In the present study, we sought to test the hypothesis that GAS was present within or on pediatric palatine tonsils and to see if we could identify evidence of biofilm formation. To achieve this, we analyzed surgically excised tonsils from 30 pediatric patients undergoing tonsillectomy due to recurrent GAS tonsillopharyngitis. Originally, we planned to examine a limited number of surgically excised tonsils from patients undergoing tonsillectomy for ATH as these patients were asymptomatic for GAS infection, and thus we thought these samples would provide a negative control for the detection of GAS. However, we readily detected the presence of GAS in these samples by immunofluorescence and scanning electron microscopy revealed the presence of biofilms made up of chains of cocci which morphologically resembled GAS. Given this result, we examined the tonsils excised from a total of 30 patients presenting with ATH in addition to the tonsils excised from 30 patients presenting with recurrent GAS tonsillopharyngitis.
We discovered that a similar proportion of tonsils from children with either ATH or recurrent GAS tonsillopharyngitis were positive for GAS by immunofluorescence (~35% positive). Of note, our method of immunofluorescence allowed the dual detection of both GAS and the cellular markers of the reticulated crypt epithelium, cytokeratin 8 & 18. By this method, we were able to show that the GAS detected had colonized throughout the tonsillar crypts in both sets of patient tonsil samples.
The finding that tonsillar tissue from children with ATH patients showed GAS in such a high percentage was unexpected. However, bacterial carriage by children with ATH is not unprecedented. Several previous studies of adenotonsillar hypertrophy have provided evidence that increased numbers of pathogenic bacteria can be recovered from homogenized hypertrophic tonsil cores compared to swabs of those tonsils alone [28
]. Indeed, it is proposed that lymphoid hyperplasia (chronic enlargement) is correlated with increased bacterial load and increased B- and T-lymphocyte proliferation [30
]. This phenomenon has been associated with a number of bacteria including Staphyloccocus aureus
, Haemophilus influenzae, S. pneumoniae
, as well as GAS [28
]. However, a review of the literature revealed that the percentage of hypertrophic tonsils positive for GAS by Brodsky et al. (16%) and Stjernquist-Desatnik et al. (20%) was lower than the 36.7% positive that we observed [28
]. Our method of detection is more sensitive (antibody based immunofluorescence vs. culture), but this difference in frequency of detection may also be due to geographic location or the fact that our study occurred almost 20 years after the reports referenced.
Our results support the hypothesis that GAS colonize pediatric palatine tonsils as a biofilm. SEM clearly reveals the presence of 3-dimensional communities of chains of cocci in tonsils that had tested positive for the presence of GAS by immunofluorescence. These structures closely resemble the in vivo GAS biofilms grown ex vivo on pig epithelium. Furthermore, Gram-staining reveals the presence of microcolonies of Gram-positive cocci indicative of biofilms in samples that tested positive for the presence of GAS. However, despite the fact that these samples were positive for GAS by immunofluorescence, we cannot rule out at this juncture that the biofilms observed by SEM or Gram-staining were not GAS.
Detailed information regarding antibiotic exposure prior to tonsillectomy was not collected; however, it is known that 100% of the children undergoing tonsillectomy for recurrent GAS tonsillopharyngitis had experienced a recent GAS infection. The finding that GAS were detected in roughly equivalent percentages in these two patient groups is consistent with the hypothesis that biofilms may be important in carrier state antibacterial resistance.
The high rate of asymptomatic GAS colonization that is presented here also has implications regarding the utility of rapid antigen tests and cell culture. The question of what is a true positive vs. a clinical false positive, especially in light of the potential for co-colonization by a viral pathogen, is a difficult one. Now that we have established the ability to rapidly detect the presence of GAS within the tonsillar crypt, our emphasis will turn to the collection and typing of these strains, analysis of the frequency of their isolation, as well as an elucidation of what makes up the GAS biofilm structure and how it is regulated. Specifically, our findings contribute to an understanding of GAS tonsillar colonization. Developing the capacity to distinguish patients with GAS tonsillopharyngitis from those with GAS colonization, or those with GAS colonization and viral tonsillopharyngitis is a clinically important goal that could greatly reduce unnecessary antibiotic use. This work may ultimately contribute to the development of clinically useful methods for identifying patients with longstanding GAS colonization.