Seroprevalence and characterisation of herpes simplex virus from human immunodeficiency virus in samples collected from two provinces in South Africa: a retrospective study

Background: Herpes simplex virus (HSV) is a widely distributed human pathogen that is known for its ulcerative lesions at the infection site. HSV can cause persistent infection in the host that is often followed by a period of latency within the neurons. Considering the high rate of HIV infection in South Africa, it is important to assess the seroprevalence of HSV with a focus to determine the epidemiological association between HSV-DNA and HIV-1 in the population. Methods: A total of 44 sera samples were screened for HSV and HIV-1 using the highly sensitive enzyme-linked immunosorbent assay (ELISA). The ELISA positive samples were characterized using polymerase chain reaction (PCR) to confirm the positivity of both viruses and to further differentiate HSV into HSV-1 and -2. Thereafter, the samples were analysed for relatedness using phylogenetic analysis. Results: Of the 44 samples, 36 (81.8%) were positive for HIV-1, while 35 (79.5%) were positive for HSV when screened with ELISA kits. The PCR results, with the use of type specific primers, showed that 4/35 (11.4%) samples were specific for HSV-1 while 30/35 (85.7%) were specific for HSV-2. Statistical analysis performed using the chi-squared goodness-of-fit test showed that there is a significant relationship between HSV-2 and HIV-1 transmission. Conclusions: There is a significant positive association between HSV-2 and HIV-1 in the study population. Our study shows that some of the HSV-2 isolates are not related to the clinical isolate SD90e from South Africa, suggesting diversity in HSV-2 viral transmission.

Accordingly, the current study aims to establish the prevalence of HSV antibodies and HSV-DNA in HIV-1 sera and further assess the evidence of HSV-2 and HIV-1 co-infection within the study cohort.

Sample collection criteria and study population
The sera samples that were used in the study were previously stored sera collected from patients who visited Bophelong Provincial Hospital in Mafikeng, North-West Province, and Inkosi Albert Luthuli Central Hospital (IALCH-NHLS) in KwaZulu-Natal Province, both in South Africa. The study participants visited the hospitals for HIV screening and management in their respective provinces. Only 25 sera samples from each hospital (50) were selected for use in the current study. However, the sample quality (serum lysis) of six sera from the total samples prevented us from using them and were excluded, reducing the total sample number to forty-four (44). The samples were anonymized with no additional data except age and gender of the study participants.

Laboratory analysis
The enzyme-linked immunosorbent assay (ELISA) was used to detect the presence of HSV and HIV-1 antibodies in the sera. Polymerase chain reaction (PCR) was used to confirm that the samples were infected with HSV and HIV-1 and to differentiate the HSV samples into HSV type 1 and 2.

ELISA
A highly sensitive ELISA test kit for HSV, Platelia HSV (1+2) (Bio-Rad, Marnes-la-Coquette, Paris, France) was used to measure HSV IgG antibody in the samples. The Genscreen Ultra HIV Ag-Ab test kit (Bio-Rad, Marnes-la-Coquette, Paris, France) was used to detect HIV-1 p24 antigen in the sera. The ELISA plate was pre-coated with biotinylated polyclonal antibody to p24 HIV-1 Ag. The manufacturer's instructions were followed with slight modifications. Briefly, the sera were diluted in a ratio 1:10 and 100 µl of the diluted sera, blank, positive, and negative controls were added to a flat-bottom 96-well plate. The plate was incubated at 37°C for 45 min and washed four times. Thereafter, the plate was incubated at room temperature for 15 min with 100 µl of substrate followed by 100 µl of stop solution (0.3M of H 2 SO 4 , Bio-Rad) to terminate the reaction. The plate was read on a microplate reader at 450 nm. The samples were run in duplicate and were considered positive for HSV IgG antibody if the ratio of the average of serum OD and the cut-off value was greater than 1.2, and negative if the ratio was less than 0.8.

DNA isolation and PCR
DNA and RNA were extracted from the sera samples using QIAamp® MinElute® Virus Spin kit and QIAamp® Viral RNA Mini kit (Whitehead Scientific, Cape Town, South Africa), respectively, following the manufacturer's protocols.  (Table A). The HSV positive samples were also tested for HIV-1 co-infection using PCR. This allowed the detection of HIV-1 and HSV-1 and/or -2 co-infections from the study samples. Briefly, 25 µl PCR reaction mixture (Quick-Load® Taq 2X Master Mix kit, Biolabs) with each primer set targeting the different regions, was prepared (Extended data, Table B) and amplification was performed with the use of a T100 TM thermal cycler (Bio-Rad, Hercules, California, United States). HSV-1 and -2 amplification was performed with the nested PCR cycling conditions outlined in the Extended data (Table B). A reverse transcription PCR was used to amplify the HIV-1 integrase gene from the extracted HIV-1 RNA samples. This was achieved with the PCR cycling conditions outlined in the Extended data (Table C) and the PCR products were analysed by gel electrophoreses.

DNA sequencing
Next-generation sequencing (NGS) was used to validate the genomes of the samples used. Only the samples that exhibited high titers of HSV-2 and HIV-1 with ELISA were sequenced. Thus, DNA sequencing was performed on four HSV-2/HIV-1 co-infected samples with the highest titers. The samples were annotated as G13, G15, G20 and G34. HSV-2 primers were used to sequence G13, G15 and G34 while the G20 sample was sequenced using HIV-1 primers since this was the only sample with a high HIV-1 titer. NGS was carried out on the Illumina MiSeq NGS platform at Inqaba Biotec (Inqaba Biotechnical Industries (Pty) Ltd, Johannesburg, SA). Phylogenetic analysis was performed by trimming and aligning the sequences using BowTie 2 v 2.3.2 (Langmead & Salzberg, 2012). All aligned data were further annotated to determine the viral genome using Prokka v 1.12 (Seemann, 2014). Thereafter, the sequence data were subjected to Molecular Evolutionary Genetics Analysis (MEGA 7) against HIV-1 and HSV-2 reference genomes obtained from National Centre for Biotechnology Information (NCBI). Phylogenetic analysis was performed to assess the evolutionary relatedness of the sequenced data in relation to the reference genomes. The HSV-2 published reference sequences that were used in the analysis were SD90e (KF781518) from South Africa, HSV-2 strain 333 (M15118), a wild-type laboratory reference strain from USA, and glycoprotein G-2 (AF141858), a European HSV-2 isolate. The HIV-1 reference sequences that were used in the analysis were two HIV-1 subtype C sequences from South Africa (HM569277 and HM569273), two HIV-1 subtype B from USA (AF203332) and Japan (LC022388), and HIV-1 subtype A (HM466997) from Europe. Thus, the reference sequences were selected to explore the relatedness of the sequenced samples with published reference strains from the country (South Africa) and from other developed countries. In addition, different HIV-1 subtypes were selected from different regions to identify the sub-type of our clinical isolate.

Statistical analysis
Data analyses were carried out on Statistical Package of Social Sciences (SPSS) software (version 25). The chi-squared goodness-of-fit test (x 2 ) was used to evaluate for an association between the categorical variables. Relationship between the demographics and the viruses was tested using Pearson correlation coefficient. The 5% significance level was considered as a significant p value in this study.

Ethical approval
The study received ethical approval from the North-West University Research Ethics Regulatory Committee (NWU-00068-15-A9).

Results
The demographics of the study population showed that majority of the study participants were female (79.5%) with a low percentage of males (20.5%) ( Table 1). The mean age and standard deviation of the study population were 33.09 ±11.94 years.
ELISA screening of the samples showed that 36/44 (81.8%) were seropositive for HIV-1 while 35/44 (79.5%) were positive for either HSV-1/2 antibody (Table 1 and Extended data, Figure A). Notably, the study participants within the age group of 21-40 years had the highest HSV and HIV-1 infection rates, as depicted in Table 1.
Furthermore, PCR was also used to amplify the integrase gene of HIV-1 genome in the samples and the data showed that 36/44 (81.8%) samples that were positive for HIV-1 p24 using ELISA were also PCR positive.
NGS was performed to confirm the genomes of HSV-2 and HIV-1 viral isolates that were detected using PCR. The samples that were sent for NGS sequencing were those that exhibited high titers with ELISA, that is, HSV-2 sequences 13, 15 and 34 and HIV-1 sequence G20. The sequenced data was compared with reference sequences obtained from NCBI that were selected based on the amplified targeted regions. A maximum likelihood phylogeny method of analysis was used with a bootstrap value of 1,000 replicates to generate evolutionary trees for HSV-2 and HIV-1, respectively. The phylogenetic analysis of HSV-2 sequence data showed that sequences 13 and 15 are more closely related to glycoprotein G2 reference strain than sequence 34, with 98-99% similarity to the G2 strain ( Figure 2).
The HIV-1 phylogenetic tree in Figure 3 shows that G20 sequence is 100% closely related to subtype A (HM466997) and distantly related to subtype B (AF203332 and LC022388) and subtype C (HM569277 and HM569273) reference genomes.
The SPSS v25 statistical software was used to detect whether there is a relationship between age, HSV, and HIV-1. SPSS was also utilized to evaluate the association between HSV and HIV-1 positive samples (Table 3). The two-tailed correlation test exhibited a statistical positive link between age and HSV-1 (0.366**) as shown in Table 3. HSV-2 and HIV-1 samples are also positively correlated (0.690**). In addition, an inverse or negative relationship between HSV-1 and HSV-2 (-0.463**) was detected.
The SPSS chi-square goodness-of-fit test was also used to assess association between HSV and HIV-1 positive samples. The data showed that there was no significant association between HSV-1 and HIV-1 (X 2 (1) = 0.138, p>0.05). However, there was a strong statistical association between HSV-2 and HIV-1 (X 2 (1) = 20.952, p< 0.05), as shown in Table 4.

Discussion
There are insufficient data to recount the prevalence of HSV genotypes among HIV-1 infected individuals in the Republic of South Africa. This study provides insight into an existing interrelation between HSV and HIV-1 and the potential risk that one of the viruses may have on the other. Our focus was to The study showed that more children will reach the age of sexual debut with no antibody protection against HSV-1. In addition, the low rate is attributable to the change in disease spread in the population since there are reduced viral HSV-1 antibodies at a very young age, a factor influencing HSV-2 acquisition.
Our study also revealed a higher rate of HSV-2/HIV-1 co-infection (13.6% males, 54.5% females) compared with HSV-1/HIV-1 (2.3% males, 4.5% females).Thus, one of the few observations drawn from this study was the relatively high prevalence of HSV-2 compared to HSV-1 in HIV-1 co-infected samples. The validity of the high prevalence of HSV-2/HIV-1 co-infection in this study is supported by previous studies    A previous study on the relationship between age and HSV-1 prevalence by Smith & Robinson (2002) reported that global HSV-1 prevalence increases with age. The statistical relationship between age, HSV and HIV-1 infection in our study revealed a similar significant relationship between age and HSV-1 (p = 0.366*). Similarly, we discovered a discreet relationship between age, HSV-2 and HIV-1. However, HSV-1 showed an inverse correlation with HSV-2 (p = -0.463 ** ). This may suggest that an increase in HSV-2 prevalence in the population will result in a decline in HSV-1 as supported by Ayoub et al. (2019). It was also discovered that a robust significant relationship exists between HSV-2 and HIV-1 (p =0.690 ** ) suggesting that a steady rise in HSV-2 contributes to an increase in HIV-1 infection in the population. A probable reason for this relationship is that the viruses share a similar route of entry and the impact of one is significant on the other as seen in the micro-ulceration of the genitalia in HSV-2 patients, which provides a port of entry for HIV-1 (Sheth et al., 2008).
The association between HIV-1 and HSV was analysed using chi-square goodness-of-fit test and it was discovered that no significant association exist between HSV-1 and HIV-1 (X 2 (1) = 0.138, p >0.05) but a strong statistical association was found between HSV-2 and HIV-1 (X 2 (1) = 20.952, p <0.05).
Although, there has been contrasting opinions on the association of HSV-2 and HIV-1, the current study is consistent

Limitations
The study is a retrospective study with the aim to determine the prevalence of HSV-DNA in HIV-1 sera. However, there were certain limitations. Primarily, the study population was small when compared to the size of the general population.
In addition, the sera samples were collected exclusively from North-West and KwaZulu-Natal Provinces, poses a challenge to generalize the outcomes of the study to the South African populace.

Conclusion
This study revealed a clustering variation of HSV-2 sequences. The clustering differentiated the sequences from the prominent sequence (SD90e) found in South Africa, which suggests diversity in the transmission of the virus. Similarly, a different HIV-1 sub-type was isolated from the study population, different from other HIV-1 sub-type C isolates in South Africa. Despite the small sample size, a high prevalence of HSV-2/HIV-1 co-infection (68.2%) was recorded which shows that a positive association exists between HSV-2 and HIV-1, suggesting that an increase in one of the viruses may influence the spread of the other. Thus, acknowledging that a relationship exists between these two viruses, and to identify how the transmission of one could affect the other, requires a larger cohort that is well described with longitudinal measurements of HSV-1, HSV-2, and HIV-1 as well as measurements of potential confounders such as condom use, partner change and other sexually transmitted diseases. This project contains the following extended data: - Table A: Primers for the detection of HSV-1, HSV-2 and HIV-1.

Data availability
- Table B: PCR reaction mixture and thermocycling conditions for HSV-1 and -2.

Bongiwe Ndlovu
HIV Pathogenesis Programme, University of KwaZulu-Natal, Durban, South Africa I am happy with the revised manuscript. Although, a few points should be addressed before the study is indexed. Conclusions should be revised, the authors mentioned that "There is a significant relationship between HSV-2 and HIV-1 in the study population". Are they referring to a positive association? This should be clarified.
In sample collection criteria and study population, the authors mentioned that "six samples were lysed and were excluded". This is confusing, are they referring to poor sample quality? This should also be clarified. Which stop solution did they use to terminate ELISA experiments? We would like to thank you for considering our manuscript (28105) entitled "Seroprevalence and characterisation of herpes simplex virus from human immunodeficiency virus in samples collected from two provinces in South Africa: a retrospective study" for publication in your reputable journal. We also appreciate the reviewers' efforts, comments, and concerns in making the manuscript better. We have gone through the reviewer's comments and edited the manuscript accordingly in the revised manuscript. The inputs made in response to the reviewer's comments are listed below.

I confirm that I have read this submission and believe that I have an appropriate level of Bongiwe Ndlovu
HIV Pathogenesis Programme, University of KwaZulu-Natal, Durban, South Africa Study Title: The title of the paper should be revised, it is unclear. The location of samples collected should be removed from the title and properly explained in the Methods section.
The introduction: The authors should restructure the sentences. Some sentences are long and they have typographical errors. Some statements are inaccurate, for example: they mentioned that transmission is asymptomatic.

Methods:
The methods section should be revised. The authors should indicate the type of samples used for the study (blood, plasma, or serum), the country of origin, and the hospitals/ institutions where the samples were collected. They should also define convenience sampling.
Laboratory Analysis: The authors should restructure the sentences. They also used abbreviations like EIA & PCR without describing them first/ writing in full.
ELISA method was not properly described. Which HIV-1/ HSV proteins did they use to measure antibody responses?
The authors also did not specify which genes of interest and primers were used to amplify the HIV-1 genome with nested PCR. This should be clearly explained and referred to supplementary information.
Sequencing: It is not clear why G13, G15, G34 were sequenced with HSV-2 primers while G20 was amplified with HIV-1 primers.

Results:
Did you measure p24 antibody titers? ○ Figure 1 is confusing, the authors used too many colors. They should consider simplifying the figure by showing a bar graph with positive and negative responses for each age group. The figure legend is too brief.

○
The authors did HSV1/2 PCR. However, the results are not shown in the article. The bar graphs are only shown in the supplementary information. These should be simplified and added as figure 2 of the manuscript. Where is the figure that shows HIV-1/HSV coinfections? These can be summarized clearly in a pie chart.

○
The interpretation of the sequencing data needs improvement. Which HIV-1 subtypes are HIV LC201873, 4533/11 and KU609428? Are they South African strains? What was the aim of selecting these strains? ○ HIV-1 LC022388 and 3793/15 KU609388) were used in constructing the phylogenetic tree. However, it is not clear why these strains were selected for analysis.

○
The relationships between age and HSV-1, HSV-2, and HIV-1 or HSV-1 and HSV-2 are not clearly explained. Are they positive or negative associations?  figure 1 represent the same data. One should be removed, as it is a duplication of data representation.