Advertisement

Evaluation of cerebrospinal fluid UCH-L1, GFAP, and NF-L as novel markers for the diagnosis of neurosyphilis among HIV-negative patients

  • Author Footnotes
    † These authors have contributed equally to this work and share first authorship
    Rui Chen
    Footnotes
    † These authors have contributed equally to this work and share first authorship
    Affiliations
    Center of Clinical Laboratory, Zhongshan Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, China
    Search for articles by this author
  • Author Footnotes
    † These authors have contributed equally to this work and share first authorship
    Li-Rong Lin
    Footnotes
    † These authors have contributed equally to this work and share first authorship
    Affiliations
    Center of Clinical Laboratory, Zhongshan Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, China

    Institute of Infectious Disease, School of Medicine, Xiamen University, Xiamen, China
    Search for articles by this author
  • Yao Xiao
    Affiliations
    Department of Hospital Infection Management, Zhongshan Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, China
    Search for articles by this author
  • Wu-Jian Ke
    Correspondence
    Corresponding author. Wu-Jian Ke, Dermatology Hospital, Southern Medical University, Guangzhou, China.
    Affiliations
    Dermatology Hospital, Southern Medical University, Guangzhou, China
    Search for articles by this author
  • Tian-Ci Yang
    Correspondence
    Corresponding author. Tian-Ci Yang, Center of Clinical Laboratory, Zhongshan Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, China, Institute of Infectious Disease, School of Medicine, Xiamen University, Xiamen, China. Phone: +86-592-2993042; Fax: +86-592-2993043.
    Affiliations
    Center of Clinical Laboratory, Zhongshan Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, China

    Institute of Infectious Disease, School of Medicine, Xiamen University, Xiamen, China

    Xiamen Clinical Laboratory Quality Control Center, Xiamen, China
    Search for articles by this author
  • Author Footnotes
    † These authors have contributed equally to this work and share first authorship
Open AccessPublished:November 15, 2022DOI:https://doi.org/10.1016/j.ijid.2022.11.013

      Abstract

      Objectives

      To evaluate the possibility of using cerebrospinal fluid ubiquitin C-terminal hydrolase L1 (CSF UCH-L1), glial fibrillary acidic protein (GFAP), and neurofilament light protein (NF-L) for the diagnosis of neurosyphilis (NS).

      Methods

      A cross-sectional study of 576 subjects was conducted at Zhongshan Hospital from January 2021 to August 2022, to evaluate the diagnostic accuracy of CSF UCH-L1, GFAP, and NF-L for NS and analyse their correlations with CSF rapid plasma reagin (RPR), white blood cells (WBC), and protein.

      Results

      NS patients had higher CSF UCH-L1, GFAP, and NF-L levels than syphilis/non-NS and non-syphilis patients. Using a cut-off point of 652.25 pg/mL, 548.89 pg/mL, and 48.38 pg/mL, CSF UCH-L1, GFAP, and NF-L had a sensitivity of 85.11%, 76.60%, and 82.98% with a specificity of 92.22%, 85.56%, and 91.11%, respectively, for NS diagnosis. Moreover, parallel and serial testing algorithms improved their sensitivity and specificity to 93.62% and 98.89%, respectively. Interestingly, levels between CSF RPR-positive and-negative NS patients did not differ and showed weak or moderate correlation with WBC and CSF protein in syphilis patients.

      Conclusion

      CSF UCH-L1, GFAP, and NF-L can be used as novel markers for the diagnosis of NS, independent of CSF RPR, WBC, and proteins.

      Keywords

      1. Introduction
      Treponema pallidum can affect the central nervous system at any time after the initial infection, leading to neurosyphilis (NS), which earned the name the Great Imitator or the Great Mimicker (Peeling et al. 2017; Ghanem, Ram, and Rice 2020). The diagnosis of NS is based on clinical symptoms and examination of the cerebrospinal fluid (CSF) using treponemal and non-treponemal tests, such as the Treponema pallidum particle assay (TPPA), fluorescent treponemal-antibody absorption (FTA-ABS), Venereal Disease Research Laboratory (VDRL), rapid plasma reagin (RPR), and elevation of inflammatory parameters (CSF white blood cells [WBC] and proteins). However, no single test can currently be used to diagnose NS in all instances (Workowski et al. 2021). A meta-analysis showed that the specificity of CSF VDRL was 74-100%, but the sensitivity was low (49-87%), leading to missed diagnosis (Tuddenham, Katz, and Ghanem 2020). Usually, NS can be diagnosed with reactive CSF VDRL, which can be replaced by CSF RPR in most instances (Janier et al. 2020), but negative CSF does not exclude NS. When CSF VDRL or CSF RPR is negative, the combination of positive CSF FTA-ABS or TPPA, increased WBC, and protein levels are usually used to assist NS diagnosis, but suffers from poor specificity leading to misdiagnosis owing to the passive transfer of immunoglobulins across the blood-CSF barrier (Zhang et al. 2013b; Park et al. 2020; Ghanem, Ram, and Rice 2020; Reiber and Peter 2001). Increased WBC and protein levels are indicative of an inflammatory reaction in the CSF and are not specific markers for NS (Workowski et al. 2021). To date, no reliable test for the diagnosis of NS has been developed. Therefore, researchers are seeking diagnostic indicators to avoid missed diagnosis or misdiagnosis for the early recognition of NS.
      Neuronal damage is an important pathological feature of NS (Li et al. 2020), thus the markers associated with neuronal damage may be helpful in the diagnosis of NS. Neuronal damage markers have been suggested for the diagnosis of multiple sclerosis and Wilson's disease (Disanto et al. 2017; Siller et al. 2019; Lin et al. 2021). Ubiquitin C-terminal hydrolase L1 (UCHL1) is a compact, nearly spherical, small ubiquitin protein highly specifically expressed by neurons, which is essential for removing abnormal proteins and preventing the accumulation of potentially toxic proteins in neuronal cells (Graham and Liu 2017; Bilguvar et al. 2013). It is associated with the occurrence and development of neurodegenerative diseases such as Alzheimer's disease, Parkinson's disease, and amyotrophic lateral disease (Setsuie and Wada 2007). Glial fibrillary acidic protein (GFAP) is a marker of astrocyte proliferation, which is involved in cytoskeleton formation and maintenance of tensile strength, and its expression is increased when neurons are damaged (Oeckl et al. 2019; Olsson et al. 2016; Heller et al. 2020). Neurofilament light chain protein (NF-L) is an important part of the neuronal cytoskeleton and plays a key role in axonal radial growth and stability, thus ensuring proper nerve signal transmission (Khalil et al. 2018). It is abnormally elevated in neurodegenerative, inflammatory, vascular, and traumatic diseases (Disanto et al. 2017; Rohrer et al. 2016). UCH-L1, GFAP, and NF-L have been shown to be indicators of neuronal damage in many central nervous diseases (Shahim et al. 2020; Bazarian et al. 2018; Newcombe et al. 2022), but the diagnostic utility of UCH-L1, GFAP, and NF-L in NS remains unclear. This study aimed to analyse the levels of CSF UCH-L1, GFAP, and NF-L in NS patients and explore their diagnostic value in NS patients to better recognise NS and avoid missed diagnosis or misdiagnosis.
      2. Materials and methods
      2.1 Study participants
      A total of 576 participants with neurological disorders were included in this cross-sectional study conducted at Zhongshan Hospital of Xiamen University, School of Medicine, Xiamen University, from January 2021 to August 2022. The participants’ clinical information, including physical examination, laboratory investigation, and medical history, were extracted and recorded. The following participants were excluded from this study: those without a lumbar puncture, returning patients, and those with HIV infection. According to the guidelines of NS in the UK, USA, Europe, and Canada (Kingston et al. 2016; Workowski et al. 2021; Janier et al. 2020; Levett et al. 2015), NS was defined as syphilis at any stage with a positive CSF TPPA result and one or more of the following findings: (1) reactive CSF RPR, (2) elevated CSF protein concentration (>500 mg/L) or WBC (>5 cells/µL) without other known causes, and (3) clinical signs or symptoms consistent with NS in the absence of other known causes of these clinical abnormalities. Syphilis/non-neurosyphilis patients were positive for syphilis serologic tests and underwent lumbar puncture to rule out neurosyphilis. Non-syphilis participants were referred to as persons with negative syphilitic serological tests (RPR and TPPA).
      2.2 Quantification of CSF UCH-L1, GFAP, and NF-L
      CSF samples were gained at the time of diagnosis, and directly transported to the Centre of Clinical Laboratory, where the samples were centrifuged. The supernatants of CSF samples were collected and stored at −80 °C until assayed (Lin et al. 2016). CSF UCH-L1 (R&D Systems, Minneapolis, MN, USA), GFAP (R&D Systems, Minneapolis, MN, USA), and NFL (Novus Biologicals, Littleton, USA) levels were measured using human ELISA kits according to the manufacturer's instructions.
      2.3 Syphilitic serologic tests
      Serological tests for syphilis were performed using RPR (Intec Products Inc., Xiamen, China) and TP-CIA (Boson Biotechnology, Xiamen, China) according to the manufacturer's instructions and our previous studies (Qiu et al. 2015; Lin et al. 2016). RPR had a sensitivity and specificity of 75.81% and 99.67%, respectively, while TP-CIA had a sensitivity and specificity of 99.85% and 99.98%, respectively, for syphilis diagnosis according to our previous study (Tong et al. 2014).
      2.4 Examination of CSF proteins and WBC
      Approximately 2 mL of each CSF sample was collected in a plain sterile tube and analysed within 1 hour to measure the protein content using a Roche-Hitachi Modular P800 analyser (Roche Diagnostics, Basel, Switzerland). The CSF WBC count was determined using an automatic blood cell XE5000 analyser (Sysmex International Reagents Co., Tokyo, Japan) (Tong et al. 2013; Lin et al. 2016).
      2.5 Statistical Analysis
      Continuous variables are presented as medians (interquartile spacing), and categorical variables are expressed as frequencies (percentages). The Mann–Whitney U test was used for continuous variables with skewed distribution, the χ2 test was used for categorical variables, and the odds ratio (OR) was calculated. Spearman rank correlation was used to analyse the correlation among all parameters (CSF UCH-L1, GFAP, NF-L, CSF WBC, and CSF protein). The correlation of the results according to their r values was categorised as extreme (0.91–1.0), strong (0.71–0.9), moderate (0.41–0.7), weak, or poor (0–0.40). Receiver operating characteristic (ROC) analysis was performed to determine the performance of CSF UCH-L1, GFAP, and NF-L in the diagnosis of NS, and the optimal cut-off point was determined to correspond to the maximal Youden's index (sensitivity+specificity−100%). The sensitivity, specificity, and predictive value of the optimal critical values were calculated. The above-mentioned statistical analyses were performed using SPSS 19.0 for Windows (SPSS Inc., Chicago, Illinois, USA). Statistical significance was defined as a two-sided P-value <0.05.
      3. Results
      3.1 Characteristics of the participants
      The clinical characteristics of the study participants are summarised in Table 1. In this study, the NS, syphilis/non-NS, and non-syphilis control groups exhibited no significant differences in age (P=0.065) or sex (P=0.339). The serum RPR, serum TP-CIA, and CSF TP-CIA titres in the NS group were significantly higher than those in the syphilis/non-NS group (P<0.001). Compared to the syphilis/non-NS and non-syphilis groups, patients with NS had higher levels of CSF WBC (P<0.001) and increased CSF protein (P<0.001)
      Table 1Clinical characteristics of the participants
      NS (n=47)Syphilis/non-NS (n=90)Non-Syphilis (n=439)PPaPbPc
      Age, median (IQR)57(48-65)50(35-64)50(36-62)0.065
      Gender ratio n(M/F)32/1550/40254/1850.339
      Serum RPR titer, median (IQR)1:16(1:4-1:64)1:2(1:1-1:8)0(0-0)< 0.001
      Serum TP-CIA titer, median (IQR)120.51(71.53-144.85)32.86(10.81-67.15)0(0-0)< 0.001
      CSF RPR titer, median (IQR)1:4(1:4-1:8)0(0-0)NA
      CSF TP-CIA titer, median (IQR)92.34(37.16-155.40)4.28(1.51-12.38)NA< 0.001
      CSF WBC, median (IQR), /μL25(4-62)2(1-7)5(2-47)< 0.001< 0.0010.0310.051
      CSF protein, median (IQR), mg/L617.10(451.10-936.30)407.00(316.55-576.52)478.20(336.75-875.60)< 0.001< 0.0010.0380.052
      Data expressed as median (IQR) as appropriate. The Wilcoxon test was used to compare age, serum RPR, TP-CIA titre, CSF RPR, TP-CIA titre, CSF WBC count, and protein levels in different groups. The chi-square test was used to compare the size of the sex ratio in different groups.
      Abbreviations: NS, neurosyphilis; M, male; F, female; RPR, rapid plasma regain; TP, Treponema pallidum; CSF, cerebrospinal fluid; WBC, white blood cell; CIA, chemiluminescence immunoassay; IQR, interquartile range; NA, not available; P, comparison among the NS, syphilis/non-NS, and non-syphilis groups; Pa, comparison between the NS and syphilis/non-NS groups; Pb, comparison between the NS and non-syphilis groups; Pc, comparison between the syphilis/non-NS and non-syphilis groups.
      3.2 CSF UCH-L1, GFAP and NF-L levels among the groups
      The level of CSF UCH-L1 was 951.92 (715.10–1,263.51) pg/mL in the NS group, which was significantly higher than that in the syphilis/non-NS group [370.30 (231.23–543.10) pg/mL] (P<0.001) and non-syphilis group [155.47 (56.29–679.33) pg/mL] (P<0.001). CSF UCH-L1 levels were higher in the syphilis/non-NS group than in the non-syphilis group (P=0.022; Figure 1A). Similarly, CSF GFAP [692.19 (549.53–950.61) pg/mL] and CSF NF-L levels [87.15 (57.21–127.60) pg/mL] in the NS group were significantly higher than levels in the syphilis/non-NS group [409.62 (381.13–503.06) pg/mL, P<0.001; 30.41 (20.11–40.43) pg/mL, P<0.001] and non-syphilis group [407.46 (405.75–453.98) pg/mL, P<0.001; 20.99 (15.61–46.27) pg/mL, P<0.001]. However, there was no statistical difference between the syphilis/non-NS and the non-syphilis groups (P=0.7696, P=0.0777) (Figure 1B, C).
      Figure 1
      Figure 1CSF levels of UCH-L1, GFAP, and NF-L in NS, syphilis/non-NS, and non-syphilis patients.
      (A) CSF UCH-L1, (B) CSF GFAP, and (C) CSF NF-L.
      The Wilcoxon test was used to compare the levels of CSF UCH-L1, GFAP, and NF-L in the different groups.
      *: P<0.05, ***: P<0.001, ns: no significance.
      Abbreviations: NS, neurosyphilis; CSF, cerebrospinal fluid; UCH-L1, ubiquitin C-terminal hydrolase L1; GFAP, glial fibrillary acidic protein; NF-L, neurofilament light protein.
      3.3 Diagnostic value of CSF UCH-L1, GFAP, and NF-L levels in NS patients
      To evaluate the use of CSF UCH-L1 levels as a screening test for NS, its screening accuracy was determined using ROC curve analysis. The area under the curve (AUC) for CSF UCH-L1 was 0.9177 (95% CI 0.8691–0.9663; Figure 2A). Comparably, the AUC of CSF GFAP and NF-L were 0.8512 and 0.8848, respectively (95% CI 0.7782–0.9242; 95% CI 0.8222–0.9473; Figure 2B, C). According to the maximal Youden's index, the optimal cut-offs for CSF UCH-L1, GFAP, and NF-L to distinguish NS from syphilis/non-NS were 652.25 pg/mL, 548.89 pg/mL, and 48.38 pg/mL, respectively. In addition, the chi-square test was used to analyse the correlation of CSF UCH-L1, GFAP, and NF-L with NS, and the results showed that syphilis patients with CSF UCH-L1≥652.25 pg/mL were 67.76 times more likely to have NS than those with CSF UCH-L1<652.25 pg/mL (odds ratio [OR], 67.76; 95% CI 22.25–206.31) (P<0.001). Likewise, syphilis patients with CSF GFAP≥548.89 pg/mL and CSF NF-L≥48.38 pg/mL were 19.39 times and 49.97 times more likely to have NS than those with low levels, respectively (OR, 19.39; 95% CI, 7.92–47.45; and OR, 49.97; 95% CI, 17.46–143.01) (P<0.001) (Table 2).
      Figure 2
      Figure 2ROC curves analysis of CSF UCH-L1, GFAP, and NF-L for NS diagnosis.
      (A) CSF UCH-L1, (B) CSF GFAP, and (C) CSF NF-L.
      Abbreviations: ROC, receiver operator characteristics; NS, neurosyphilis; CSF, cerebrospinal fluid; UCH-L1, ubiquitin C-terminal hydrolase L1; GFAP, glial fibrillary acidic protein; NF-L, neurofilament light protein.
      Table 2Diagnostic value of CSF UCH-L1, GFAP and NF-L levels in NS patients.
      Optimal cut-offNS, n (%)Syphilis/non-NS, n (%)OR (95% CI)P
      CSF UCH-L1 < 652.25 pg/mL7 (14.89)83 (92.22)1.0
      CSF UCH-L1 ≥ 652.25 pg/mL40 (85.11)7 (7.78)67.76 (22.25-206.31)< 0.001
      CSF GFAP < 548.89 pg/mL11 (23.40)77 (85.56)1.0
      CSF GFAP ≥ 548.89 pg/mL36 (76.60)13 (14.44)19.39 (7.92-47.45)< 0.001
      CSF NF-L< 48.38 pg/mL8 (17.02)82 (91.11)1.0
      CSF NF-L ≥ 48.38 pg/mL39 (82.98)8 (8.89)49.97 (17.46-143.01)< 0.001
      The chi-square test was used to compare the diagnostic values of CSF UCH-L1, GFAP, and NF-L levels for NS diagnosis in the different groups.
      Abbreviations: NS, neurosyphilis; CSF, cerebrospinal fluid; OR, odds ratio; CI, confidence interval; UCH-L1, ubiquitin C-terminal hydrolase L1; GFAP, glial fibrillary acidic protein; NF-L, neurofilament light protein.
      3.4 Evaluation of the diagnostic accuracy of CSF UCH-L1, GFAP, and NF-L compared with the clinical diagnosis in NS
      Based on the aforementioned ROC curve and optimal cut-off values established above, CSF UCH-L1 demonstrated a sensitivity of 85.11%, a specificity of 92.22%, and a negative predictive value (NPV) of 92.22% for NS diagnosis by using diagnostic criteria of NS as the reference standard (Table 3). Similarly, the sensitivity and specificity of CSF GFAP for NS diagnosis were 76.60% and 85.56%, respectively. Moreover, the sensitivity and specificity of CSF NF-L for NS diagnosis were 82.98% and 91.11%, respectively.
      Table 3The diagnostic accuracy of CSF UCH-L1, GFAP and NF-L compared with clinical diagnosis in NS.
      CSF UCH-L1

      ≥ 652.25 pg/mL
      CSF GFAP

      ≥ 548.89 pg/mL
      CSF NF-L

      ≥ 48.38 pg/mL
      Sen (%)Spe (%)PPV (%)NPV

      (%)
      PLRNLRκ valueAgreement

      (%)
      Single indicator positive
      +//85.1192.2285.1192.2210.940.160.7789.78
      /+/76.6085.5673.4785.505.300.270.6282.48
      //+82.9891.1182.9891.119.330.190.7488.32
      Parallel testing format: one of two or three indicators positive
      +/-+/-/83.6780.0069.4992.314.180.200.6382.48
      +/-/+/-93.6291.1184.6296.4710.530.070.8391.97
      /+/-+/-91.4978.8969.3594.674.330.110.6583.21
      +/-+/-+/-93.6277.7862.5095.894.210.080.6683.21
      Combination testing format: two or three indicators positive
      ++/74.4797.7894.5988.0033.550.260.7689.78
      +/+74.4792.2283.2287.379.570.280.6886.13
      /++68.0998.8994.1285.4461.340.320.7187.59
      +++65.9698.8993.9484.6259.420.340.6986.86
      + means adding the indicator; / means that the indicator is not added. +/- means that the indicator can be added, but has one of two or three positive indicators.
      Abbreviations: CSF, cerebrospinal fluid; UCH-L1, ubiquitin C-terminal hydrolase L1; GFAP, glial fibrillary acidic protein; NF-L, neurofilament light protein; PPV, positive predictive value; NPV, negative predictive value; PLR, positive likelihood ratio; NLR, negative likelihood ratio.
      To further improve the sensitivity of NS diagnosis, CSF UCH-L1, GFAP, and NF-L were combined in a parallel testing format (Table 3), which referred to one of two or three positive indicators. Using this algorithm, sensitivity increased more than that observed with a single indicator, reaching 83.67–93.62%, while specificity decreased, although it was still above 77.78%. The combination of CSF UCH-L1 and CSF NF-L was most optimal, yielding a high sensitivity of 93.62% and specificity of 91.11% in identifying NS. Moreover, the NPV and κ values increased to 96.47% and 0.83, respectively (Table 3).
      To further improve the specificity of NS diagnosis, a series combination testing format was used, which referred to two or three indicators that were all positive. By this algorithm, the specificity increased as high as 92.22–98.89%, but the sensitivity decreased significantly to 65.96–74.47%. The combination of CSF GFAP≥548.89 pg/mL and CSF NF-L≥48.38 pg/mL both had a specificity and PPV of 98.89% and 94.12% for the diagnosis of NS, respectively (Table 3).
      3.5 CSF UCH-L1, GFAP, and NF-L levels in NS patients with CSF RPR positive, negative, and syphilis/non-NS patients
      To evaluate the association of CSF UCH-L1, GFAP, and NF-L levels with CSF RPR, the NS group was further divided into CSF RPR-positive and-negative subgroups, and the syphilis/non-NS group (all CSF RPR-negative) was used as the control. The results showed that the levels of CSF UCH-L1, CSF GFAP, and NF-L were all higher in the NS group, regardless of positive or negative RPR, than in the syphilis/non-NS group. There was no difference in CSF UCH-L1 levels between CSF RPR positive NS patients and CSF RPR negative NS patients (median: 960.00 pg/mL versus 922.02 pg/mL, P>0.9999; Figure 3A); however, CSF UCH-L1 levels in CSF RPR positive and CSF RPR negative NS patients were higher than those in syphilis/non-NS patients (median: 960.00 pg/mL versus 370.30 pg/mL, P<0.001; 922.02 pg/mL versus 370.30 pg/mL, P<0.001; Figure 3A). Similarly, CSF GFAP and NF-L levels in CSF RPR positive NS patients were not significantly different from those in CSF RPR negative NS patients (P>0.999, P>0.9999; Figure 3B, C), and both were higher than those in syphilis/non-NS patients (median: 682.82 pg/mL versus 409.62 pg/mL, P<0.001; 802.93 pg/mL versus 409.62 pg/mL, P<0.001; median: 81.04 pg/mL versus 30.42 pg/mL, P<0.001; 92.17 pg/mL versus 30.42 pg/mL, P<0.001; Figure 3B, C). These results indicate that CSF UCH-L1, GFAP, and NF-L may be indicators for NS diagnosis independent of CSF RPR.
      Figure 3
      Figure 3CSF UCH-L1, CSF GFAP, and CSF NF-L levels in NS patients with CSF RPR positive, negative, and syphilis/non-NS patients.
      (A) CSF UCH-L1, (B) CSF GFAP, and (C) CSF NF-L.
      The Wilcoxon test was used to compare the levels of CSF UCH-L1, GFAP, and NF-L in the different groups.
      ***: P<0.001, ns: no significance.
      Abbreviations: NS, neurosyphilis; CSF, cerebrospinal fluid; UCH-L1, ubiquitin C-terminal hydrolase L1; GFAP, glial fibrillary acidic protein; NF-L, neurofilament light protein; RPR, rapid plasma reagin.
      3.6 Correlations of CSF UCH-L1, GFAP, and NF-L levels with CSF WBC
      Spearman rank correlation was used to analyse the correlation between CSF UCH-L1, GFAP, and NF-L levels and CSF WBC count. The results showed that CSF UCH-L1, GFAP, and NF-L levels were moderately or weakly correlated with CSF WBC counts (rs=0.3991; rs=0.4359; rs=0.3845; P<0.001, respectively; Figures 4A, B, and C) in all participants. Furthermore, the NS group was analysed independently and there was no significant correlation with CSF WBC among CSF UCH-L1, CSF GFAP, and CSF NF-L in the NS group (rs=0.1083, P=0.4688; rs=0.1593, P=0.2847; rs=-0.0889, P=0.5523; Figure 4D, E and F).
      Figure 4
      Figure 4Correlations of CSF UCH-L1, GFAP, and NF-L levels with CSF WBC.
      (A) The correlation between CSF UCH-L1 and CSF WBC in all participants. (B) The correlation between CSF GFAP and CSF WBC in all participants. (C) The correlation between CSF NF-L and CSF WBC in all participants. (D) The correlation between CSF UCH-L1 and CSF WBC in NS patients. (E) The correlation between CSF GFAP and CSF WBC in NS patients. (F) The correlation between CSF NF-L and CSF WBC in NS patients.
      Spearman rank correlation was used to analyse the correlation of CSF UCH-L1, GFAP, and NF-L levels with the CSF WBC count.
      Abbreviations: NS, neurosyphilis; CSF, cerebrospinal fluid; UCH-L1, ubiquitin C-terminal hydrolase L1; GFAP, glial fibrillary acidic protein; NF-L, neurofilament light protein; WBC, white blood cell.
      3.7 Correlations of CSF UCH-L1, GFAP, and NF-L levels with CSF proteins concentration
      Similarly, Spearman's rank correlation was used to analyse the correlation between CSF UCH-L1, GFAP, NF-L levels, and CSF protein concentration. The results showed that CSF UCH-L1, GFAP, and NF-L levels were moderately or weakly correlated with CSF protein levels in all participants (rs=0.3776; rs=0.4196; rs=0.3976; P<0.001, respectively; Figure 5A, B, and C). Moreover, CSF UCH-L1, CSF GFAP, and CSF NF-L levels were moderately or weakly correlated with CSF protein levels in the NS group (rs=0.5014, P<0.001; rs= 0.3113, P= 0.0332; rs=0.2481, P=0.0426, respectively; Figure 5D, E, and F).
      Figure 5
      Figure 5Correlations of CSF UCH-L1, GFAP, and NF-L levels with CSF protein concentration.
      (A) The correlation between CSF UCH-L1 and CSF proteins in all participants. (B) The correlation between CSF GFAP and CSF proteins in all participants. (C) The correlation between CSF NF-L and CSF proteins in all participants. (D) The correlation between CSF UCH-L1 and CSF proteins in NS patients. (E) The correlation between CSF GFAP and CSF proteins in NS patients. (F) The correlation between CSF NF-L and CSF proteins in NS patients.
      Spearman rank correlation was used to analyse the correlation between CSF UCH-L1, GFAP, and NF-L levels and CSF protein concentration.
      Abbreviations: NS, neurosyphilis; CSF, cerebrospinal fluid; UCH-L1, ubiquitin C-terminal hydrolase L1; GFAP, glial fibrillary acidic protein; NF-L, neurofilament light protein.
      4. Discussion
      NS is one of the most feared syndromes of syphilis owing to its serious, irreversible sequelae (Liu et al. 2019). To date, the diagnosis of NS has relied on a combination of neurological manifestations in patients with reactive serological anti-T. pallidum test and CSF abnormalities (Liu et al. 2012; Zeng et al. 2013; Lin et al. 2016). Usually, NS can be diagnosed with reactive CSF VDRL in conjunction with neurological signs or symptoms, but the test is insensitive. For patients with suspected NS but negative CSF VDRL, there is a lack of specific indicators to identify NS. Thus, the diagnosis of NS remains challenging and new indicators are urgently needed. In this study, we focused on markers associated with neuronal damage and evaluated the possibility of CSF UCH-L1, GFAP, and NF-L as novel markers for the diagnosis of NS. We found that the CSF UCH-L1, GFAP, and NF-L levels in NS patients were significantly higher than those in syphilis/non-NS patients and non-syphilis participants. Interestingly, syphilis patients with CSF UCH-L1≥652.25 pg/mL, CSF GFAP≥548.89 pg/mL, and CSF NF-L≥48.38 pg/mL were 67.76, 19.39, and 49.97 times more likely to have NS than those with low levels, respectively. Additionally, CSF UCH-L1, GFAP, and NF-L showed a sensitivity of 85.11%, 76.60%, and 82.98%, with a specificity of 92.22%, 85.56%, and 91.11%, respectively, for the diagnosis of NS using diagnostic criteria of NS as the reference standard.
      NS may manifest early on, within the first few weeks to months of the appearance of meningitis or, more seriously, as parenchymal inflammation of the brain and spinal cord. Parenchymal inflammation is more common than meningitis and leads to irreversible CNS damage, which may be attributed to neuronal damage and loss (Zhang et al. 2013a). Astroglial and neuronal proteins have been detected in the CSF of patients with NS, indicating possible glial and neuronal damage in the CNS parenchyma (Xu et al. 2020; Stroffolini et al. 2021). In this present study, we similarly found that the neuronal damage markers, CSF UCH-L1, GFAP, and NF-L, were higher in NS patients than in syphilis/non-NS patients and non-syphilis participants, confirming that NS patients may have varying degrees of neuronal damage. Furthermore, we learnt that CSF UCH-L1, GFAP, and NF-L had superior screening performance, with AUCs of 0.9177, 0.8512, and 0.8848, respectively, in the identification of NS. Of course, the sole use of UCH-L1, GFAP, or NF-L may not adequately predict NS. Therefore, we employed a parallel testing format and found that the combination of CSF UCH-L1≥652.25 pg/mL and CSF NF-L≥48.38 pg/mL had a high sensitivity of 93.62% in NS screening, achieving a low-probability of a missed diagnosis of NS. Parallel testing also provided an NPV of 96.47%, that is, 96.47% of patients excluded from the parallel testing format were correctly identified as non-NS. Moreover, we also implemented a serial testing format and found that the serial combination of CSF GFAP≥548.89 pg/mL and CSF NF-L≥48.38 pg/mL had a specificity of 98.89%, and PPV of 94.12%, i.e., 94.12% of patients identified from the serial testing format will be correctly identified as NS, thus avoiding misdiagnosis.
      The CSF RPR test plays an important role in the diagnosis of NS, along with the VDRL test. The specificity of CSF RPR is high (74-100%), whereas its sensitivity is approximately 49–87% (Tuddenham, Katz, and Ghanem 2020). Consequently, a negative CSF RPR does not exclude NS. When CSF-RPR is negative, but clinical signs of neurosyphilis, reactive serological test results, and abnormal inflammatory parameters (CSF cell count and/or protein) are present, NS should be considered. However, CSF cell counts and protein levels may falsely increase due to a traumatic tap. In this instance, additional evaluation of the CSF may be warranted. In this study, the levels of CSF UCH-L1, CSF GFAP, and NF-L were all higher in the NS group, regardless of a positive or negative RPR result, than in the syphilis/non-NS group; moreover, they were weakly or moderately correlated with CSF WBC and protein levels. The results indicated that CSF UCH-L1, GFAP, and NF-L, reflecting neuronal damage, may be indicators for NS diagnosis independent of CSF RPR or inflammatory parameters and could be novel indicators for NS diagnosis. Notably, the UCH-L1, GFAP, or NF-L were measured using human ELISA kits. ELISA can be very sensitive and specific analytical techniques. It is also cost-affordable and fairly easy to perform. Additionally, there is a lower reporting error since the reading is taken by the ELISA reader, compared to the manual test, which may have subjective variation. Moreover, the pre-analytical errors such as sample interchange and post-analytic transcriptional errors can be eliminated because the automated system can be linked to the hospital information system (Zhdanov et al. 2018).
      Neurosyphilis is demonstrated by the presence of an increased number of human B cells and an abnormal B cell response in the central nervous system, and Chemokine ligand 13 (CXCL13) is regarded as B cell attracting chemokine 1 (Gudowska-Sawczuk and Mroczko 2020; Reiber and Peter 2001). CXCL13 has been examined and strongly indicated as a possible diagnostic tool in the diagnosis of neurosyphilis (Gudowska-Sawczuk and Mroczko 2020; Marra 2021; Stroffolini et al. 2021). A previous study in our laboratory also showed that CXCL13 has superior screening performance in the identification of NS (Zeng et al. 2016). Further comparison of the diagnostic performance of CXCL13 with UCH-L1, GFAP, or NF-L is required in future studies.
      The limitations of this study should be acknowledged. First, the sample size was relatively small. Second, the NS group was not further classified according to clinical symptoms or signs to analyse the levels of neuronal damage in various types of NS. Third, this study was designed as a cross-sectional survey and no follow-up data for CSF UCH-L1, GFAP, and NF-L concentrations were gained to evaluate the effect of therapy for NS patients. Fourth, guidelines and criteria are different for NS/HIV-positive and NS/HIV-negative patients (Workowski et al. 2021). Among enrolled participants, only two were coinfected with HIV; therefore, we excluded the two HIV-positive patients from this research to reflect the population accurately, and evaluated the diagnostic value of CSF UCH-L1, GFAP, and NF-L in NS among the HIV-negative population. The interesting issue regarding HIV-positive patients should be investigated in future studies. The fifth, previous study found that GFAP expression is increased in neurodegenerative diseases(Oeckl et al. 2019; Olsson et al. 2016; Heller et al. 2020). However, NS is not neurodegenerative per se, our results showed that GFAP was significantly higher in NS patients, which is very worthy of further investigation. Finally, NS is present as a neuroinflammatory disease, and the diagnostic value of neuroinflammatory relative markers such as CSAR, IgG Ratio/reibergram, beta amyloid, neopterin, total and phosphorylated tau proteins, 14.3.3 protein, and s100beta protein (Reiber and Peter 2001, Stroffolini et al. 2021), require further study.
      Our study demonstrated that CSF UCH-L1, GFAP, and NF-L can be used as novel markers for the diagnosis of NS, independent of CSF RPR, WBC, and proteins. Parallel testing could improve sensitivity, and serial testing could improve specificity to avoid missed diagnosis.
      Conflict of interest
      All authors declare that they have no conflicts of interest.
      Funding
      This work was supported by the National Natural Science Foundation of China [grant numbers 82172331, 81973104, 81972028, 81772260, 81471967, 82003512 and 82072321], the Key Projects for Province Science and Technology Program of Fujian Province, China [grant number 2020D017, 2018D0014], the Natural Science Foundation of Fujian Province, China [grant number 2021J02055], The Basic and Clinical Research Training Project of Dermatology Hospital, Southern Medical University [grant number C2019001] and the Guangdong Medical Research Fund, Programme for Appropriate Technology in Health Project [grant number 202107031024288992].
      Ethical approval
      This study was approved by the Ethics Committee of Zhongshan Hospital of Xiamen University, School of Medicine, Xiamen University in the People's Republic of China, and was in compliance with national legislation and the Declaration of Helsinki guidelines. Written informed consent to participate in this study was provided by the participants or their legal guardian/next of kin.
      Acknowledgments
      The authors would like to thank fellows at Zhongshan Hospital of Xiamen University and 576 participants for their support.
      Author Contributions
      Tian-Ci Yang and Wu-Jian Ke contributed to the study conception and design. Materials preparation, data collection and analysis were performed by Rui Chen, Li-Rong Lin, Yao Xiao. The first draft of the manuscript was written by Rui Chen, Li-Rong Lin and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
      Declaration of interests
      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
      References
      Bazarian, J. J., P. Biberthaler, R. D. Welch, L. M. Lewis, P. Barzo, V. Bogner-Flatz, et al. 2018. 'Serum GFAP and UCH-L1 for prediction of absence of intracranial injuries on head CT (ALERT-TBI): a multicentre observational study', Lancet Neurol, 17: 782-89.·http://doi:10.1016/S1474-4422(18)30231-X
      Bilguvar, K., N. K. Tyagi, C. Ozkara, B. Tuysuz, M. Bakircioglu, M. Choi, et al. 2013. 'Recessive loss of function of the neuronal ubiquitin hydrolase UCHL1 leads to early-onset progressive neurodegeneration', Proc Natl Acad Sci U S A, 110: 3489-94.·http://doi:10.1073/pnas.1222732110
      Disanto, G., C. Barro, P. Benkert, Y. Naegelin, S. Schadelin, A. Giardiello, et al. 2017. 'Serum Neurofilament light: A biomarker of neuronal damage in multiple sclerosis', Ann Neurol, 81: 857-70.·http://doi:10.1002/ana.24954
      Ghanem, K. G., S. Ram, and P. A. Rice. 2020. 'The Modern Epidemic of Syphilis', N Engl J Med, 382: 845-54.·http://doi:10.1056/NEJMra1901593
      Graham, S. H., and H. Liu. 2017. 'Life and death in the trash heap: The ubiquitin proteasome pathway and UCHL1 in brain aging, neurodegenerative disease and cerebral Ischemia', Ageing Res Rev, 34: 30-38.·http://doi:10.1016/j.arr.2016.09.011
      Gudowska-Sawczuk, M., and B. Mroczko. 2020. 'Chemokine Ligand 13 (CXCL13) in Neuroborreliosis and Neurosyphilis as Selected Spirochetal Neurological Diseases: A Review of Its Diagnostic Significance', Int J Mol Sci, 21.·http://doi:10.3390/ijms21082927
      Heller, C., M. S. Foiani, K. Moore, R. Convery, M. Bocchetta, M. Neason, et al. 2020. 'Plasma glial fibrillary acidic protein is raised in progranulin-associated frontotemporal dementia', J Neurol Neurosurg Psychiatry, 91: 263-70.·http://doi:10.1136/jnnp-2019-321954
      Janier, M., M. Unemo, N. Dupin, G. S. Tiplica, M. Potocnik, and R. Patel. 2020. '2020 European guideline on the management of syphilis', J Eur Acad Dermatol Venereol.·http://doi:10.1111/jdv.16946
      Khalil, M., C. E. Teunissen, M. Otto, F. Piehl, M. P. Sormani, T. Gattringer, et al. 2018. 'Neurofilaments as biomarkers in neurological disorders', Nat Rev Neurol, 14: 577-89.·http://doi:10.1038/s41582-018-0058-z
      Kingston, M., P. French, S. Higgins, O. McQuillan, A. Sukthankar, C. Stott, et al. 2016. 'UK national guidelines on the management of syphilis 2015′, Int J STD AIDS, 27: 421-46.·http://doi:10.1177/0956462415624059
      Levett, P. N., K. Fonseca, R. S. Tsang, K. Kadkhoda, B. Serhir, S. M. Radons, et al. 2015. 'Canadian Public Health Laboratory Network laboratory guidelines for the use of serological tests (excluding point-of-care tests) for the diagnosis of syphilis in Canada', Can J Infect Dis Med Microbiol, 26 Suppl A: 6A-12A.·http://doi:10.1155/2015/983425
      Li, W., H. Chang, W. Wu, D. Xu, M. Jiang, J. Gao, et al. 2020. 'Increased CSF Soluble TREM2 Concentration in Patients With Neurosyphilis', Front Neurol, 11: 62.·http://doi:10.3389/fneur.2020.00062
      Lin, J., Y. Zheng, Y. Liu, Y. Lin, Q. Wang, X. H. Lin, et al. 2021. 'Higher Concentration of Plasma Glial Fibrillary Acidic Protein in Wilson Disease Patients with Neurological Manifestations', Mov Disord, 36: 1446-50.·http://doi:10.1002/mds.28509
      Lin, L. R., D. H. Lin, M. L. Tong, L. L. Liu, J. Y. Fan, X. Z. Zhu, et al. 2016. 'Macrophage migration inhibitory factor as a novel cerebrospinal fluid marker for neurosyphilis among HIV-negative patients', Clin Chim Acta, 463: 103-08.·http://doi:10.1016/j.cca.2016.10.018
      Liu, L. L., W. N. Liu, X. Y. Jiang, Ren Jun, M. H. Chen, Z. J. Liu, et al. 2019. 'Changes of T lymphocyte subsets in patients with HIV-negative symptomatic neurosyphilis', Microb Pathog, 130: 213-18.·http://doi:10.1016/j.micpath.2019.03.008
      Liu, L. L., W. H. Zheng, M. L. Tong, G. L. Liu, H. L. Zhang, Z. G. Fu, et al. 2012. 'Ischemic stroke as a primary symptom of neurosyphilis among HIV-negative emergency patients', J Neurol Sci, 317: 35-9.·http://doi:10.1016/j.jns.2012.03.003
      Marra, C. M. 2021. 'Alternatives to the Cerebrospinal Fluid Venereal Disease Research Laboratory Test for Neurosyphilis Diagnosis', Sex Transm Dis, 48: S54-S57.·http://doi:10.1097/OLQ.0000000000001450
      Newcombe, V. F. J., N. J. Ashton, J. P. Posti, B. Glocker, A. Manktelow, D. A. Chatfield, et al. 2022. 'Post-acute blood biomarkers and disease progression in traumatic brain injury', Brain.·http://doi:10.1093/brain/awac126
      Oeckl, P., P. Weydt, P. Steinacker, S. Anderl-Straub, F. Nordin, A. E. Volk, et al. 2019. 'Different neuroinflammatory profile in amyotrophic lateral sclerosis and frontotemporal dementia is linked to the clinical phase', J Neurol Neurosurg Psychiatry, 90: 4-10.·http://doi:10.1136/jnnp-2018-318868
      Olsson, B., R. Lautner, U. Andreasson, A. Ohrfelt, E. Portelius, M. Bjerke, et al. 2016. 'CSF and blood biomarkers for the diagnosis of Alzheimer's disease: a systematic review and meta-analysis', Lancet Neurol, 15: 673-84.·http://doi:10.1016/S1474-4422(16)00070-3
      Park, I. U., A. Tran, L. Pereira, and Y. Fakile. 2020. 'Sensitivity and Specificity of Treponemal-specific Tests for the Diagnosis of Syphilis', Clin Infect Dis, 71: S13-S20.·http://doi:10.1093/cid/ciaa349
      Peeling, R. W., D. Mabey, M. L. Kamb, X. S. Chen, J. D. Radolf, and A. S. Benzaken. 2017. 'Syphilis', Nat Rev Dis Primers, 3: 17073.·http://doi:10.1038/nrdp.2017.73
      Qiu, X. H., Y. F. Zhang, Y. Y. Chen, Q. Zhang, F. Y. Chen, L. Liu, et al. 2015. 'Evaluation of the Boson Chemiluminescence Immunoassay as a First-Line Screening Test in the ECDC Algorithm for Syphilis Serodiagnosis in a Population with a High Prevalence of Syphilis', Journal of clinical microbiology, 53: 1371-74.·http://doi:10.1128/Jcm.00069-15
      Reiber, H., and J. B. Peter. 2001. 'Cerebrospinal fluid analysis: disease-related data patterns and evaluation programs', J Neurol Sci, 184: 101-22.·http://doi:10.1016/s0022-510x(00)00501-3
      Rohrer, J. D., I. O. Woollacott, K. M. Dick, E. Brotherhood, E. Gordon, A. Fellows, et al. 2016. 'Serum neurofilament light chain protein is a measure of disease intensity in frontotemporal dementia', Neurology, 87: 1329-36.·http://doi:10.1212/WNL.0000000000003154
      Setsuie, R., and K. Wada. 2007. 'The functions of UCH-L1 and its relation to neurodegenerative diseases', Neurochem Int, 51: 105-11.·http://doi:10.1016/j.neuint.2007.05.007
      Shahim, P., A. Politis, A. van der Merwe, B. Moore, V. Ekanayake, S. M. Lippa, et al. 2020. 'Time course and diagnostic utility of NfL, tau, GFAP, and UCH-L1 in subacute and chronic TBI', Neurology, 95: e623-e36.·http://doi:10.1212/WNL.0000000000009985
      Siller, N., J. Kuhle, M. Muthuraman, C. Barro, T. Uphaus, S. Groppa, et al. 2019. 'Serum neurofilament light chain is a biomarker of acute and chronic neuronal damage in early multiple sclerosis', Mult Scler, 25: 678-86.·http://doi:10.1177/1352458518765666
      Stroffolini, G., G. Guastamacchia, S. Audagnotto, C. Atzori, M. Trunfio, M. Nigra, et al. 2021. 'Low cerebrospinal fluid Amyloid-betaeta 1-42 in patients with tuberculous meningitis', Bmc Neurology, 21: 449.·http://doi:10.1186/s12883-021-02468-2
      Tong, M. L., L. R. Lin, L. L. Liu, H. L. Zhang, S. J. Huang, Y. Y. Chen, et al. 2014. 'Analysis of 3 algorithms for syphilis serodiagnosis and implications for clinical management', Clin Infect Dis, 58: 1116-24.·http://doi:10.1093/cid/ciu087
      Tong, M. L., L. R. Lin, H. L. Zhang, S. J. Huang, G. L. Liu, W. H. Zheng, et al. 2013. 'Spectrum and characterization of movement disorders secondary to neurosyphilis', Parkinsonism Relat Disord, 19: 441-5.·http://doi:10.1016/j.parkreldis.2013.01.003
      Tuddenham, S., S. S. Katz, and K. G. Ghanem. 2020. 'Syphilis Laboratory Guidelines: Performance Characteristics of Nontreponemal Antibody Tests', Clin Infect Dis, 71: S21-S42.·http://doi:10.1093/cid/ciaa306
      Workowski, K. A., L. H. Bachmann, P. A. Chan, C. M. Johnston, C. A. Muzny, I. Park, et al. 2021. 'Sexually Transmitted Infections Treatment Guidelines, 2021′, Mmwr Recommendations and Reports, 70: 1-187.·http://doi:10.15585/mmwr.rr7004a1
      Xu, D. M., S. N. Cai, R. Li, Y. Wu, S. A. Liu, and W. H. Lun. 2020. 'Elevation of Cerebrospinal Fluid Light and Heavy Neurofilament Levels in Symptomatic Neurosyphilis', Sex Transm Dis, 47: 634-38.·http://doi:10.1097/OLQ.0000000000001236
      Zeng, Y. L., Y. Q. Lin, N. N. Zhang, C. N. Zou, H. L. Zhang, F. Peng, et al. 2016. 'CXCL13 chemokine as a promising biomarker to diagnose neurosyphilis in HIV-negative patients', Springerplus, 5: 743.·http://doi:10.1186/s40064-016-2462-4
      Zeng, Y. L., W. J. Wang, H. L. Zhang, F. Y. Chen, S. J. Huang, G. L. Liu, et al. 2013. 'Neuropsychiatric disorders secondary to neurosyphilis in elderly people: one theme not to be ignored', Int Psychogeriatr, 25: 1513-20.·http://doi:10.1017/S1041610213000896
      Zhang, H. L., L. R. Lin, G. L. Liu, Y. L. Zeng, J. Y. Wu, W. H. Zheng, et al. 2013a. 'Clinical spectrum of neurosyphilis among HIV-negative patients in the modern era', Dermatology, 226: 148-56.·http://doi:10.1159/000347109
      Zhdanov, A., J. Keefe, L. Franco-Waite, K. R. Konnaiyan, and A. Pyayt. 2018. 'Mobile phone based ELISA (MELISA)', Biosensors & Bioelectronics, 103: 138-42.·http://doi:10.1016/j.bios.2017.12.033