The immunological response to syphilis differs by HIV status; a prospective observational cohort study

Values are summarized as medians and interquartile ranges (IQR). In keeping with previous studies, we found that the immunological profiles of syphilis differed between primary and secondary stage syphilis combined group (PSS) and early and late latent stage syphilis combined group (LS) [20, 21]. All results are therefore presented stratified by PSS versus LS. There were only two HIV-uninfected patients diagnosed with LS and no comparisons are therefore made with this group. Comparison of the inter-individual data between the PSS, and LS groups was performed using the Mann-Whitney U-test. Wilcoxon signed rank test was used to evaluate intra-individual changes from baseline to six-month visit. Correlations were assessed via Spearman’s rank correlation coefficient. Due to the small sample size and the fact that this is the first study of its kind, the study was conceived as being exploratory rather than hypothesis testing. This, combined with the fact that the cyto- and chemokines under investigation are known to be inter-dependent, led us to not utilize Bonferroni corrections [28‐30]. All analyses were performed in Stata 13 (StataCorp LP, College Station, TX, USA). A P-value of less than 0.05 was considered statistically significant.

At the baseline visit, compared to the control group, concentrations of the anti-inflammatory cytokine IL-10 were significantly elevated in both the HIV-infected and uninfected groups (Fig. 1; Table 2). The level of IL-10 was strikingly higher in the HIV-infected PSS compared to the HIV-uninfected PSS group (25.3 pg/mL (IQR, 4.56–41.76) versus 2.73 pg/mL (IQR, 1.55–9.02), P = 0.0192). In the HIV-infected PSS group (but not the HIV-infected LS or HIV-uninfected PSS groups) the chemokines IP-10, MIP-1β, IL-8 and the pro-inflammatory cytokine IL-6 were raised significantly compared to the controls. Of these only the MIP-1β level was significantly higher in the HIV-infected PSS than uninfected PSS group (P = 0.0205).

In seven out of 91 individuals with syphilis we had baseline but no 6-month samples. At 6 months post syphilis treatment the levels of IL-10 had declined in all groups, and most strikingly in the HIV-infected PSS group. IL-10 remained significantly elevated compared to the control group (P < 0.0001 to 0.0018). There was no longer a difference in IL-10 levels between the HIV-infected and uninfected PSS groups. Another interesting observation was that IFNα levels declined in all groups and at 6 months attained significantly lower levels than the HIV-infected control group (P = 0.0007 to 0.0104).

At the baseline visit there were differences between the HIV-infected and uninfected groups in the immunological markers that were correlated with IL-10 level. In the HIV-infected PSS group, IL-10 was significantly positively correlated with IL-12p40, IL-12p70, MIP-1α, MIP-1β, IL-5, IL-6, IL-7 and IL-8 (r = 0.36, 0.67, 0.59, 0.61, 0.75, 0.79, 0.83, 0.53, respectively; Additional file 1: Table S2a). In the HIV-uninfected group there were no significant correlations with IL-10 (Additional file 1: Table S2b). There were however only 10 individuals in this group.

None of the cohort was diagnosed with treatment failure or neurosyphilis subsequent to the baseline visit and we were thus unable to assess if any particular immunological profile was predictive of these adverse outcomes.

In sensitivity analyses we excluded the one woman from the analyses and repeated the analyses with the PSS group split into primary syphilis (PS) and secondary syphilis (SS) and the LS group split into early latent and late latent syphilis. We also repeated the analyses with the HIV-infected group limited to those taking ART. These made little difference to the results (results not shown). Scatterplots revealed that the difference in IL-6 between controls and the HIV-infected PSS group was largely driven by two outlier values. Repeating the comparison between these two groups excluding these two outliers, there was no longer a statistically higher IL-6 level in the HIV-infected group.

The initial immune response to T. pallidum involves a robust cell mediated response characterized by a Th1 predominance that results in the clearance of the vast majority of treponemes at the lesion of the primary stage [15, 32]. We mainly found increased levels of IL-10 and chemokines in the PSS group. The positive correlations we found between IL-10 and IL-12p40, IL-12p70, MIP-1α, MIP-1β, IL-5, IL-6, IL-7 and IL-8 in the HIV-infected PSS group could be interpreted as evidence that the increased IL-10 is acting as a feedback response to limit inflammation. We did not see any particular increase of Th1 cytokines as reported by two other studies [15, 29].

An alternative/complementary explanation is that T. pallidum is playing a role in the IL-10 elevation. In a proportion of cases, T. pallidum is able to evade the immune system and set up secondary and chronic latent stages of infection [14]. Poor antigenicity and antigenic variation have been shown to play an important role in this regard [33‐35]. A further mechanism is that an enhanced regulatory T-cell (Treg) response in early syphilis may down-regulate the immune response to T. pallidum and thereby facilitate its survival [17, 31, 36]. IL-10 plays an important role in the establishment of the Treg anti-inflammatory effect that is important in preventing overwhelming inflammatory responses [37]. These immunosuppressive properties of IL-10 have also been shown to be harnessed by a number of pathogens to facilitate persistent infection: Borrelia burgdorferi [38], Plasmodium spp. [39], Leishmania spp. [40] and Mycobacterium tuberculosis [37].

Babolin et al. provided experimental evidence that the T. pallidum protein, TpF1 bacterioferrin, is able to induce a Treg response in patients with secondary syphilis via the induction of IL-10 and TGF-β [41]. Podwinska and colleagues investigated the ability of lymphocytes to produce cytokines from patients at different stages of syphilis in response to stimulation with T. pallidum antigen [20]. They noted a stepwise increase in IL-10 production from primary syphilis to secondary and latent syphilis that correlated inversely with the ability to produce IL-2 (a Th1 cytokine). In a study of 531 syphilis patients with and without neurosyphilis, Li and colleagues found evidence that Tregs play a role in treponemal persistence and the genesis of neurosyphilis in HIV-uninfected individuals [17]. They demonstrated that secondary and serofast syphilis patients had increased Treg percentages in their peripheral blood compared with healthy controls. Patients with neurosyphilis had a higher frequency of Tregs in peripheral blood compared to non-neurosyphilis patients. In a similar vein, Pastuszczak et al., found that IL-10 was considerably higher in the CSF of patients with neurosyphilis than syphilis patients without neurological involvement [19]. These findings are also consistent with the clinical study by Knudsen et al., which was able to demonstrate with a high degree of probability that T. pallidum infection was the cause of the IL-10 elevation [21].

HIV-infection is characterized by a Th1 to Th2 cytokine shift during the course of the infection [42] which includes an enhanced potential for robust IL-10 and associated Treg response [43‐45]. If HIV infection leads to a more pronounced Treg response to infection with T. pallidum, then this could explain a number of the atypical features of syphilis in HIV-infected individuals mentioned in the introduction. The recently established correlation between Treg activity and risk of serofast syphilis and neurosyphilis [17] in HIV-uninfected individuals suggests the need for more investigations along these lines in HIV co-infected patients. The discovery of an immunological profile strongly associated with asymptomatic neurosyphilis would be of considerable clinical utility.

In addition, the small sample size and the fact that almost the entire cohort were MSM, with high-risk behavior, thereby reducing the generalizability of the results. We did not resample the control group at 6-months and thus did not have a control group for the 6-month samples. The fact that those diagnosed with syphilis reported a higher number of sexual partners than the controls raises the possibility that they were co-infected with other undiagnosed STIs. Given the fact that there was no statistically significant difference in the number of partners, or the diagnosis of STIs between the PSS HIV-infected and uninfected subjects, this explanation is unlikely to explain the immunological differences found between these two groups. Our HIV-infected patients included both those receiving and not receiving ART, which may have influenced our results, however, we consider this as unlikely as repeating the analyses restricted to those on ART did not substantively change the results. We did not assess patients’ genotypes for IL-10 and other cytokines. A number of IL-10 single nucleotide polymorphisms have been found to be associated with both IL-10 inducibility [19] and susceptibility to infectious diseases such as tuberculosis [46] and neurosyphilis [19]. For a number of the chemo- and cytokines, only a small proportion of samples had levels above the level of detection (Additional file 1: Table S1), particularly in the cases of IL1b, IL-6, IL-17A, IL-7, IL-12p40, IL-4 and IL-5. A number of these low abundance chemo- and cytokines had low correlations between technical duplicates (Additional file 1: Table S1). The results for these chemo- and cytokines need to be interpreted with considerable caution. Furthermore, we did not correct for multiple comparisons which may have led to type I errors [28]. The strengths of the study include its setting within a prospective observational cohort study wherein the patients had their behavioral, clinical and laboratory characteristics collected in a detailed and standardized fashion.