Detection and quantification of residual low-level HIV viremia in plasma based on extracellular particle co-isolation

To the Editor: Current combined antiretroviral therapy (cART) can effectively suppress the plasma viral load in most people living with human immunodeficiency virus-1 (HIV, PLWH). However, approximately 20–30% of PLWH undergoing cART can have very low but detectable plasma viral loads (<20 copies/mL, the limit of quantification in commercially available assays), known as residual HIV-1 low-level viremia (LLV).[1] Residual HIV-1 LLV detection can reflect the source and dynamics of the HIV-1 reservoir in PLWH, thus playing an important role in both HIV-1 management and basic research.[2] Despite real-time quantitative polymerase chain reaction (RT-qPCR) methods like the COBAS AmpliPrep COBAS TaqMan HIV-1 Test version 2.0 (CAP/CTM v2.0) (Roche Diagnostics, Almere, Nederland) achieving a limit of detection (LOD) of 20 copies/mL, they fail to meet the requirements for detecting extremely low levels of viremia.[3] HIV-1 viral particles, which share size and structural similarities with extracellular particles (EPs), can be co-isolated with EPs using methods such as ultracentrifugation or size exclusion chromatography. However, while improving the LOD of plasma HIV-1 RNA, these methods are hindered by significant drawbacks like low EPs recovery, large sample volume, and high labor intensity, limiting their clinical use. In this study, we attempted to address this question using an easy-to-operate medical device developed by us to co-isolate HIV-1 RNA from EPs (EP-HIV RNA) for detecting residual HIV-1 LLV. This retrospective, single-center study was conducted at Guangzhou Eighth People’s Hospital from January 2020 to August 2021 and was approved by the Ethics Committee of Guangzhou Eighth People’s Hospital. We consecutively and randomly recruited 398 HIV-1 patients from the outpatient department, each with recent follow-up data provided in Supplementary File 1, https://links.lww.com/CM9/C174. All participants signed an informed consent form. The enrichment of EPs was performed by a medical device for the clinical isolation of exosomes (medical device filing certificate number: Guangdong Guangzhou machinery No. 20170482). The device had a 0.22 μm upper chamber to keep particles bigger than ~200 nm in diameter from flow-through, while particles with similar size to exosomes and HIV-1 were captured in the middle chamber [Supplementary Figure 1, https://links.lww.com/CM9/C173]. The size, distribution of EPs, and particle concentrations of EPs were measured using a NanoFCM nano flow cytometer (Xiamen Flow Biotechnology, Xiamen, China). HIV-1 DNA and cell-associated RNA (CA-RNA) were quantified using HIV-1 DNA Quantitative Assay Kit (Supbio, Guangzhou, China) and HIV-1 Cell-Associated RNA Quantitative Assay Kits (Supbio), respectively, following the manufacturer’s protocols. The HIV-1 DNA genotypic drug resistance mutation (DRM) assay was conducted as previously described.[4] The protease/reverse transcriptase (PI/RT) and integrase (IN) regions were amplified from HIV DNA and sequenced using Sanger sequencing. Results were analyzed with the HIVdb program (Stanford University HIV Drug Resistance Database) following the International AIDS Society–USA Consensus Guidelines. Drug resistance was scored as follows: 0–9 points (sensitive), 10–14 points (potential resistance), 15–29 points (low resistance), 30–59 points (moderate resistance), and ≥60 points (high resistance). Data analysis was performed using GraphPad Prism software version 9.0 (GraphPad Software, San Diego, USA). Continuous variables are shown as medians and interquartile ranges (IQR), whereas percentages, sex, initial ART, and other factors are indicated as categorical variables. Intergroup comparisons were performed using the Student’s t-test. P <0.05 was deemed a significant difference. To evaluate the efficiency of HIV-1 co-isolation, we analyzed the HIV-1 RNA load of the filtrates in the middle chamber and the flow-through in the lower chamber using plasma samples acquired from three donors. The results showed that the number of EPs enriched in the middle chamber was 4–5 orders of magnitude higher than that in the lower filtrate, with nearly 99% of the HIV-1 particles being enriched in the middle chamber [Supplementary Table 1, https://links.lww.com/CM9/C175]. These results suggest that HIV-1 particles can be efficiently co-isolated with EPs. To evaluate the linear amplification range of the EP-HIV RNA method, we prepared concentration gradients of 1E7, 1E6, 1E5, 1E4, 1E3, 1E2, and 20 IU/mL of HIV-1 pseudoviruses. Linear regression analysis showed that the measured and theoretical concentrations were significantly co-related 20 at 1E7 IU/mL [Supplementary Figure 2, https://links.lww.com/CM9/C173]. Using repeated tests (10–40 times) on standard samples at concentrations of 1E5, 1E4, 1E3, 1E2, 20 IU/mL, and 10 IU/mL, we determined that the EP-HIV RNA method had an LOD of 20 IU/mL, equivalent to 12 copies/mL [Supplementary Table 2, https://links.lww.com/CM9/C176]. Furthermore, the stability, precision [Supplementary Table 3, https://links.lww.com/CM9/C177], and specificity (tested negative for Hepatitis B virus, Mycobacterium tuberculosis, influenza A virus, hepatitis A virus, hepatitis C virus, human cytomegalovirus, and Epstein–Barr virus) of the HIV-1 RNA quantification kit were largely unaffected by the co-isolation of HIV-1 RNA with EPs. Next, we compared the performance of the EP-HIV RNA method with that of the widely used CAP/CTM v2.0, using plasma samples obtained from a cross-sectional cohort of 398 PLWHs. The results showed that the CAP/CTM v2.0 test was positive in 106 cases and negative in 292 cases, with a detection rate of 26.63%. However, the EP-HIV RNA method was positive in 261 cases and negative in 137 cases, with a detection rate of 65.57%. Among the 292 cases that tested negative with the CAP/CTM v2.0, 172 were positive with the EP-HIV RNA method. In contrast, among the 137 cases that tested negative using the EP-HIV RNA method, only 17 were detected as positive using the CAP/CTM v2.0 method [Supplementary Table 4, https://links.lww.com/CM9/C178]. Among the 398 samples, 77 test results were within the quantitative analysis range (20–1E7 IU/mL) for both methods. For these 77 samples, the EP-HIV RNA and CAP/CTM v2.0 test results showed a significant and robust linear correlation (R2 = 0.834, P <0.0001) [Supplementary Figure 3, https://links.lww.com/CM9/C173]. Therefore, we obtained samples that tested negative by CAP/CTM v2.0 but were positive for EP-HIV RNA, which are consistent with the characteristics of residual LLV. As more LLV could be detected with the EP-HIV RNA method than with CAP/CTM v2.0, we posited that the transcriptional activity and size of HIV-1 reservoirs had distinct features in subjects with residual LLV detected by the EP-HIV RNA method only. Fortunately, we observed significantly higher levels of CA-HIV RNA in CAP/CTM v2.0 test-negative and EP-HIV RNA test-positive samples, as compared with samples that tested negative by both methods (P <0.001) [Figure 1A]. Additionally, among all samples that were detected as negative using the CAP/CTM v2.0 method, EP-HIV RNA-positive samples had significantly higher levels of HIV-1 DNA in peripheral blood mononuclear cells (PBMCs) than EP-HIV RNA-negative samples (P <0.001) [Figure 1B].Figure 1: Comparison of HIV-1 CA-RNA (A) and HIV-1 DNA (B) quantification between EP-HIV-RNA-negative and -positive samples (n = 172 vs. n = 120). P-values were calculated using an unpaired Student’s t-test. * P <0.001. CA-RNA: Cell-associated RNA; EP: Extracellular particle; EP-HIV RNA: HIV-1 RNA from EP; PBMCs: Peripheral blood mononuclear cells.The presence of drug-resistant viral clones is a significant cause of unsuppressed viremia (>200 copies/mL).[5] To address the prevalent DRMs at the HIV-1 DNA level, we performed Sanger sequencing on PBMCs of EP-HIV RNA test-positive but CAP/CTM v2.0 test-negative samples. Considering the successful amplification of both the PI/RT and IN regions, the success rates were 85% (34/40) and 67.5% (27/40) in the EP-HIV RNA-positive and -negative samples, respectively [Supplementary Table 5, https://links.lww.com/CM9/C179]. For either EP-HIV RNA-positive or -negative samples, 5% failed to amplify either the PI/RT or IN region [Supplementary Table 5, https://links.lww.com/CM9/C179]. The proportion of successful amplifications in the PI/RT regions was over 80% in EP-HIV RNA-positive and -negative samples [Supplementary Table 5, https://links.lww.com/CM9/C179]. We observed 15 and 11 DRMs at the HIV-1 DNA level in EP-HIV RNA-positive and -negative samples, respectively [Supplementary Table 6, https://links.lww.com/CM9/C180]. DRMs found in EP-HIV RNA-negative samples were classified as potentially resistant or low-level resistant. In contrast, EP-HIV RNA-positive samples had two highly correlated resistance sites, with resistance scores ranging from 190 to 240 [Supplementary Table 6, https://links.lww.com/CM9/C180]. In this study, we report the feasibility of co-isolating HIV-1 particles with EPs to increase the sensitivity of LLV measurements. The EP-HIV RNA detection method meets the clinical demands for the precise evaluation of cART and adherence. This method involves only one additional step of EP co-isolation with a well-characterized medical device prior to regular HIV-1 RNA load tests, which can be easily performed by under-trained personnel in clinical laboratories. Based on the results of this study, the EP-HIV RNA method did not change the stability, precision, or specificity but significantly increased the sensitivity of the HIV-1 RNA test. It is known that plasma contains PCR-interfering ingredients, whereas EP co-isolation may reduce such ingredients, which may be a possible reason for increased sensitivity. The detection threshold of the EP-HIV RNA method was slightly superior to that of commonly used commercially available assays such as CAP/CTM v2.0. In the clinical cohort validation, we found that the detection rate of commercial kits was only 26.63% (106/398), whereas our method achieved a detection rate of 65.57% (261/398). Moreover, among the 292 samples not detected by commercial kits, 58.90% (172/292) were detected using our method. This demonstrates that the sensitivity of the EP-HIV RNA method is significantly higher than that of commonly used commercial detection methods. Furthermore, we found that EP-HIV RNA-positive samples had a significantly higher HIV-1 reservoir size and HIV-1 DNA transcriptional activity. The relationship between residual LLV and HIV reservoirs has also been reported.[6] These residual low-level viruses often exhibit high clonality as detected by integration site analysis or env sequencing of plasma viruses. This suggests that these residual viruses did not originate from new rounds of infection and did not evolve. This may be due to the lower concentrations of antiretroviral drugs in tissues and reduced HIV-specific CD8+ T cell responses.[6] In addition, HIV-1 DNA DRM analysis indicated that patients with residual LLV were at a higher risk of drug resistance. Whether this residual LLV progresses to the risk of uncontrolled viremia remains to be verified. However, our results also remind us to pay attention to this subset of patients to maintain continuous and effective viral control. This study has certain limitations. Most proviruses in the HIV-1 reservoir are defective (95%), with only a fraction of intact proviruses capable of producing replicable viruses. The assays performed for HIV-1 DNA testing are based on total proviral DNA and do not distinguish intact proviral DNA. The virus outgrowth assay (VOA) is a classical method used to assess the size of HIV-1 reservoirs. The standard HIV-1 drug resistance test is usually performed using viral RNA in plasma. However, a low viral load often results in a high failure rate during amplification. Hence, the HIV-1 proviral DNA has been used in HIV-1 drug resistance tests. In conclusion, the detection of residual LLV is clinically significant. In this study, we introduced an ultrafiltration-based EP enrichment method to enhance the sensitivity and efficiency of plasma HIV-1 detection and demonstrated the practicality and reliability of this method. Numerous patients with residual viral loads caused by ongoing transcription from HIV-1 reservoirs have been identified using this method. Furthermore, apart from plasma HIV-1 particles, the EP enrichment method provides novel directions for the sensitive detection of other pathogenic particles, such as hepatitis B virus and hepatitis C viru. Funding This work was supported by grants from the Natural Science Foundation of China (No. 82072265), the National Natural Science Foundation of China (No. 82302513), Guangdong Basic and Applied Basic Research Foundation (Nos. 2022A1515110546, 2024A1515013281), and Science and Technology Project of Guangzhou (No. 2023A03J0801), the Science and Technology Project of Guangzhou (No. 202201020276), and the Fund of Innovation and Entrepreneurship Leading Team Project of Guangzhou (No. 201809010009).