Alterations of the upper respiratory microbiome among children living with HIV infection in Botswana
====================================================================================================

* Sweta M. Patel
* John Farirai
* Mohamed Z. Patel
* Sefelani Boiditswe
* Leabaneng Tawe
* Shimane Lekalake
* Mogomotsi Matshaba
* Andrew P. Steenhoff
* Tonya Arscott-Mills
* Kristen A. Feemster
* Samir S. Shah
* Nathan Thielman
* Coleen K. Cunningham
* Lawrence A. David
* David Murdoch
* Matthew S. Kelly

## ABSTRACT

**Background** Children living with HIV (CLWH) are at high risk of colonization and infection by bacterial respiratory pathogens. Microbes in the upper respiratory microbiome can prevent colonization by these pathogens. The impact of HIV infection on development of the upper respiratory microbiome during childhood is poorly understood.

**Methods** We enrolled healthy CLWH (<5 years) and age- and sex-matched HIV-exposed, uninfected (HEU) and HIV-unexposed, uninfected (HUU) children in a cross-sectional study conducted in Botswana. We used shotgun metagenomic sequencing to compare the nasopharyngeal microbiomes of children by HIV status.

**Findings** Of 143 children, 44 were CLWH, 49 were HEU, and 50 were HUU. Nasopharyngeal microbiome composition differed by HIV status (p=0·043, R2=0·019). The relative abundance of *Corynebacterium pseudodiphtheriticum* was lower in CLWH compared to HEU and HUU children (p=0·01). Among CLWH, a low (<25%) CD4+ cell percentage was associated with microbiome composition (p=0·009, R2=0·042) and lower relative abundances of *Corynebacterium propinquum* (p=0·003), *C. pseudodiphtheriticum* (p=0·007), and *Dolosigranulum pigrum* (p=0·004). The relative abundances of *C. propinquum, C. pseudodiphtheriticum* and *D. pigrum* in the nasopharyngeal microbiome were negatively correlated with the abundances of *Streptococcus pneumoniae* and *Staphylococcus aureus*.

**Interpretation** CLWH with HIV-associated immunosuppression have altered nasopharyngeal microbiome composition and lower abundances of bacterial species associated with respiratory health during childhood. These findings suggest that the upper respiratory microbiome may contribute to the high risk of bacterial respiratory infections among CLWH.

**Funding** National Institutes of Health, Duke Center for AIDS Research, Penn Center for AIDS Research

**Evidence before this study** We searched PubMed for research articles published from database inception through November 20, 2022, using the terms (“nasopharyngeal” OR “nasal” OR “upper respiratory”) AND (“HIV” OR “human immunodeficiency virus”) AND (“microbiome” OR “microbiota”) AND (“pediatric” OR “child” OR “children” OR “infants”). This search returned five articles, three of which collected nasopharyngeal specimens from children living with HIV (CLWH). The objective of two of these articles was pathogen identification using culture- and polymerase chain reaction-based methods. The remaining study characterized the nasopharyngeal microbiomes of children with pneumonia, children with upper respiratory infections, and healthy children in Botswana using 16S rRNA sequencing. Genera associated with respiratory health were less abundant in CLWH with pneumonia, but no data was available for healthy CLWH. Thus, it remained unknown if the microbiome alterations observed in CLWH were associated with HIV or with pneumonia.

**Added value of this study** To our knowledge, this is the first study to investigate the nasopharyngeal microbiome in healthy CLWH using metagenomic sequencing. To account for shifts in the microbiome that occur with age, we enrolled age- and sex-matched HIV-exposed, uninfected and HIV-unexposed children for each CLWH. The use of shotgun metagenomic sequencing allowed us to investigate differences in the microbiome at the species level. We found that HIV infection and HIV-associated immunosuppression were associated with an altered nasopharyngeal microbiome and a lower abundance of species associated with respiratory health and resistance to colonization by common bacterial respiratory pathogens.

Implications of all the available evidence

These findings suggest that HIV-associated alterations in the nasopharyngeal microbiome may predispose CLWH to colonization by bacterial respiratory pathogens responsible for invasive infection and death. Strategies to reduce pathogen colonization through modification of the microbiome hold promise for reducing infectious morbidity and mortality in CLWH.

Keywords
*   pediatric nasopharyngeal microbiota
*   bacterial colonization resistance
*   childhood respiratory infections
*   *Corynebacterium* species
*   *Dolosigranulum pigrum*
*   shotgun metagenomic sequencing

## INTRODUCTION

Despite widespread use of the 13-valent pneumococcal conjugate (PCV-13) and *H. influenzae* type B (Hib) vaccines, bacterial respiratory pathogens such as *Streptococcus pneumoniae* (pneumococcus), *Haemophilus influenzae*, and *Staphylococcus aureus* account for considerable child morbidity and mortality worldwide.1,2 The nasopharynx is an important reservoir for these pathogens, and nasopharyngeal colonization precedes the development of invasive infection.3 Establishment of colonization is influenced both by host-microbe interactions and interactions with other microbes residing within the nasopharynx. For example, *Corynebacterium accolens* hydrolyzes human triacylglycerols to produce free fatty acids that inhibit pneumococcal growth.4 Given the role that microbes in the nasopharynx play in resisting colonization and infections caused by bacterial respiratory pathogens, targeted manipulation of the nasopharyngeal microbiome is a promising strategy for the prevention or treatment of these infections.

The nasopharyngeal microbiome may represent a particularly important therapeutic target in children living with HIV (CLWH). CLWH have higher incidences of nasopharyngeal colonization and invasive infections caused by pneumococcus and *S. aureus*.5,6 There are likely multiple mechanisms that contribute to the high risks of bacterial pathogen colonization and infection among CLWH. Maternal HIV infection impairs placental transfer of antibodies to several respiratory pathogens, including pneumococcus and *H. influenzae*.*7* Socioeconomic factors, replacement feeding, and undernutrition may further contribute but fail to fully explain the rates of infections observed in this population. Variations in the upper respiratory microbiome may also affect the risk of bacterial respiratory pathogen colonization and infection among CLWH. Prior studies demonstrated that the gut and oral microbiomes are altered in CLWH.8,9 However, there are few data on associations between HIV infection and the nasopharyngeal microbiome.10

In this study, we sought to identify associations between HIV infection and the nasopharyngeal microbiomes of 153 healthy children in Botswana. We used shotgun metagenomic sequencing to compare the nasopharyngeal microbiomes of CLWH, HIV-exposed, uninfected (HEU) children, and HIV-unexposed, uninfected (HUU) children. We identified HIV-related factors and environmental exposures associated with nasopharyngeal microbiome composition during childhood.

## METHODS

### Setting

This study was conducted from February 2020 to February 2021 at four public clinics in Gaborone, Botswana. The country’s HIV prevalence among pregnant women is 24%.11 The government of Botswana provides antiretroviral therapy (ART) at no cost, and 98% of pregnant women receive ART for prevention of vertical HIV transmission. The country’s childhood immunization program is fully funded by the government and includes vaccines against common respiratory pathogens, including Hib, measles, diphtheria, pertussis, and pneumococcus. Botswana incorporated a pentavalent vaccine containing Hib (Serum Institute of India) into its immunization program in November 2010 and PCV-13 (Prevnar-13; Pfizer) in July 2012; both vaccines are administered in 3+0 schedules with doses at 2, 3, and 4 months of age. The estimated national coverage rates among infants in 2020 for the complete Hib and PCV-13 vaccine series were 95% and 90%, respectively.12

### Study Population

Children under five years of age with documentation of child and maternal HIV testing were eligible for inclusion. We excluded children with symptoms of clinical pneumonia (cough with rapid breathing or lower chest wall indrawing), craniofacial abnormalities, or chronic pulmonary disease (including asthma). For each child living with HIV, we enrolled an HEU and an HUU child, each sex- and age-matched (±3 months). Additionally, we enrolled any HEU siblings under five years of age with the same biological mother who lived in the same household as enrolled CLWH. This was done to more precisely identify associations between HIV infection and microbiome composition, given that environmental exposures among these sibling pairs were expected to be similar. Written informed consent was obtained for all study participants. This study was approved by the Health Research and Development Committee (Ministry of Health, Botswana) and institutional review boards at the University of Botswana, the Botswana-Baylor Children’s Clinical Centre of Excellence, the University of Pennsylvania, and the Duke University Health System.

### Data collection

We collected sociodemographic and clinical data through physical examination, a face-to-face questionnaire with the child’s guardian, and review of medical records. Children with a positive HIV PCR were classified as CLWH, while children born to women with HIV infection at or before delivery were classified as HEU if they had documentation of a negative HIV PCR after 6 weeks of age if exclusively formula fed or at least 6 weeks after breastfeeding cessation. Children born to mothers with negative testing for HIV during or after pregnancy were classified as HUU. Using the most recent results available at the time of enrollment, we defined viral suppression as an HIV RNA <400 copies/mL, a normal CD4+ cell percentage as >25%, and a low CD4+ cell percentage as <25%.

### Laboratory testing

Nasopharyngeal specimens were collected using flocked swabs and MSwab medium (Copan Italia, Brescia, Italy). Samples were transported to the National Health Laboratory in Gaborone, transferred to cryovials, and stored at -80°C before shipment on dry ice to Duke University (Durham, NC, USA). The Duke Microbiome Core Facility extracted DNA from nasopharyngeal specimens using DNeasy PowerSoil Pro kits (Qiagen, Hilden, Germany). Sequencing libraries were generated using DNA Prep kits (Illumina, San Diego, CA, USA). Libraries were pooled at equimolar concentrations and sequenced as 150 base pair paired-end reads on a NovaSeq6000 instrument (Illumina). Negative extraction controls were included on the sequencing run.

### Bioinformatics processing

Initial filtering and quality assessment of the sequencing data were performed using the default settings in the KneadData toolkit, which filters reads based on nucleotide quality, adapter content, contamination from human reads, and low-information tandem repeats using Trimmomatic and Bowtie 2. Kaiju was used to assign taxonomy to sequencing reads using the proGenomes database version 2·1 as the reference, maximum exact match mode, and a default minimum match length of m=11. Taxa with fewer than 10 sequencing reads across the dataset were filtered, as were non-bacterial reads. We used the *decontam* R package version 1·14 to remove presumed reagent contaminants (n=53) based on negative correlation with DNA concentration using the frequency method (threshold=0·1; Table S1). Samples with fewer than 2,500 high-quality paired-end reads were excluded from further analyses. *Cutibacterium acnes* was the only species identified in the negative extraction control.

### Statistical analysis

We compared characteristics of CLWH, HEU children, and HUU children using Chi-squared or Fisher’s exact tests for categorical variables and Kruskal-Wallis tests for continuous variables. We used multivariable linear regression to evaluate associations between HIV status and the Shannon index and log-transformed Chao1 richness of the microbiome. We used the *microbiome* R package version 1·16 to generate centered log-ratio (CLR)-transformed sample counts to evaluate between-sample compositional differences. We used permutational analysis of variance (PERMANOVA) to compare overall microbiome composition by HIV status. We used paired PERMANOVA testing to compare microbiome composition within sibling pairs. After filtering taxa with fewer than 50 counts across the dataset, we fit multivariable general linear models using MaAsLin 2·0 to evaluate for differences in the relative abundances of individual bacterial species by HIV status.13

For analyses limited to CLWH, we filtered taxa as described above and used the linear discriminant effect size (LEfSe) method to identify differentially abundant species by CD4+ cell percentage, viral load, use of trimethoprim-sulfamethoxazole (TMP-SMX) prophylaxis, and antibiotic treatment in the preceding three months.14 We chose LEfSe for these analyses over multivariable general linear models due to the small sample size. Given the small number of samples for analysis, we used Wilcoxon signed-rank tests to compare the relative abundances of the ten most common bacterial species within sibling pairs. Lastly, we used Spearman’s rank correlation coefficients to evaluate associations between the relative abundances of specific bacterial species. Unless otherwise noted, we adjusted all analyses for potential confounding variables identified based on a review of the literature: season, age, household use of wood for cooking or heating, receipt of antibiotics within the preceding three months, and upper respiratory infection symptoms (nasal congestion, nasal discharge, or cough) in the preceding one month.15-17 We conducted all statistical analyses using R software version 4·1·2.

### Role of the funding source

The study sponsors had no role in study design, data collection or analysis, manuscript preparation, or the decision to submit the manuscript for publication.

## RESULTS

### Characteristics of the study population

Of 187 children screened for eligibility, 153 were enrolled (Figure 1). Sufficient microbial reads for analysis were obtained from the nasopharyngeal samples of 143 (93%) children, including 44 (31%) CLWH, 49 (34%) HEU, and 50 (35%) HUU (Table 1). Median (interquartile range, IQR) age was 32 (18, 49) months, and 52% of children were female. Maternal age differed by child HIV status (p=0·0002), but no other differences in sociodemographic characteristics were observed by HIV status. Of CLWH, 42 (95%) were prescribed ART with the most frequent regimen being abacavir-lamivudine-lopinavir/ritonavir (n=38). Median (IQR) CD4+ cell percentage was 30% (24-36%), and 27 of 41 (66%) children with HIV RNA measurements had undetectable viral loads (<400 copies/mL) collected a median (IQR) of 12 (4, 18) weeks prior to enrollment. Eleven of 42 children (26%) had a low CD4+%. Fifteen (36%) CLWH were receiving trimethoprim-sulfamethoxazole (TMP-SMX) prophylaxis. Seven (16%) had received antibiotic treatment in the preceding 3 months.

![Figure 1.](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2022/12/20/2022.12.19.22283664/F1.medium.gif)

[Figure 1.](http://medrxiv.org/content/early/2022/12/20/2022.12.19.22283664/F1)

Figure 1. 
Screening and enrollment of children in Gaborone, Botswana, February 2020 to February 2021.

View this table:
[Table 1.](http://medrxiv.org/content/early/2022/12/20/2022.12.19.22283664/T1)

Table 1. 
Baseline characteristics of the study population by HIV status

### HIV infection is associated with alterations in nasopharyngeal microbiome composition

A total of 6,709 bacterial species from 116 phyla and 2,146 genera were detected in the 143 samples included in this analysis. Median (IQR) Shannon and Chao1 indices were 1·85 (1·51, 2·26) and 137 (68, 261), respectively. There were no significant associations between child HIV status (Figure S1) or other child characteristics and the alpha diversity of the nasopharyngeal microbiome. The composition of the nasopharyngeal microbiome differed by HIV status (Figure 2A, PERMANOVA; p=0·043, R2=0·019), as well as by child age, season of enrollment, and recent URI symptoms (Table S2). To further characterize these compositional differences in the nasopharyngeal microbiome, we used MaAsLin2 to fit generalized linear models evaluating associations between sociodemographic and environmental factors and the abundances of specific bacterial species in the nasopharyngeal microbiome (Figure 3, Table S3). The relative abundance of *Corynebacterium pseudodiphtheriticum* was lower in CLWH compared to HEU and HUU children (Figure 2B, Figure 3B); there were no differences in species abundances between HEU and HUU children. The relative abundance of *Micrococcus luteus* was higher and the relative abundances of multiple *Moraxella* species lower among children enrolled during the rainy season (Figure 3A). In contrast, recent URI symptoms were associated with a higher relative abundance of *Moraxella cuniculi* (Figure 3C). Finally, increasing age was associated with higher relative abundances of multiple non-pneumococcal *Streptococcus* species and *M. luteus* (Figure 3D).

![Figure 2.](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2022/12/20/2022.12.19.22283664/F2.medium.gif)

[Figure 2.](http://medrxiv.org/content/early/2022/12/20/2022.12.19.22283664/F2)

Figure 2. Composition of the nasopharyngeal microbiome by HIV status and HIV-associated immunosuppression among children in Botswana.
A. Principal components plot based on Euclidean distances showing distinct nasopharyngeal microbiome composition by HIV status among children in Botswana (n=143, PERMANOVA: p = 0·043, R2 = 0·019). B. Relative abundances of the ten most abundant bacterial species by HIV status in nasopharyngeal samples from children in Botswana (n=143). C. Principal components plot based on Euclidean distances showing distinct nasopharyngeal microbiome composition by CD4+ cell percentage among children living with HIV (CLWH) in Botswana (n=44, PERMANOVA: p=0·009, R2 = 0·042). D. Relative abundances of the ten most abundant bacterial species by CD4+ cell percentage in nasopharyngeal samples from CLWH in Botswana (n=44).

![Figure 3.](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2022/12/20/2022.12.19.22283664/F3.medium.gif)

[Figure 3.](http://medrxiv.org/content/early/2022/12/20/2022.12.19.22283664/F3)

Figure 3. Associations between sociodemographic and clinical factors and environmental exposures and the composition of the microbiome among children in Botswana.
MaAsLin2 was used to fit log-transformed generalized linear mixed models evaluating associations between subject factors and the relative abundances of bacterial species within the nasopharyngeal microbiome. The coefficients from these models, which correspond to the relative effect sizes of associations, are shown for significant associations (q < 0·20) identified for (A) rainy season, (B) HIV infection, (C) recent URI symptoms, and (D) age. Bacterial species that were higher in relative abundance with the exposure are shown as blue bars. Bacterial species that were lower in relative abundance with the exposure are shown as red bars. Standard errors are represented by the error bars. URI, upper respiratory symptoms

We additionally compared the nasopharyngeal microbiomes of CLWH to their HEU siblings to more precisely identify associations between HIV infection and microbiome composition, given that environmental exposures among sibling pairs in the same household were presumably similar. The nine sibling pairs for whom we had sufficient bacterial sequencing reads did not differ in demographic or clinical characteristics (Table S4). Additionally, overall nasopharyngeal microbiome composition did not differ among HIV-HEU sibling pairs (paired PERMANOVA; p=0·20, R2=0·075). However, multiple species were differentially abundant between CLWH and their HEU siblings (Figure S2). CLWH had lower relative abundances of *Corynebacterium propinquum* (p=0·008), *C. pseudodiphtheriticum* (p=0·008), and *Dolosigranulum pigrum* (p=0·03), and higher relative abundances of *Moraxella lincolnii* (p=0·004) and pneumococcus (p=0·02) than their HEU siblings.

### HIV-associated immunosuppression is associated with altered nasopharyngeal microbiome composition among CLWH

We next sought to identify factors associated with the diversity and composition of the nasopharyngeal microbiome among CLWH. There were no associations between viral suppression status, CD4+%, TMP-SMX exposure, or antibiotic treatment and the alpha diversity of the nasopharyngeal microbiome. Microbiome composition differed between children with low and normal CD4+ cell percentages (Figures 2C and 2D; PERMANOVA; p=0·01, R2=0·042), but did not differ by viral suppression status (p=0·79, R2=0·02), TMP-SMX exposure (p=0·13, R2=0·031), or recent antibiotic treatment (p=0·19, R2=0·029). In differential abundance analyses, *C. propinquum, C. pseudodiphtheriticum, D. pigrum*, and *Alloiococcus otitis* were more abundant in children with normal CD4+% compared to children with low CD4+%, and *Neisseria lactamica* was more abundant in children with low CD4+% (Figure S3A; LEfSe). TMP-SMX exposure was associated with a higher relative abundance of *S. aureus* (Figure S3B). There were no differentially abundant species by viral suppression status or recent antibiotic exposure.

To further evaluate the association between HIV-associated immunosuppression and nasopharyngeal microbiome composition, we compared CLWH and low CD4+% to HEU and HUU children. Microbiome composition differed in children with HIV and low CD4+% compared to HEU and HUU children (PERMANOVA; p=0·003, R2=0·032). *C. propinquum* and *C. accolens* were more abundant in HEU children compared to immunosuppressed CLWH and HUU children, while *C. pseudodiphtheriticum* and *C. aurimucosum* were more abundant in HUU children compared to immunosuppressed CLWH and HEU children (Figure S3C; LEfSe). In contrast, overall microbiome composition (PERMANOVA; p=0·15, R2=0·018) and individual species relative abundances did not differ between CLWH with normal CD4+% and HEU and HUU children. These findings indicate that the observed differences in nasopharyngeal microbiome composition between CLWH and children without HIV were limited primarily to children with HIV-associated immunosuppression.

### Identification of species negatively associated with *S. pneumoniae* and *S. aureus* abundance

Finally, we sought to determine if differences in pneumococcal and *S. aureus* colonization existed by HIV status and how the abundances of other common species were associated with the prevalence and abundance of these two important respiratory pathogens. Pneumococcus was identified in 126 (88%) children and *S. aureus* was identified in 45 (31%) children. There were no differences in pneumococcal colonization prevalence (Chi-squared test; p=0·09) or relative abundance (Kruskal-Wallis test; p=0·55) by HIV status. Conversely, *S. aureus* colonization was more common among HEU children than CLWH or HUU children (p=0·03), though the relative abundance did not differ by HIV status (p=0·05). Given that other microbes in the nasopharynx contribute to colonization resistance to pneumococcus and *S. aureus*, we evaluated correlations between the relative abundances of these potential pathogens and the abundances of other species associated with HIV or HIV-associated immunosuppression. The relative abundances of *C. propinquum, C. pseudodiphtheriticum, and D. pigrum* were negatively correlated with the relative abundances of pneumococcus and *S. aureus* (Figure 4).

![Figure 4.](http://medrxiv.org/http://medrxiv.stage.highwire.org/content/medrxiv/early/2022/12/20/2022.12.19.22283664/F4.medium.gif)

[Figure 4.](http://medrxiv.org/content/early/2022/12/20/2022.12.19.22283664/F4)

Figure 4. Negative interspecies correlations identified in the nasopharyngeal microbiomes of children in Botswana.
Scatter plots depict the correlation between the relative abundances of (A) *C. propinquum* and *S. pneumoniae*, (B) *C. propinquum* and *S. aureus*, (C) *C. pseudodiphtheriticum* and *S. pneumoniae*, (D) *C. pseudodiphtheriticum* and *S. aureus*, (E) *D. pigrum* and *S. pneumoniae*, and (F) *D. pigrum* and *S. aureus*. The Spearman’s rank correlation coefficient and p value for associations are displayed in each panel. Smoothed conditional means are represented by red lines.

## DISCUSSION

In this study, we described nasopharyngeal microbiome composition of healthy children in Botswana and identified associations between HIV infection and microbiome composition. We found that differences in microbiome composition among CLWH were driven primarily by HIV-associated immunosuppression. Finally, we demonstrated negative correlations between species that were less abundant in CLWH and the potential pathogens *S. pneumoniae* and *S. aureus*.

HIV infection is associated with a reduction of potentially protective bacterial species and an increase of potential pathogens in the nasopharynx. For example, CLWH with pneumonia in Botswana exhibited an unexpected near-complete absence of *D. pigrum* in nasopharyngeal samples.10 Additionally, other studies identified a higher prevalence of nasopharyngeal colonization with *S. pneumoniae, S. aureus*, and *H. influenzae* among CLWH.5,6 However, these studies were limited by enrollment of children with pneumonia and the use of culture-based identification methods. Importantly, we identified an association between HIV infection and overall nasopharyngeal microbiome composition among healthy children in Botswana using shotgun metagenomic sequencing. CLWH in our cohort also had a lower relative abundance of *C. pseudodiphtheriticum* compared to HEU and HUU children. Lastly, though we only had sequencing data for nine serodiscordant sibling pairs, the higher relative abundances of *C. propinquum, C. pseudodiphtheriticum* and *D. pigrum* observed in the HEU siblings of CLWH and the higher abundance of pneumococcus in CLWH further support a potential causal association between HIV and these species given the high likelihood of similar environmental exposures within the same household. As members of the *Corynebacterium* and *Dolosigranulum* genera are associated with protection from respiratory infections in children, their lower relative abundance in CLWH may play a role in the increased incidence of pneumonia seen in this population.18

*D. pigrum* and multiple *Corynebacterium* species are common inhabitants of the upper respiratory tract and can influence colonization resistance to bacterial pathogens. *C. propinquum* inhibits growth of *S. pneumoniae* and some staphylococcal species *in vitro*, while *C. pseudodiphtheriticum* increases resistance to pneumococcal infection in mouse models and was successfully used for *S. aureus* eradication via nasal instillation in adults.19-22 *D. pigrum* also inhibits *in vitro* growth of *S. aureus*, and cocultivation of *C. pseudodiphtheriticum* and *D. pigrum* inhibits pneumococcal growth.23 *S. pneumoniae* and *S. aureus* are responsible for significant childhood morbidity and mortality, and their ability to cause invasive disease is predicated on successful nasopharyngeal colonization.1-3 Among children in our study, the relative abundances of *C. propinquum, C. pseudodiphtheriticum* and *D. pigrum* were negatively correlated with the relative abundances of *S. pneumoniae* and *S. aureus*. Within the context of the existing literature, our findings suggest that the lower nasopharyngeal abundance of *C. pseudodiphtheriticum* in CLWH and the lower relative abundances of *C. propinquum* and *C. pseudodiphtheriticum* in children with HIV-associated immunosuppression may contribute to the increased risk of invasive pneumococcal and *S. aureus* infections in this population.

Several factors likely contribute to the alterations of the nasopharyngeal microbiome observed among CLWH, including immune dysregulation, medication effects, and viremia. First, HIV is associated with altered immunity and several prior studies suggest that HIV-associated immunosuppression influences microbiome composition. Among people with HIV, CD4+ cell counts are associated with the relative abundances of *Streptococcus* and *Lactobacillus* in the salivary microbiome, while low CD4+ cell counts are associated with reduced alpha diversity and higher relative abundances of genera within the *Enterobacteriaceae* family in the gut microbiome.24,25 Similarly, low CD4+% was associated with alterations in nasopharyngeal microbiome composition in our cohort and CLWH and normal CD4+% did not exhibit significant compositional differences compared to HEU and HUU children. Second, medications like ART may impact nasopharyngeal microbiome composition, as they are associated with alterations of the gut microbiome in CLWH.8,26 CLWH are also frequently exposed to antibiotics for both prophylaxis against opportunistic pathogens and treatment of acute infections, which can affect nasopharyngeal microbiome composition.27 In our cohort, receipt of TMP-

SMX prophylaxis was associated with a higher relative abundance of *S. aureus*. Finally, while associations exist between HIV RNA measurements and composition of other microbial communities, viral suppression was not associated with nasopharyngeal microbiome composition in our cohort.24 Transient elevations in HIV RNA measurements can occur in children on ART and this may have been responsible for the detectable HIV RNA measurements among the children in our study, though the cross-sectional study design precludes us from assessing the duration of HIV viremia.28

Our study is one of few utilizing shotgun metagenomic sequencing to characterize the nasopharyngeal microbiome during childhood.29,30 Use of shotgun metagenomic sequencing to characterize respiratory microbial communities presents unique challenges, due in large part to the low microbial biomass in the respiratory tract. Over 99% of the reads generated by this project were human in origin. However, given the sequencing depth we achieved, we were still able to generate sufficient bacterial reads for analyses of microbiome diversity and composition from 93% of samples. We encountered additional challenges with species identification. The amount of publicly available microbiome sequencing data from the gut outstrips sequencing data from other body sites, which is reflected in the National Center for Biotechnology Information (NCBI) sequencing repository and in the marker gene databases developed using NCBI data. Additionally, the reliance on exact matching of marker genes with relatively few microbial reads identified only a median (IQR) of 4 (3, 6) species per sample using MetaPhlAn3, which led us to develop a tailored bioinformatics pipeline that used Kaiju for read classification. Our most abundant species were consistent with genus-level data from prior cohorts in Botswana, which strengthens our confidence in the validity of our results.10,15

Our study has several limitations. First, this study was cross-sectional, and we were thus unable to determine if the differences we observed in nasopharyngeal microbiome composition were associated with increased risk of subsequent respiratory infections in our study population. Additionally, the potential for residual confounding from unmeasured variables exists. For example, we lacked data on delivery mode, can affect upper respiratory microbiome composition and pneumonia risk in children. Enrollment in this study was restricted to southeastern Botswana, and our results may not be generalizable to other settings, especially other low- and middle-income countries where ART is not universally available for CLWH and environmental and sociodemographic differences exist. Our small sample size also limited our ability to adjust for potential confounders in analyses restricted to CLWH and sibling pairs. Lastly, our analyses focused on sample relative abundances, and it is unclear if the associations that we observed reflect differences in the absolute abundances of bacterial species in the nasopharynxes of children.

In conclusion, HIV infection and HIV-associated immunosuppression were associated with alterations of the nasopharyngeal microbiomes of children in Botswana, including lower relative abundances of several putatively beneficial species whose abundances were negatively correlated with the abundances of the common pathogens *S. pneumoniae* and *S. aureus*. Given the high burden of invasive bacterial infections among CLWH, there is an urgent need to develop strategies to modify the upper respiratory microbiome for the reduction or exclusion of potential pathogens and the prevention of respiratory infections in this population.

## Supporting information

Supplemental Material [[supplements/283664_file07.docx]](pending:yes)

## Data Availability

All microbial genetic sequences generated as part of this study will be uploaded to the National Center for Biotechnology Information. The R code used for analysis is available for review at [https://github.com/swetamp/HIV\_mb](https://github.com/swetamp/HIV_mb). Deidentified participant data can be made available via a shared GitHub repository after investigator approval of a proposal.

[https://github.com/swetamp/HIV\_mb](https://github.com/swetamp/HIV_mb) 

## CONTRIBUTORS

SMP and MSK designed the study and statistical methods. JF, SB, LT, and SL recruited participants, collected data, and completed initial specimen processing. MSK performed bioinformatics processing of the sequencing data. SMP led the data analysis, with support from MSK and LAD, and prepared the first draft of the paper. MZP, MM, APS, TAM, KAF, SSS, CKC, NT, and DM provided critical review of study design and implementation, data analysis, and manuscript preparation. SMP and MSK had access to the raw data. All authors provided feedback on the analysis and the manuscript and have read and approved the final version. SMP made the final decision to submit for publication and holds final responsibility for the content.

## DATA SHARING

All microbial genetic sequences generated as part of this study will be uploaded to the National Center for Biotechnology Information. The R code used for analysis is available for review at [https://github.com/swetamp/HIV\_mb](https://github.com/swetamp/HIV_mb). Deidentified participant data can be made available via a shared GitHub repository after investigator approval of a proposal.

## DECLARATION OF INTERESTS

KAF is an employee of Merck Research Laboratories, Merck & Co., Inc. No other authors have any conflicts of interest to report.

## FUNDING

SMP was supported by an Institutional Training Grant in Pulmonary and Critical Care Medicine (T32 HL 007538) and a VECD Global Health Fellowship, funded by the NIH Office of AIDS Research and Fogarty International Center (D43 TW009337). SMP, MSK, and CKC received financial support from the Duke University Center for AIDS Research (CFAR), an NIH funded program (5P30-AI064518). MSK was supported by a NIH Career Development Award (K23-AI135090). APS and TAM received financial support from the NIH through the Penn Center for AIDS Research (P30-AI045008).

## ACKNOWLEDGMENTS

We thank Copan Italia (Brescia, Italy) for the donation of the MSwab media and flocked swabs used in the collection of nasopharyngeal specimens and the Duke University School of Medicine for the use of the Microbiome Core Facility, which performed the DNA extractions and metagenomic sequencing for this research. Additionally, we thank the Botswana Ministry of Health, Gaborone District Health Management Team, and clinic matrons for their support. Lastly, we offer our sincere gratitude to the children and families who participated in this research.

*   Received December 19, 2022.
*   Revision received December 19, 2022.
*   Accepted December 20, 2022.


*   © 2022, Posted by Cold Spring Harbor Laboratory

This pre-print is available under a Creative Commons License (Attribution-NonCommercial-NoDerivs 4.0 International), CC BY-NC-ND 4.0, as described at [http://creativecommons.org/licenses/by-nc-nd/4.0/](http://creativecommons.org/licenses/by-nc-nd/4.0/)

## References

1.  1.Benet T, Sanchez Picot V, Messaoudi M, et al. Microorganisms Associated With Pneumonia in Children <5 Years of Age in Developing and Emerging Countries: The GABRIEL Pneumonia Multicenter, Prospective, Case-Control Study. Clin Infect Dis 2017; 65(4): 604–12.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/cid/cix378&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 

2.  2.Collaborators GBDLRI. Estimates of the global, regional, and national morbidity, mortality, and aetiologies of lower respiratory infections in 195 countries, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Infect Dis 2018; 18(11): 1191–210.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/s1473-3099(18)30310-4&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 

3.  3.Bogaert D, De Groot R, Hermans PW. Streptococcus pneumoniae colonisation: the key to pneumococcal disease. Lancet Infect Dis 2004; 4(3): 144–54.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S1473-3099(04)00938-7&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=14998500&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000189356700018&link_type=ISI) 

4.  4.Bomar L, Brugger SD, Yost BH, Davies SS, Lemon KP. Corynebacterium accolens Releases Antipneumococcal Free Fatty Acids from Human Nostril and Skin Surface Triacylglycerols. mBio 2016; 7(1): e01725–15.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1128/mBio.01725-15&link_type=DOI) 

5.  5.Madhi SA, Adrian P, Kuwanda L, Cutland C, Albrich WC, Klugman KP. Long-term effect of pneumococcal conjugate vaccine on nasopharyngeal colonization by Streptococcus pneumoniae-- and associated interactions with Staphylococcus aureus and Haemophilus influenzae colonization--in HIV-Infected and HIV-uninfected children. J Infect Dis 2007; 196(11): 1662–6.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1086/522164&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18008250&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000250965900013&link_type=ISI) 

6.  6.McNally LM, Jeena PM, Gajee K, et al. Lack of association between the nasopharyngeal carriage of Streptococcus pneumoniae and Staphylococcus aureus in HIV-1-infected South African children. J Infect Dis 2006; 194(3): 385–90.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1086/505076&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16826488&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000239120900017&link_type=ISI) 

7.  7.Jones CE, Naidoo S, De Beer C, Esser M, Kampmann B, Hesseling AC. Maternal HIV infection and antibody responses against vaccine-preventable diseases in uninfected infants. JAMA 2011; 305(6): 576–84.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/jama.2011.100&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=21304083&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000287080500022&link_type=ISI) 

8.  8.Abange WB, Martin C, Nanfack AJ, et al. Alteration of the gut fecal microbiome in children living with HIV on antiretroviral therapy in Yaounde, Cameroon. Sci Rep 2021; 11(1): 7666.
    
    

9.  9.Starr JR, Huang Y, Lee KH, et al. Oral microbiota in youth with perinatally acquired HIV infection. Microbiome 2018; 6(1): 100.
    
    

10. 10.Kelly MS, Surette MG, Smieja M, et al. The Nasopharyngeal Microbiota of Children With Respiratory Infections in Botswana. Pediatr Infect Dis J 2017; 36(9): e211–e8.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1097/INF.0000000000001607&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28399056&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 

11. 11.Kapoor A, Mussa A, Diseko M, et al. Cross-sectional trends in HIV prevalence among pregnant women in Botswana: an opportunity for PrEP? J Int AIDS Soc 2022; 25(3): e25892.
    
    

12. 12.Botswana: WHO and UNICEF estimates of immunization coverage: 2020 revision, 2021.
    
    

13. 13.Mallick H, Rahnavard A, McIver LJ, et al. Multivariable Association Discovery in Population-scale Meta-omics Studies. bioRxiv 2021: 2021.01.20.427420.
    
    

14. 14.Cao Y, Dong Q, Wang D, Zhang P, Liu Y, Niu C. microbiomeMarker: an R/Bioconductor package for microbiome marker identification and visualization. Bioinformatics 2022.
    
    

15. 15.Kelly MS, Plunkett C, Yu Y, et al. Non-diphtheriae Corynebacterium species are associated with decreased risk of pneumococcal colonization during infancy. ISME J 2022; 16(3): 655–65.
    
    

16. 16.Salter SJ, Turner C, Watthanaworawit W, et al. A longitudinal study of the infant nasopharyngeal microbiota: The effects of age, illness and antibiotic use in a cohort of South East Asian children. PLoS Negl Trop Dis 2017; 11(10): e0005975.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1371/journal.pntd.0005975&link_type=DOI) 

17. 17.Vanker A, Nduru PM, Barnett W, et al. Indoor air pollution and tobacco smoke exposure: impact on nasopharyngeal bacterial carriage in mothers and infants in an African birth cohort study. ERJ Open Res 2019; 5(1).
    
    

18. 18.Biesbroek G, Tsivtsivadze E, Sanders EA, et al. Early respiratory microbiota composition determines bacterial succession patterns and respiratory health in children. Am J Respir Crit Care Med 2014; 190(11): 1283–92.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1164/rccm.201407-1240OC&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=25329446&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 

19. 19.Kiryukhina NV, Melnikov VG, Suvorov AV, Morozova YA, Ilyin VK. Use of Corynebacterium pseudodiphtheriticum for elimination of Staphylococcus aureus from the nasal cavity in volunteers exposed to abnormal microclimate and altered gaseous environment. Probiotics Antimicrob Proteins 2013; 5(4): 233–8.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1007/s12602&ndash;013&ndash;9147&ndash;x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26783069&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 

20. 20.Yan M, Pamp SJ, Fukuyama J, et al. Nasal microenvironments and interspecific interactions influence nasal microbiota complexity and S. aureus carriage. Cell Host Microbe 2013; 14(6): 631–40.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.chom.2013.11.005&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24331461&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 

21. 21.Stubbendieck RM, May DS, Chevrette MG, et al. Competition among Nasal Bacteria Suggests a Role for Siderophore-Mediated Interactions in Shaping the Human Nasal Microbiota. Appl Environ Microbiol 2019; 85(10).
    
    

22. 22.Xu L, Earl J, Pichichero ME. Nasopharyngeal microbiome composition associated with Streptococcus pneumoniae colonization suggests a protective role of Corynebacterium in young children. PLoS One 2021; 16(9): e0257207.
    
    

23. 23.Brugger SD, Eslami SM, Pettigrew MM, et al. Dolosigranulum pigrum Cooperation and Competition in Human Nasal Microbiota. mSphere 2020; 5(5).
    
    

24. 24.Lewy T, Hong BY, Weiser B, et al. Oral Microbiome in HIV-Infected Women: Shifts in the Abundance of Pathogenic and Beneficial Bacteria Are Associated with Aging, HIV Load, CD4 Count, and Antiretroviral Therapy. AIDS Res Hum Retroviruses 2019; 35(3): 276–86.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1089/AID.2017.0200&link_type=DOI) 

25. 25.Monaco CL, Gootenberg DB, Zhao G, et al. Altered Virome and Bacterial Microbiome in Human Immunodeficiency Virus-Associated Acquired Immunodeficiency Syndrome. Cell Host Microbe 2016; 19(3): 311–22.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.chom.2016.02.011&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26962942&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 

26. 26.Nguyen QT, Ishizaki A, Bi X, et al. Alterations in children’s sub-dominant gut microbiota by HIV infection and anti-retroviral therapy. PLoS One 2021; 16(10): e0258226.
    
    

27. 27.Henares D, Rocafort M, Brotons P, et al. Rapid Increase of Oral Bacteria in Nasopharyngeal Microbiota After Antibiotic Treatment in Children With Invasive Pneumococcal Disease. Front Cell Infect Microbiol 2021; 11: 744727.
    
    

28. 28.Lee KJ, Shingadia D, Pillay D, et al. Transient viral load increases in HIV-infected children in the U.K. and Ireland: what do they mean? Antivir Ther 2007; 12(6): 949–56.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17926649&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2022%2F12%2F20%2F2022.12.19.22283664.atom) 

29. 29.Binia A, Siegwald L, Sultana S, et al. The Influence of FUT2 and FUT3 Polymorphisms and Nasopharyngeal Microbiome on Respiratory Infections in Breastfed Bangladeshi Infants from the Microbiota and Health Study. mSphere 2021; 6(6): e0068621.
    
    

30. 30.Manenzhe RI, Dube FS, Wright M, et al. Longitudinal changes in the nasopharyngeal resistome of South African infants using shotgun metagenomic sequencing. PLoS One 2020; 15(4): e0231887.