Characteristics of Brain Structure in HIV Male Patients with Syphilis Co-Infection: A Cross-Sectional Study.

Background To investigate the effect of syphilis infection on the microstructure of white matter (WM) in HIV-infected male patients through comparing the differences of WM between HIV-infected male patients with and without syphilitic infection using diffusion tensor imaging (DTI). Methods 27 HIV-infected male patients with current syphilis or a history of syphilis (HIV+/syphilis+) and 29 HIV-infected male patients without syphilis co-infection history (HIV+/syphilis-) were enrolled. All patients received DTI and comprehensive neuropsychological assessment. Clinical data were compared between the two groups with T-test, Mann-Whitney U Test and Chi-square Test. Tract-based spatial statistics (TBSS) was adopted to analyze the DTI metrics. Multiple linear regression analysis was conducted to investigate the relationships between DTI metrics and clinical variables and cognitive performance.


Background
When the human immunode ciency virus (HIV) enters the human body, it can invade and destroy the immune system, and can also invade the central nervous system (CNS) at an early stage of infection.
One study showed that viral RNA could be detected in the cerebrospinal uid(CSF) of infected person within as early as eight days of HIV infection [1] . After invaded, the astrocytes, microglia, oligodendrocytes and other cells can be destroyed by the virus [2] , causing some changes in the microstructure of white matter (WM). Sustained damage caused by HIV can lead to multiple cognitive impairments in patients known as HIV-related neurocognitive disorders (HAND). HAND is classi ed into three levels according to severity: asymptomatic neurocognitive impairment (ANI), mild neurocognitive disorder (MND) and HIVassociated dementia(HAD) [3][4] . With the widespread use of antiretroviral therapy (ART), the ratio of MND and HAD decreased while the ratio of ANI increased [5] . Although the rate of the neurocognitive decline is slow, if the treatment is neglected or not properly controlled, ANI eventually deteriorates into MND and even into HAD [6] . Therefore, it is of signi cance to paying attention to the neurocognitive function of HIVinfected patients.
In recent years, the incidence of syphilis in some European regions has increased [7][8] , which mainly occurs in men who have sex with men (MSM) [9] . Syphilis, caused by the spirochaete bacterium Treponema pallidum has the same group of patients with HIV, the incidence of syphilis in HIV-infected patients is 77 times greater than in the general population [10][11] . There may be a cooperative relationship between syphilis and HIV: syphilis had a negative impact on the CD4+T cell counts in HIV-positive patients, regardless of treatment [12] . HIV-infected patients with CD4+T counts <350 μl/L were more likely to acquire neurosyphilis [13] , and when immune system is seriously impaired, neurosyphilis tends to progress [14] . Syphilis co-infection may cause neurocognitive impairment. A study showed that patients with prior syphilis had a greater number of impaired NP test domains and more impaired in the NP learning domain compared with HIV-infected patients without a history of syphilis [15] ; another study showed HIV infected subjects with a history of syphilis had a poor performance in the cognitive domains of memory and learning [16] . However, the mechanism of cognitive impairment caused by syphilis and HIV co-infection remains unclear. Some scholars believe that increased HIV viral load in CSF of syphilis and HIV co-infection patients may be related to impaired neurocognitive function [14,17] .
The damage of microstructure in WM can re ect the change of cognitive function to some extent.
Diffusion tensor imaging (DTI) can re ect the diffusion characteristics of water molecules in WM, and it is extremely sensitive to changes in the microstructure of WM. There are four parameters for DTI: fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD) and radial diffusivity (RD). FA is related to the integrity of the WM, decreased FA indicates that the integrity of WM is impaired [18][19][20] . MD can re ect the diffusion rate of water molecules. Decreased FA and increased MD mean the damage of neuronal ber bundles [21] . AD re ects the condition of axonal damage, decreased AD indicates acute axonal injury while increased AD suggests chronic axonal injury [22] . Increased RD indicates the absence of myelin sheath and re ects demyelination [23] . Several studies have investigated the relationship between the damage to WM and cognitive impairment in HIV-infected patients. Chang [24] found that for antiretroviral-stable HIV patients, the decline in the global de cit score was related to the increased MD in the genu of callosum and decreased FA in the parietal and frontal WM and putamen. Cysique [25] studied the HIV-infected patients who were clinically stable with successful viral control and showed that compared to both control group and ANI patients, patients with MND or HAD had lower FA values in cingulum and fornix, higher MD values in the fornix and external capsule. Li [26] found increased AD and MD in periventricular WM in untreated ANI patients through comparing them with normal controls, the result suggested that the cause of microstructure damage in untreated HIV patients with ANI was axonal injury rather than demyelination.
However, There was no study on the characteristics of its microstructure changes in WM of patients with syphilis and HIV co-infection at present. The incidence of syphilis and HIV co-infection is higher in males [27][28][29] , so this study aims to investigate the effect of syphilis infection on the microstructure of WM in HIV-infected male patients through comparing the differences of WM between HIV-infected male patients with and without syphilitic infection. We hypothesized that the degree of damage of microstructure in WM in HIV-infected male patients with syphilis was more severe than in HIV-infected male patients without syphilis. And this study may help us understand the mechanism of cognitive decline in patients with HIV and syphilis co-infection.

Participant selection
Between June 2015 to June 2017, 27 HIV-infected male patients with current syphilis or a history of syphilis (HIV+/syphilis+) and 29 HIV-infected male patients without syphilis co-infection history (HIV+/syphilis-) were enrolled in this study. This study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the ethics committee of Beijing Youan Hospital, Capital Medical University and informed consent was obtained from all patients or their relatives (guardians). The inclusion criteria were as follows: (1) Patients with HIV was diagnosed by an enzyme-linked immunosorbent assay and western blot analysis; (2) Patients recieved T pallidum particle agglutination (TP-PA) and the Rapid Plasma Reagin (RPR); (3) Male; (4) Age had to be from 18 to 50 years; (5) Patients underwent a comprehensive neuropsychological assessment and had complete clinical and imaging data. The exclusion criteria were as follows: (1) HIV/AIDS patients transmitted by drug abuse or mother-to-child vertical transmission; (2) Anxiety, depression,alcoholism, stimulant use, drug side effects, metabolic encephalopathy, vitamin B12 de ciency and drug interaction; (3) Central nervous system diseases: tumors, cerebrovascular diseases and other visible diseases on MRI (T1WI and T2-uid attenuated inversion recovery (FLAIR)); (4) Patients with MR contraindications: pacemakers, de brillators, implanted electronic systems, vascular clips, mechanical heart valves or cochlear implants; (5) Patients with a sudden onset of illness.
To assess the cognitive status, all patients underwent a comprehensive neuropsychological assessment which included six cognitive domains. The neurocognitive evaluation surveys contained: ne motor skills(Grooved Pegboard Test), information processing speed (Trail Marking Test A, TMT-A), memory(Hopkins Verbal Learning Test, HVLT-R; Brief Visuospatial Memory Test, BVMT-R), abstraction and executive function (Wisconsin Card Sorting Tests, WCST-64), attention and working memory (Continuous Performance Test-Identical Pair, CPT-IP; Wechsler Memory Scale, WMS-III; Paced Auditory Serial Addition Test, PASAT) and verbal/language (Animal Verbal Fluency Test, AFT). Raw scores for each test is converted to T-scores, which is adjusted for age, and education level. T-scores on over one test for one cognitive domain were averaged to get a mean T-scores.
Clinical and demographic data including age, duration of infection, CD4 count, the CD4/CD8 ratios, treatment condition and HIV viral load were collected. The duration of HIV infection was described by the patient's own account. The recent CD4+ counts and the CD4/CD8 ratios were collected within 2 weeks of the MRI.
All HIV-infected patients received TP-PA and RPR to diagnose the current status of syphilis infection or a history of syphilis. All the patents with current syphilis infection or a history of syphilis were administered with penicillin during diagnosis.

MRI data processing
DTI data were processed using FSL5.0.9(http://neuro.debian.net/). There were three steps for data preprocessing. Firstly, eddy current correction with FDT diffusion module was adopted to eliminate head movement and deformation caused by head movement and eddy current during scanning. Secondly, brain mask image was performed using the Brain Extraction Tool, the fractional intensity threshold was 0.2. Finally, the tensor were calculated using the function of DTIFIT with FDT diffusion module in FSL, and the four parameters: FA, MD, AD and RD were obtained. There were four steps for tract-based spatial statistics (TBSS) processing. Firstly, all subjects' FA is registered to a target brain image template (FMRIB58_FA) using nonlinear registration criteria. Secondly, the mean of all FA images was skeletonized to generate a mean FA skeleton image. Then skeletonized FA map of all subjects were constructed by projecting all FA data onto the mean FA skeleton image. The skeletonized maps of other parameters (MD, AD and RD) were obtained similarly. Finally, the skeleton-based statistical analysis was conducted for the four parameters.

Statistical analysis
Analysis was conducted with IBM SPSS Statistics25.0. All normally distributed variables were reported as mean ±SD, while non-normally distributed variables were reported as median (25th-75 th percentile). T-test was conducted for normally distributed data and Mann-Whitney U Test was conducted for non-normally distributed data. Chi-square Test was conducted for categorical data. P<0.05 was considered statistically signi cant.
For TBSS analysis, voxel-wise statistics of the four DTI parameters (FA, MD, AD and RD) between the two groups were compared using general linear model (GLM) module in FSL. Then, randomise in FSL is adopted to conduct the displacement test (5000 random permutations). The threshold-free cluster enhancement (TFCE) approach was adopted for multiple comparison correction. All the results were considered signi cant with a family-wise error (FWE) correction at the level p<0.05. The Johns Hopkins University (JHU)-ICBM-DTI-81 WM Label Atlas was adopted to identify the differences in white matter bers.
Multiple linear regression analysis was conducted to investigate the relationships between DTI metrics and clinical variables and cognitive performance.

Clinical and demographic data
This study enrolled 27 HIV+/syphilis+ male patients, 29 HIV+/syphilis-male patients. The clinical and demographic data were listed in Table 1. In our study, there were two kinds of patients in HIV+/syphilis+ group: 11 patients had a history of primary syphilis and 16 patients were diagnosed as primary syphilis. There were no signi cant differences in age, CD4 count, CD4/CD8 ratios, duration of infection treatment conditions or ratio of target not detected (TND) of HIV viral load. Table 2 showed the differences in neuropsychological assessment between the group of HIV+/syphilis+ and HIV+/syphilis-. There were no signi cant differences in verbal uency (VF), attention/working memory (A/WM), executive function (EF), memory (learning/delayed recall, LDR) and speed of information processing (SIP). But there was signi cant differences in complex motor skills (CMS) (p=0.009).

TBSS results
All patients enrolled in this study had nished MRI scanning. After confounding variables, there were no signi cant differences in MD between HIV+/Syphilis+ group and HIV+/syphilis-group.  (Fig. 1).
Compared with HIV+/syphilis-group, there was a decrease in AD in the HIV+/syphilis+ group in these areas: the right SCR (SCR-R) and the BCC. (Fig. 2).
Compared with HIV+/syphilis-group, there was an increase in RD in the HIV+/syphilis+ group in these areas: the bilateral PCR, the right PTR (PTR-R), the left SCR (SCR-L), SCC and BCC. (Fig. 3).

Correlations between DTI metrics and clinical variables
AD in BCC were negatively correlated with CD4/CD8 ratios(P<0.05) and SCR-R were positively correlated with CD4/CD8 ratios(p<0.1). The FA values in different regions have different effects on the clinical data (Fig. 4).

Correlations between DTI metrics and cognitive performance
RD in SCC and SCR-L was negatively correlated with CMS(P<0.05); RD in PTR-R was positively correlated with CMS(P<0.1). AD in SCR-R was positively correlated with CMS(P<0.05). The FA values in different regions have different effects on the functions of cognitive performance (Fig. 5).

Discussion
In our study, we investigate the effects of syphilis infection on brain structural changes in HIV-infected patients and showed that compared with the HIV+/syphilis-group, the FA is widely decreased in the HIV+/syphilis + group. In addition, the HIV+/syphilis + group showed decreased AD in SCR-R and BCC, increased RD in the bilateral PCR, PTR-R, SCR-L, SCC and BCC.
HIV can damage the microstructure of WM in HIV-infected individuals and is more sensitive to WM bers around the lateral ventricles, such as CC, PCR, PTR, SCR and ACR [26,30−34] , but the exact reason is unclear. One possible reason is that the virus can be present in the CSF of the lateral ventricle, when the blood CSF barrier is broken, the virus can invade adjacent brain tissue [26,35] . In our study, compared with the HIV+/syphilis-group, widespread decreased FA were found in the area surrounding the lateral ventricle in the HIV+/syphilis + group, the altered areas of AD and RD were also found around the lateral ventricles. The distribution characteristics of abnormal regions are similar to HIV-infected patients. Patients with HIV and syphilis co-infections can have a higher CSF HIV viral load [17] , the CSF HIVRNA was positively correlated with the levels of various in ammatory cytokines [36] , thus a higher CSF HIV viral load may indicate a more active intracranial immune activation. ART introduction can result in reducing CSF HIV viral load but no respond to immune activation [37] . This may illustrate why the brain tissue around the lateral ventricle are more severely damaged despite treatment.
Factors in uencing diffusion anisotropy include pathways of neuronal ber bundles, diameter of neuraxon, whether the neuraxon are tightly packed, the thickness of myelin sheath and structures other than the axonal bers [38] . RD may be the predominant factor contributing to the changes of FA [26,39−40] , both changes in AD and RD can lead to increased in MD [26,36] . However, our study showed a widespread decreased FA in the whole brain without any differences in MD, the abnormal regions for AD and RD were also relatively smaller. The reason maybe that the destruction of integrity of WM in this group was caused not only by the combined action of demyelination and axonal injury, but also by the damage of the collagenous perivascular brous alae [38] . Widespread decreased FA suggested compared to patients who are simply infected with HIV, the integrity of WM is more seriously impaired in HIV-infected patients with syphilis co-infection. Decreased AD are concentrated in the BCC and SCR and increased RD are concentrated in the SCR, PTR, PCR and SCC, this result may indicate syphilis infection leads to demyelination changes in the posterior region of the lateral ventricles and acute axonal injury in the upper region of the lateral ventricles in this study group, but the exact mechanism is unclear, future studies will be needed to better understand them.
A few studies have explored the relationship between structure of WM and cognitive function, and found that compared to patients without cognitive impairment, patients with cognitive impairment had a decreased FA of whole brain and more extensive increased MD [41][42] . In our study, decreased FA in different regions of the internal capsule have a negative effect on the function of cognitive performance. The internal capsule connects the cerebral cortex to the brainstem and spinal cord. Damage to this region may result in motor and sensory impairment and then affects cognitive performance. It is worth noting that the increased RD in SCC and SCR-L can affect the ability of CMS, decreased AD in SCR-R can damage the ability of CMS. The SCC connects parts of parietal bers and may be functionally associated with visual area of the occipital cortex. SCR is part of projection bers connecting the internal capsule to the cerebral cortex. Damage to these two regions may affect CMS. Furthermore, motor skill learning requires active central myelination, myelin can increase the speed of electrical communication among neurons [43] . So demyelination may affect motor skills, and this may be the reason why increased RD was correlated with the degradation of CMS in our study.
Our study also showed that CD4 count in HIV+/Syphilis + group was higher than that in HIV+/Syphilisgroup, but the difference was not statistically signi cant. This may be due to the fact that all enrolled syphilis patients were treated with penicillin. Though syphilis had a negative impact on the CD4T cell counts in HIV-positive patients, it can be restored after using penicillin [12] . We also found AD in BCC was negatively correlated with CD4/CD8 ratios, AD in SCR-R was positively correlated with CD4/CD8 ratios.
Lower CD4/CD8 ratios is associated with persistent in ammation and immunosenescence [41] . Our result demostrated that immunologic dysfunction may cause axonal injury of WM around the lateral ventricle. More studies needs to be explored in the future to understand the mechanism of immune status affecting brain structure.
Our study has some limitations. Firstly, the limited sample size may result in lower power, thus a study of larger samples should be done in the future; secondly, our study includes only male gender, which may prevent the results applicable to all populations or female patients. thirdly, a longitudinal follow-up study should be done to investigate the dynamic changes of WM between the two groups.

Conclusions
This study showed that compared to male patients simply infected with HIV, the integrity of WM is more seriously impaired in HIV-infected male patients with syphilis co-infection, and it may accelerate the impairment of cognitive function, so more attention should be paid to this group and more aggressive measures are needed to delay or halt neurological deterioration in this group. This study may provide

Declarations
Ethical approval and consent to participate This research was approved by the ethics committee of Beijing Youan Hospital. All participants provided written informed consent prior to enrolment.

Consent for publication
Not Applicable.

Availability of data and materials
The datasets used in this manuscript will be available from the corresponding author on reasonable request.

Competing Interests
Authors in this article have no con icts of interest to declare.  The differences of FA between HIV+/Syphilis+ group and HIV+/syphilis-group. Areas in red regions were signi cantly increased (P<0.05). The number below each brain image indicates the Z coordinate in the Montreal Neurological Institute (MNI) space.

Figure 2
The differences of AD between HIV+/Syphilis+ group and HIV+/syphilis-group. Areas in blue regions were signi cantly increased (P<0.05). The number below each brain image indicates the Z coordinate in the MNI space.

Figure 3
The differences of RD between HIV+/Syphilis+ group and HIV+/syphilis-group. Areas in green regions were signi cantly increased (P<0.05). The number below each brain image indicates the Z coordinate in the MNI space. Figure 4