Sinusitis in HIV-infected patients: a clinical and radiographic review.

PURPOSE: To describe the clinical, radiographic, and laboratory features of sinus disease in human immunodeficiency virus (HIV)-infected individuals. PATIENTS: Seventy-two patients with a history of sinusitis identified from 1,461 consecutive admissions (667 patients) to the HIV ward at The Johns Hopkins Hospital. METHODS: Retrospective chart review. SETTING: The Johns Hopkins Hospital. RESULTS: Sinusitis was identified in 72 HIV-infected patients, predominantly individuals with a CD4 cell count of less than 200/mm3. A history of respiratory infections such as bacterial pneumonia, bronchitis, and otitis media was common. Although nasal congestion and postnasal drainage were found in the majority of patients, symptoms of sinusitis were often nonspecific and the diagnosis was incidental in 28 patients (33%). Magnetic resonance imaging or computed tomography was significantly more sensitive than plain radiography (p less than 0.001) in defining the extent of the disease, particularly with posterior sinus involvement, which occurred in the majority of the patients. The number of radiologically abnormal sinuses correlated inversely with the CD4 count. Although the majority of patients responded at least partially to antibiotic therapy, only 15% had complete resolution of clinical symptoms. Fifty-eight percent of patients had clinical and/or radiographic evidence of recurrent/persistent sinus infection, and chronicity correlated with a CD4 count less than 200/mm3 (p less than 0.001). CONCLUSIONS: Sinusitis in HIV-infected patients is common, severe, and difficult to treat. Patients with CD4 counts less than 200/mm3 are prone to disease involving multiple sinuses that responds incompletely to antibiotic therapy, often resulting in chronic sinusitis. Unlike the immunocompetent host, the majority of the HIV-infected patients with advanced immunodeficiency develop posterior sinus disease.

Human colonic epithelial cells, HT29/C1, treated with crude Bacteroides fragilis enterotoxin dramatically alter their morphology.

Bacteroides fragilis has been associated with causation of diarrheal disease in livestock and humans. To date, conventional tissue culture and animal assays used to detect the biologic activity of bacterial enterotoxins have failed with enterotoxigenic B. fragilis. Although enterotoxigenic B. fragilis stimulates intestinal secretion in lamb and calf ligated intestinal loops, infant rabbits, and adult rabbits with ligated ceca, these animal systems are costly and complicated, which limits their usefulness for identification of enterotoxigenic B. fragilis strains. Using the cloned human colonic-epithelial-cell line HT29/C1, we have developed an in vitro assay that is 89% sensitive and 100% specific in detecting enterotoxigenic B. fragilis strains as defined by the lamb ligated-intestinal-loop assay. Subconfluent HT29/C1 cells treated with concentrated bacterium-free culture supernatants of enterotoxigenic B. fragilis strains develop specific and striking morphologic changes including loss of cell-to-cell attachments, rounding, swelling, and, in some cases, pyknosis. These morphologic changes are initially visible at 1 h after treatment and progress over at least the first 24 h. This tissue culture assay should prove useful in epidemiologic studies of enterotoxigenic B. fragilis and may facilitate basic studies to identify the B. fragilis toxin(s) and its mechanism of action.

Characterization of the synergistic interaction of Escherichia coli heat-stable toxin and carbachol.

STa, the heat-stable enterotoxin of Escherichia coli, is a specific activator of membrane-bound guanylyl cyclase and stimulates secretion of Cl- in a human colonic carcinoma cell line (T84). We investigated the effect of the cholinergic agent carbachol on the secretory response to STa. T84 cell monolayers were studied under voltage-clamped conditions in modified Ussing chambers. Simultaneous addition of STa and carbachol resulted in a biphasic synergistic response characterized by a brief peak in short-circuit current (Isc) followed by a prolonged plateau phase lasting up to 90 min. A synergistic response was also seen with sequential addition of the agonists, and was altered by the order and timing of agonist addition. Pretreatment with STa enhanced the synergistic response to carbachol, while the reverse order of additions produced synergy only when STa was added during or immediately after the Isc response to carbachol. Synergy occurred only with a concentration of STa sufficient to produce an Isc response alone. However, a concentration of carbachol that caused neither an increase in Isc nor intracellular Ca2+ mobilization was sufficient to evoke a synergistic response. Addition of 8-bromoguanosine 3',5'-cyclic monophosphate also produced a synergistic Isc response with carbachol, although maximal synergism was seen with simultaneous addition. Augmentation of the intracellular Ca2+ response to carbachol by STa is not the mechanism of synergy. Although the mechanism of synergy is not understood, these studies suggest that STa-induced cGMP interacts with other second messengers to produce the synergistic response, and that multiple intracellular mediators may influence the ability of STa to cause disease.

Intestinal cryptosporidiosis: pathophysiologic alterations and specific cellular and humoral immune responses in rnu/+ and rnu/rnu (athymic) rats.

In order to develop an experimental model of symptomatic cryptosporidiosis in an immunosuppressed mammal, we investigated the pathophysiology of infection with Cryptosporidium and the humoral and cellular host responses in rnu/rnu (athymic) rats and their heterozygous (rnu/+) littermates by challenging suckling rats with greater than or equal to 2.5 x 10(6) Cryptosporidium oocytes oro-gastrically. Normal and immunodeficient animals were followed for onset and duration of infection (fecal oocysts), physiologic consequences (diarrhea, impaired weight gain, brush-border enzyme activities), and immunologic response (both B- and T-lymphocyte-mediated). Homozygosity for the rnu gene was associated with protracted cryptosporidial infections; shedding for up to 52 days occurred, and delay in weight gain was noted in rnu/rnu-infected compared with rnu/rnu-uninfected rats (p less than 0.05). In contrast, cryptosporidial challenge of rnu/+ rats resulted in self-resolving infections, occasionally with transient diarrhea lasting four days or less occurring 10-15 days after oro-gastric challenge. The latter animals mounted a cell-mediated immune response to Cryptosporidium: three months after challenge, five of five rnu/+ rats demonstrated positive skin test responses to a subcutaneous 3.5 micrograms dose of cryptosporidial antigen. Further, sera from 6 rnu/+ rats taken two to three months after oro-gastric oocyst challenge exhibited specific anticryptosporidial immunoglobulin binding (A405 = 0.96), compared to that of seven uninfected rnu/+ controls (A405 = 0.09, P less than 0.02). Macromolecules of 150, 105, and 88 kD in the Cryptosporidium antigen preparation were bound by serum immunoglobulin from previously infected, recovered rnu/+ rats. Two brush-border enzymes (lactase and alkaline phosphatase) were markedly reduced in the ileum 8-10 days after oro-gastric challenge in rats with diarrhea and oocyst shedding. We find the rnu/rnu (athymic, nude) rat provides a useful model for study of prolonged cryptosporidial infection with impaired weight loss, brush-border enzyme alteration and intermittent diarrhea. These studies further suggest that a T-lymphocyte population is involved in recovery from Cryptosporidium infection and that this recovery is associated with both cellular and humoral immune responses to specific cryptosporidial antigenic macromolecules. This model should open further avenues for the study of the pathogenesis and protective immunity in cryptosporidial infection.

Carbachol mimics phorbol esters in its ability to enhance cyclic GMP production by STa, the heat-stable toxin of Escherichia coli.

STa, the heat-stable enterotoxin of Escherichia coli, stimulates membrane-bound guanylate cyclase in enterocytes, elevates cyclic GMP, and results in intestinal secretion of ions and fluid. Using the T84 colon carcinoma cell line as a model. Weikel et al. reported that phorbol esters enhance STa-stimulated cyclic GMP production by 60-140% [(1990) Infect. Immun. 58, 1402-1407]. In the present report we demonstrate that the acetylcholine analog carbachol enhanced toxin-stimulated cyclic GMP accumulation in intact T84 cells by 50-100% and that this effect was blocked by 10 microM atropine and 10 microM sphingosine. Pertussis toxin treatment of the T84 cells did not affect the subsequent response to carbachol. Carbachol, which elevates intracellular calcium in these cells, may act through protein kinase C to enhance cyclic GMP production.

Phorbol esters enhance the cyclic GMP response of T84 cells to the heat-stable enterotoxin of Escherichia coli (STa).

We examined the effect of protein kinase C (PKC) activation on the cyclic GMP response to heat-stable enterotoxin (STa) in a colonic carcinoma intestinal epithelial cell line, T84 cells. Our results demonstrate that the active phorbol ester analog, phorbol dibutyrate, but not the inactive alpha-phorbol dibutyrate, acts synergistically with STa to elevate cyclic GMP in intact T84 cells. The effect is observed in the absence or presence of the phosphodiesterase inhibitor, isobutylmethylxanthine, which suggests that phorbol dibutyrate modifies cyclic GMP synthesis rather than cyclic GMP degradation. In contrast to several systems in which prolonged treatment with phorbol ester desensitizes PKC-mediated responses, the cyclic GMP response in T84 cells is not diminished by prolonged treatment of T84 cells with phorbol dibutyrate. Also, transient treatment of T84 cells with phorbol dibutyrate enhances subsequent STa-stimulated cyclic GMP accumulation. These observations suggest that PKC activation produces a long-lived signal in T84 cells which enhances cyclic GMP accumulation in response to STa. Second messenger "cross talk" [T. Yoshimasa, D. R. Sibley, M. Bouvier, R. J. Lefkowitz, and M. G. Caron, Nature (London) 327:67-70, 1987] may be important in the pathogenesis of diarrheal disease.

Failure of pertussis toxin to inhibit activation of guanylate cyclase by the heat-stable enterotoxin of Escherichia coli (STa) in the T84 cell line.

The heat-stable enterotoxin (STa) of Escherichia coli causes intestinal secretion by stimulating guanylate cyclase, an enzyme believed to be distinct from the STa receptor. Pertussis toxin (PT) has been reported to block the ability of STa to stimulate guanylate cyclase in rat intestinal mucosa (S. A. Epstein, R. A. Giannella, and H. J. Brandwein, FEBS Lett. 203:44-48, 1986). This suggested that a guanine nucleotide regulatory protein (G protein) coupled the STa receptor to guanylate cyclase, a function not previously recognized for G proteins. We sought to explore this phenomenon and, if possible, to identify this G protein. Initial experiments with the human colon carcinoma cell line T84 revealed that higher-than-expected concentrations (1 micrograms/ml) of PT were needed to intoxicate cells, as assessed by ADP-ribosylation of endogenous PT substrate, but that 99 to 100% intoxication could be achieved. Homogenates made from fully intoxicated cells did not differ from controls in basal or STa-stimulated guanylate cyclase activity, and cyclic GMP accumulation in intact T84 cells was not changed by PT treatment. We conclude that a PT-sensitive G protein is not involved in the stimulation of cyclic GMP production by the enterotoxin STa.

Phorbol esters specifically enhance the cytolytic activity of Entamoeba histolytica.

Entamoeba histolytica causes invasive amebiasis by lysis of host tissue and inflammatory cells. The in vitro cytolysis of target Chinese hamster ovary (CHO) cells by axenic E. histolytica trophozoites (strain HM1:IMSS) is a calcium- and phospholipase A-dependent event initiated by the binding to the target cell of the galactose-inhibitable surface lectin of the parasite. We utilized phorbol esters as a probe to determine whether an amebic protein kinase C has a role in the cytolytic event. The addition of phorbol 12-myristate 13-acetate (PMA) at 10(-6) or 10(-7) M resulted in a greater than twofold enhancement of amebic killing of target CHO cells over 30 min (P less than 0.01). Prior exposure of only the amebae, but not the CHO cells, to PMA produced a similar effect (P less than 0.01). The inactive analog 4-alpha-phorbol had no effect on amebic killing of CHO cells. The PMA-mediated enhancement of amebic cytolysis persisted for up to 60 min after a 5-min exposure; however, after a 30-min exposure to PMA (10(-6) M) there was no augmentation of amebic killing of CHO cells. PMA (10(-6) M) did not promote adherence of parasites to CHO cells but did enhance amebic cytolysis of previously adherent target cells (P less than 0.01). Sphingosine, a specific inhibitor of protein kinase C, abolished both the PMA-stimulated and the basal cytolytic activity of E. histolytica. PMA enhanced CHO cell cytolysis by the less virulent wild-type strain H-303:NIH (P less than or equal to 0.02) but did not augment the activity of the less virulent strain H-200:NIH or two avirulent clones of HM1 (L6 and C919). In summary, these experiments with the phorbol esters and sphingosine as probes to modulate the activity of protein kinase C indicate participation of a parasite protein kinase C in the cytolytic activity of virulent, axenic E. histolytica trophozoites and thus in the pathogenesis of amebiasis.

In vivo and in vitro effects of a novel enterotoxin, STb, produced by Escherichia coli.

Escherichia coli may produce a heat-labile enterotoxin (LT) or two heat-stable enterotoxins (STa, STb). STb consistently causes secretion in vivo in 5-hr weaned-pig intestinal loops (P less than .0001). In Ussing chambers in vitro, crude culture filtrates of STb initiate a prompt increase in short circuit current (SCC; P less than or equal to .0001) and potential difference (P less than or equal to .0001) when compared with nontoxigenic culture filtrates. In bidirectional in vitro studies of ion flux (22Na and 36Cl), STb did not alter 22Na or 36Cl unidirectional or net fluxes. The calculated residual ion flux increased significantly (P less than .03), however, in tissues treated with STb and fully accounted for the STb-induced increase in SCC. Measurement of the electrolyte content of ligated intestinal segments in vivo further suggested that STb stimulated bicarbonate secretion. Relative to controls, significant accumulation of Na and Cl also occurred intraluminally in vivo. These data indicate that STb is a unique enterotoxin that causes net secretion in pig jejunum in vivo. In vitro and in vivo studies show that STb stimulates active secretion of nonchloride anion. We postulate that STb causes active bicarbonate secretion in weaned-piglet jejunum.

Species specificity and lack of production of STb enterotoxin by Escherichia coli strains isolated from humans with diarrheal illness.

Escherichia coli strains produce at least two heat-stable enterotoxins, STa and STb. STa is well known to be important in the pathogenesis of human diarrheal disease; the role of STb has not been defined. Fifty-two E. coli strains recovered from human diarrheal illness in northeast Brazil or Bangladesh were examined in weaned porcine ligated intestinal segments for STb activity. A total of 113 E. coli strains from human diarrheal disease in northeast Brazil and 28 E. coli strains from Bangladesh were examined for DNA hybridization to a STb gene probe. None of these strains produced STb as detected by enterotoxic activity or by the gene probe. We also examined adult human ileal mucosa for responses to STb in the Ussing chamber in vitro. In contrast to piglet jejunum, which consistently responds electrogenically to crude STb, human ileal tissue showed no response to STb but responded electrogenically to theophylline (10 mM). These results suggest that STb-producing E. coli strains are not a major cause of diarrheal illness in humans.

Stimulation of porcine jejunal ion secretion in vivo by protein kinase-C activators.

Microbial toxins act through cyclic nucleotide dependent (cAMP or cGMP) or cyclic nucleotide independent pathways to cause intestinal ion secretion. To explore the calcium dependent, cyclic nucleotide independent pathway that is postulated to involve protein kinase C activation, we measured protein kinase C activity and phorbol ester binding in isolated intestinal epithelial cells and examined the effects of the C-kinase activators, phorbol myristate acetate, phorbol dibutyrate, and 4-beta-phorbol-12,13-didecanoate, in weaned pig jejunum in vivo. We demonstrated both protein kinase C activity and specific phorbol ester binding in porcine jejunal epithelial cells. Phorbol myristate acetate, phorbol dibutyrate, and 4-beta-phorbol-12,13-didecanoate (10(-5) M) each caused striking secretory responses at 5 h with accumulation of Na+, K+, Cl-, and HCO3- intraluminally. In contrast, 4-alpha-phorbol and 4-alpha-phorbol-12,13-didecanoate, which do not affect protein kinase C, allowed normal net absorption of all electrolytes from the intestinal lumen equivalent to controls with only Ringer's lactate. Time course studies revealed significant secretion within 30 min after exposure to the C-kinase activators. These data suggest an important role for protein kinase C activation in intestinal ion secretion.

Unexplained persistent lymphadenopathy in homosexual men and the acquired immune deficiency syndrome.

Ninety-three homosexual men with persistent lymphadenopathy were followed at the Memorial Sloan-Kettering Cancer Center for a mean period of 20.8 months. Histories and serologic evidence of a number of previous infections were common, but the lymphadenopathy was not due to recognizable infections or neoplastic disease. Leukopenia, lymphopenia, granulocytopenia, monocytopenia, decreased ratios of T-helper to T-suppressor cells, decreased natural killer cell activity and increased serum immunoglobulin concentrations were common. Lymph node biopsies showed reactive hyperplasia without any unique histopathologic features. Antibody to the human T-lymphotropic virus-III (HTLV-III or LAV), a newly described retrovirus believed to be the etiologic agent of the acquired immune deficiency syndrome (AIDS), was detected in 91.4%. Over a 3-year period, 11 cases of AIDS were recognized in these patients: Kaposi's sarcoma developed in 7 and opportunistic infections in 4. The lymphadenopathy resolved in six patients and the others remained unchanged. Although most of these patients are asymptomatic and remain well, the risk of AIDS in this group of men was higher than in other groups of homosexual men in New York.

STb enterotoxin of Escherichia coli: cyclic nucleotide-independent secretion.

Escherichia coli may produce a heat-labile enterotoxin (LT) or two heat-stable enterotoxins (STa, STb). Experimentally, STb is consistently active only in 5 h-weaned pig intestinal loops (WPIL), an effect that is largely removable by rinsing. At least three mechanisms initiate small intestinal secretion: cyclic AMP (LT), cyclic GMP (STa) and calcium (A23187). All three increase short-circuit current (SCC) in Ussing chambers by stimulating net Cl- secretion. STb significantly increases SCC within 2-5 minutes in Ussing chambers and is independent of cyclic AMP and cyclic GMP. When compared to crude culture filtrates of a non-toxigenic strain of E. coli, crude culture filtrates of STb did not alter Na+ or Cl- undirectional or net fluxes. However, the calculated residual ion flux (JRnet) increased significantly in STb-treated tissues and appeared to largely account for the STb-induced increase in SCC. Furosemide applied serosally (10(-3) M), the removal of extracellular calcium, and lanthanum chloride (10(-3) M) did not inhibit the effect of STb on SCC. Chlorpromazine (0.4 mM) completely inhibited STb-induced secretion in porcine loops. This inhibition was a non-specific reversal of the STb effect because in Ussing chambers, chlorpromazine simply induced an equal and opposite effect on SCC. These results indicate that STb initiates intestinal secretion in porcine jejunum in vitro by stimulating primarily non-chloride anion secretion in the absence of extracellular calcium. We postulate that STb causes bicarbonate secretion by a mechanism distinct from those of previously studied enterotoxins.

Does enteropathogenic Escherichia coli produce heat-labile enterotoxin, heat-stable enterotoxins a or b, or cholera toxin A subunits?

Although most enteropathogenic Escherichia coli strains do not produce recognized enterotoxins, we wished to examine whether they produce any factors like heat-stable enterotoxin b or cholera toxin active subunits that might be missed by conventional assay methods. E. coli strains E851 (O142) and E2348 (O127) that had caused diarrhea in volunteers were negative for heat-labile enterotoxin and heat-stable enterotoxin a in Chinese hamster ovary cell and suckling mouse assays, failed to cause secretion in ligated small bowel loops from 6- to 8-week-old pigs after 4 to 5 h (used to show heat-stable enterotoxin b), and did not activate adenylate cyclase in pigeon erythrocyte lysates (used to demonstrate cholera toxin A subunit). We conclude that crude, unconcentrated culture filtrates and sonicates do not mimic heat-labile or heat-stable enterotoxins or cholera toxin or its A subunit and that enteropathogenic strains of E. coli probably have yet another mechanism or group of mechanisms by which they cause diarrhea.