bacteria revealed differences in the degree of ADP-ribosylation by ExoS, with H-Ras being modified most extensively. ADP-ribosylation of oncogenic forms of Ras was examined in vivo using cancer lines expressing mutant forms of H-, N- or K-Ras. The mutant Ras proteins were modified in a manner qualitatively similar to their normal counterparts. Using Ras/Raf-1 co-immunoprecipitation after co-culture, it was found that exposure to ExoS-producing bacteria caused a decrease in the amount of Raf-1 associated with EGF-activated Ras and oncogenic Ras. The results from this study indicate that ExoS ADP-ribosylates both normal and mutant Ras proteins in vivo and inhibits signalling through Ras.

Interruption of multiple cellular processes in HT-29 epithelial cells by Pseudomonas aeruginosa exoenzyme S.

Exoenzyme S (ExoS), an ADP-ribosylating enzyme produced by the opportunistic pathogen Pseudomonas aeruginosa, is directly translocated into eukaryotic cells by bacterial contact. Within the cell, ExoS ADP-ribosylates the cell signaling protein Ras and causes inhibition of DNA synthesis and alterations in cytoskeletal structure. To further understand the interrelationship of the different cellular effects of ExoS, functional analyses were performed on HT-29 epithelial cells after exposure to ExoS-producing P. aeruginosa 388 and the non-ExoS-producing strain 388DeltaS. Two different mechanisms of morphological alteration were identified: (i) a more-transient and less-severe cell rounding caused by the non-ExoS-producing strain 388DeltaS and (ii) a more-severe, long-term cell rounding caused by ExoS-producing strain 388. Long-term effects of ExoS on cell morphology occurred in conjunction with ExoS-mediated inhibition of DNA synthesis and the ADP-ribosylation of Ras. ExoS was also found to cause alterations in HT-29 cell function, leading to the loss of cell adhesion and microvillus effacement. Nonadherent ExoS-treated cells remained viable but had a high proportion of modified Ras. While microvillus effacement was detected in both 388- and 388DeltaS-treated cells, effacement was more prevalent and rapid in cells exposed to strain 388. We conclude from these studies that ExoS can have multiple effects on epithelial cell function, with more severe cellular alterations associated with the enzymatic modification of Ras. The finding that ExoS had greater effects on cell growth and adherence than on cell viability suggests that ExoS may contribute to the P. aeruginosa infectious process by rendering cells nonfunctional.

p53 and apoptosis alterations in keloids and keloid fibroblasts.

Keloids are the result of a dysregulated wound-healing process and are characterized by formation of excess scar tissue that proliferates beyond the boundaries of the inciting wound. In this study, we investigated the expression of key proteins involved in regulating apoptosis in keloids. Twenty archival paraffin-embedded keloid samples were randomly selected for an immunoperoxidase assay with antibodies against fas, p53, bcl-2, and bcl-x proteins using the target antigen-retrieval technique. Apoptosis was assessed in keloids and normal skin and in keloid and normal fibroblasts by the TdT-mediated dUTP nick-end labeling (tunel) assay on tissue sections, fibroblast cultures, and by flow cytometry for cell suspensions. We found that 18 of 20 keloids expressed p53 protein; bcl-2 was expressed by keloid fibroblasts in 19 of 20 keloids, and all specimens had prominent fas expression throughout the tissue. The distribution of these three antigens was regional within each lesion and followed a consistent pattern of p53 and bcl-2 expression colocalized to the hypercellular, peripheral areas of each keloid in a perinuclear pattern (p < .001). In contrast, an inverse distribution of fas expression was shown, with staining being more diffuse across the cell surfaces and limited to the central, more hypocellular regions in16 of 17 keloids (p < .001). There was no specific staining pattern in these keloids with antihuman bcl-x. In vitro studies on cultured keloid fibroblasts (derived from six patients) revealed maintenance of the p53+, bcl-2+ phenotype up to passage 10. Neither neonatal nor normal adult skin fibroblasts expressed either antigen but could be induced to express p53 by exposure to adriamycin. Keloid lesions and keloid fibroblasts were found to have lower rates of apoptosis than normal controls. Keloid fibroblasts displayed enhanced apoptosis rates in response to hydrocortisone, gamma interferon, and hypoxia treatment as compared with normal adult fibroblasts. Focal dysregulation of p53 combined with upregulation of bcl-2 may help produce a combination of increased cell proliferation and decreased cell death in the younger, hypercellular areas of the keloid. This phenotype is reversed in the older areas of the keloid and may prevent malignant degeneration, thus favoring normal apoptosis as evidenced by prominent fas expression.

Modification of Ras in eukaryotic cells by Pseudomonas aeruginosa exoenzyme S.

Genetic and functional data suggest that Pseudomonas aeruginosa exoenzyme S (ExoS), an ADP-ribosyltransferase, is translocated into eukaryotic cells by a bacterial type III secretory mechanism activated by contact between bacteria and host cells. Although purified ExoS is not toxic to eukaryotic cells, ExoS-producing bacteria cause reduced proliferation and viability, possibly mediated by bacterially translocated ExoS. To investigate the activity of translocated ExoS, we examined in vivo modification of Ras, a preferred in vitro substrate. The ExoS-producing strain P. aeruginosa 388 and an isogenic mutant strain, 388DeltaexoS, which fails to produce ExoS, were cocultured with HT29 colon carcinoma cells. Ras was found to be ADP-ribosylated during coculture with 388 but not with 388DeltaexoS, and Ras modification by 388 corresponded with reduction in HT29 cell DNA synthesis. Active translocation by bacteria was found to be required, since exogenous ExoS, alone or in the presence of 388DeltaexoS, was unable to modify intracellular Ras. Other ExoS-producing strains caused modification of Ras, indicating that this is not a strain-specific event. ADP-ribosylation of Rap1, an additional Ras family substrate for ExoS in vitro, was not detectable in vivo under conditions sufficient for Ras modification, suggesting possible ExoS substrate preference among Ras-related proteins. These results confirm that intracellular Ras is modified by bacterially translocated ExoS and that the inhibition of target cell proliferation correlates with the efficiency of Ras modification.

Effects of differential expression of the 49-kilodalton exoenzyme S by Pseudomonas aeruginosa on cultured eukaryotic cells.

Production of the ADP-ribosylating enzyme exoenzyme S (ExoS) by Pseudomonas aeruginosa has been associated with increased virulence. Previous studies, however, have been unable to confirm an effect of soluble ExoS in cell culture or animal model systems. To determine if bacteria must come in contact with target cells in order for an effect of ExoS to be observed, coculture systems were developed to compare the effects of ExoS- and non-ExoS-producing bacteria on eukaryotic cell function. The two P. aeruginosa strains used in these studies, 388 and 388delta exoS, maintained genetic identity, with the exception that strain 388delta exoS lacked production of the 49-kDa form of ExoS. When bacteria were cocultured with Detroit 532 fibroblastic cells, ExoS-producing 388 bacteria caused a significant decrease in DNA synthesis and viability compared to the decrease caused by non-ExoS-producing 388delta exoS bacteria. Maximal differences between the two strains were observed when 10(4) to 10(7) CFU of bacteria/ml were cocultured with Detroit cells for 4 or 6 h. Both strains were effective in eliminating Detroit cell DNA synthesis after a 20-h coculture period. Secreted ExoS had no effect on Detroit cell growth and viability, indicating that bacteria must have contact with target cells for the effect of ExoS on cellular function to be observed. Similar effects on cell proliferation and viability were observed when the two strains were cocultured with the KB epithelioid cell line. ExoS-associated decreases in eukaryotic cell viability were not found to be mediated by an inhibition of protein synthesis. These studies confirm that the 49-kDa ExoS contributes to the cellular pathogenesis of P. aeruginosa by interfering with eukaryotic cell growth and viability. In addition, the coculture system developed which recognizes this effect should provide a means for defining the function of ExoS in vivo.

The elastase propeptide functions as an intramolecular chaperone required for elastase activity and secretion in Pseudomonas aeruginosa.

Several proteases are secreted by Pseudomonas aeruginosa including elastase, an abundantly secreted neutral zinc-metalloprotease. Elastase (encoded by lasB) is first synthesized with a relatively large propeptide (18 kDa) domain. Here, we present evidence that this propeptide functions as an intramolecular chaperone (IMC) essential for proper maturation of elastase into a hydrolytically active enzyme. An altered elastase allele (lasB6) that encoded an elastase precursor with a precise propeptide deletion was expressed in Escherichia coli, and disrupted cells contained only inactive elastase. However, co-expression of an allele (lasB7) expressing the propeptide as an independent, non-covalently linked protein rescued about one-third of the hydrolytic activity when compared with that obtained with wild-type lasB. Thus, the propeptide was essential for elastase activity and so defined elastase as an IMC-containing protease. We examined the possibility that the propeptide of elastase also plays a role in the localization of the mature protein past the outer bacterial membrane. Expression of lasB6 in P. aeruginosa (lasB delta) in the absence of the propeptide resulted in production of inactive elastase that accumulated within the cell and was not secreted to the culture medium. When lasB7 co-expressed the non-covalently linked propeptide in the same cell with lasB6, efficient secretion was restored and active elastase was then found in the supernatant. Thus, the propeptide was needed for secretion of the mature protein as well as enzymatic activity. This chaperone-like activity of the propeptide appears to involve a direct interaction between the mature and propeptide sequences, and evidence for this was obtained by demonstrating that the non-covalently attached 18 kDa propeptide was co-precipitated with elastase using elastase antibodies. These results are consistent with a hypothesis that the propeptide domain acts as an IMC to control both enzymatic activity and competence for secretion.

Flow cytometric detection of human immunodeficiency virus type 1 proviral DNA by the polymerase chain reaction incorporating digoxigenin- or fluorescein-labeled dUTP.

Serological assays are routinely used in the laboratory diagnosis of human immunodeficiency virus type-1 (HIV-1) infection, but the polymerase chain reaction (PCR) is ultimately the most sensitive and direct method for establishing definitive diagnosis. As an alternative to the conventional radioactive PCR procedure we have developed and evaluated a pair of rapid nonradioisotopic flow cytometric detection methods. Using heminested PCR we directly incorporated fluorescein-12-dUTP (fluo-dUTP) or digoxigenin-11-dUTP (dig-dUTP) into the PCR-amplicons. The labeled amplicons were hybridized with biotinylated antisense and sense probes, followed by capture of the hybrid DNA using streptavidin-coated beads which were finally analyzed in a flow cytometer by 1) direct detection of the fluorescence intensity of the amplicons incorporating fluo-dUTP and 2) immunodetection of the amplicons incorporating dig-dUTP by anti-digoxigenin IgG labeled with fluorescein isothiocyanate (FITC). Although both assays were functionally comparable with radiolabeled probe in reliably detecting as low as five copies of HIV-1 proviral DNA sequences, the immunodetection of dig-dUTP consistently yielded higher mean channel fluorescence and gave a stable signal over an extended period of 12-14 weeks. In testing a panel of 20 pedigreed PBMC specimens from blood donors with or without HIV-1 infection, the results of both flow cytometric assays were identical with those of the conventional radioactive procedure. Therefore, we conclude that the dig-dUTP incorporation in amplicons, hybridization with a pair of sense-antisense biotinylated probes and immunodetection of hybrids by flow cytometric analyses is the nonisotopic method of choice for PCR-diagnosis of HIV-1 infection.

Retinal hemorrhage as a consequence of epidural steroid injection.

OBJECTIVE: To report that retinal hemorrhage is an important but uncommon complication of epidural steroid injection, a procedure frequently employed to alleviate chronic back pain. DESIGN: Patients who complained of visual field defects or blurred vision after receiving the injection received complete eye examinations, including fundus photography and fluorescein angiography. They were then followed up as their clinical situation dictated. SETTING: Patients were examined in an outpatient setting in our offices. PATIENTS: Between 1989 and 1992, five eyes of four female patients ranging in age from 35 to 81 years were evaluated and followed up. RESULTS: Examination and photographic documentation showed that these eyes had retinal hemorrhages, often distributed in a petaloid pattern, which resulted in transient loss of vision from 6 weeks to 6 months. CONCLUSIONS: Retinal hemorrhage is an uncommon but significant and previously unemphasized complication of epidural steroid injection for chronic back pain. This procedure should be carefully considered, particularly in one-eyed patients and those with bleeding problems, and patients should be fully informed that this can happen despite appropriate care and technique.