Characterizing genetic transmission networks among newly diagnosed HIV-1 infected individuals in eastern China: 2012-2016

Abstract

We aimed to elucidate the characteristics of HIV molecular epidemiology and identify transmission hubs in eastern China using genetic transmission network and lineage analyses. HIV-TRACE was used to infer putative relationships. Across the range of epidemiologically-plausible genetic distance (GD) thresholds (0.1-2.0%), a sensitivity analysis was performed to determine the optimal threshold, generating the maximum number of transmission clusters and providing reliable resolution without merging different small clusters into a single large cluster. Characteristics of genetically linked individuals were analyzed using logistic regression. Assortativity (shared characteristics) analysis was performed to infer shared attributes between putative partners. 1,993 persons living with HIV-1 were enrolled. The determined GD thresholds within subtypes CRF07_BC, CRF01_AE, and B were 0.5%, 1.2%, and 1.7%, respectively, and 826 of 1,993 (41.4%) sequences were linked with at least one other sequence, forming 188 transmission clusters of 2-80 sequences. Clustering rates for the main subtypes CRF01_AE, CRF07_BC, and B were 50.9% (523/1027), 34.2% (256/749), and 32.1% (25/78), respectively. Median cluster sizes of these subtypes were 2 (2-52, n = 523), 2 (2-80, n = 256), and 3 (2-6, n = 25), respectively. Subtypes in individuals diagnosed and residing in Hangzhou city (OR = 1.423, 95% CI: 1.168-1.734) and men who have sex with men (MSM) were more likely to cluster. Assortativity analysis revealed individuals were more likely to be genetically linked to individuals from the same age group (AIage = 0.090, P<0.001) and the same area of residency in Zhejiang (AIcity = 0.078, P<0.001). Additionally, students living with HIV were more likely to be linked with students than show a random distribution (AI student = 0.740, P<0.01). These results highlight the importance of Hangzhou City in the regional epidemic and show that MSM comprise the population rapidly transmitting HIV in Zhejiang Province. We also provide a molecular epidemiology framework for improving our understanding of HIV transmission dynamics in eastern China.

Fig 1. Phylogenetic trees for CRF01_AE, CRF07_BC and B.

The phylogenetic tree is computed by Mega v 5.05. Neighbor-joining phylogenetic tree constructed based on partial pol genes from individuals who were newly diagnosed between 2012–2016 in Zhejiang. The reference sequences obtained from the Los Alamos HIV database are marked as solid triangle. The CRF01_AE, CRF07_BC, and B reference sequences are labeled with red, blue and green solid upward-pointing triangles (▲). The other reported reference sequences are labeled with a black solid upward-pointing triangle (▲).

Fig 2. Genetic distance thresholds.

(A–C) HIV transmission cluster amount and maximum cluster size with a range of GDs (0.1–2.0%) among subtypes CRF01_AE, CRF07_BC, and B. The red line depicts the maximum cluster size with GDs varying from 0.1% to 2.0%, whereas the blue line represents the number of clusters. The epidemiologically plausible range of thresholds for each subtype is highlighted with dashes.

Fig 3. HIV transmission network.

HIV transmission clustering in Zhejiang by risk group (A) and student status (B). Edges indicate genetic linkage (< optimal GD threshold).

Fig 4. Assortativity analysis among linked individuals.

The random value is 1000 times, and “1” indicates linked individuals originating from the same category, whereas “-1” indicated those from different category. Assortativity analysis by age group, student status, city of residence, marital status, and year of diagnosis, respectively. Assortativity index for above were AI_age = 0.090, AI_city = 0.078, AI_marital = 0.076, AI_student = 0.740, AI_year = 0.140, and AI_risk = 0.185.