We have developed four in-house CA-p24 ELISAs using commercial antibodies to identify diverse HIV-1 isolates and detect HIV infections in vitro. These protocols were named based on the commercial source of the antibodies: Aalto Bio Reagents (ABR), Anogen (ANG), Sino Biological (SB), and R&D Systems (RND). The costs per each in-house assay range between €25 and €30 per one 96-well plate, which is 15 × lower than medium-cost commercial kits, 25 × lower than high-cost commercial kits (q = 0.0065, Kruskal–Wallis’ test), and 45 × lower than very high-cost commercial kits (q = 0.0048) (Additional file 1: Fig. S1 and Additional file 2: Table S1).
To optimize the CA-p24 ELISA protocols, we initially conducted an assessment of our in-house reagents, including PBS buffers and LumiPhos substrates (see Table 1). We also evaluated the performance of our equipment, specifically the BioTek ELx405 Select washer and the GloMax® plate reader. During this process, we determined the ideal working volume for each assay, with most assays showing optimal results using either 25 µL or 50 µL (as indicated in Table 1). Note we have utilized a standardized CA-p24 antigen across all assays to ensure consistency (Tables 3, 4, 5, 6). The working concentrations of antibodies, incubation periods, washing buffers, and blocking buffers were all employed according to the recommendations provided by the manufacturers (see “Methods” and Table 1). Lastly, we assessed the linear range of each assay, resulting in three assays with an upper limit of 5 ng/mL (ABR, ANG, and SB) and one assay with a limit of 3 ng/mL (RND) (Table 1).
Here we described the ELISA protocols optimized in-house (Fig. 1, Table 1) and the in vitro validations performed to determine the sensitivity, specificity, reactivity, and reproducibility of these assays towards distinct HIV-1 isolates.
CA-p24 ELISA protocols
Half-area white 96-well plates were pre-coated with capture antibody and incubated at room temperature overnight (Fig. 1, Table 1, see “Methods”). Next, plates were washed three times with wash buffer and blocked with blocking buffer. The ABR-protocol does not have an initial blocking step in our protocol and requires a TBS-based washing buffer for this step. After 1 h incubation at room temperature while shaking, the plates were washed. Diluted samples, CA-p24 standard proteins and controls (PBS) were then added to the wells, and the plates were incubated for 2 h at room temperature while shaking. Thereafter, plates were washed, and the alkaline phosphatase (AP)- or horseradish peroxidase (HRP)-conjugated detection antibody (ABR-, ANG-, and SB-protocols) or an unconjugated detection antibody (RND-protocol) was added to the wells. ABR-detection antibody solution with skim milk was used as a blocking step. Following 1–2 h incubation at room temperature while shaking, the plates were washed (Table 1). Note, the ABR protocol requires a PBS-based wash buffer for this step while the RND-protocol needs conjugation with Streptavidin-HRP, followed by incubation at room temperature for 20 min while shaking. Each conjugate (AP or HRP) catalyzed an enzymatic reaction by the addition of LumiPhos Plus or LumiPhos-HRP to the wells. After brief incubation (Table 1) with the LumiPhos solution, the luminescence generated was immediately read on GloMax® plate reader (Fig. 1, see “Methods”). Standard curves were created, and data analyses were performed. One set of standard curves obtained through all tested ELISAs are shown in Additional file 1: Fig. S2. HIV isolates (Table 2) were measured in duplicate by two independent ELISA runs per ELISA protocol.
Detection of HIV-1 isolates
We have previously used the ABR-CA-p24 ELISA to measure the CA-p24 concentration in HIV infection experiments for over 20 years and therefore considered this assay the standard reference [19,20,21,22]. When testing the 22 HIV isolates (Table 2), we observed that the ABR-, ANG-, and SB-CA-p24 ELISAs detected all isolates and only RND-CA-p24 ELISA was unable to detect the isolate 92UG029 (Fig. 2). Overall, higher CA-p24 values were detected by ABR-CA-p24 ELISA in comparison to the other CA-p24 ELISAs. When analyzing the CA-p24 concentrations by the HIV subtype (Table 2), ANG- and RND-CA-p24 ELISAs exhibited low reactivity for detecting non-B subtype isolates (Fig. 2).
Reactivity of ELISA systems towards HIV-1 A and B subtypes
We also observed that the measured CA-p24 concentrations of several HIV-1 isolates differed with each ELISA system (Fig. 2) and therefore calculated the CA-p24 fold-change of ANG-, SB-, and RND-CA-p24 ELISAs relative to the concentrations obtained by ABR-CA-p24 ELISA to determine their reactivity towards each isolate (Fig. 3). We focused the analyses on HIV-1 A and B subtypes because all systems performed well on these isolates (Fig. 2). Compared to ABR-CA-p24 ELISA, we observed that ANG-CA-p24 ELISA showed the lowest reactivity towards the isolates, detecting 10 out of 16 isolates (62.5%) with more than a fivefold reduction in CA-p24 concentration (Fig. 3). Overall, RND-CA-p24 ELISA exhibited a better reactivity towards most isolates when compared to ANG-CA-p24 ELISA, however, the system still showed poor reactivity when compared to ABR-CA-p24 ELISA by detecting 6 out of 16 isolates (37.5%) with more than a fivefold reduction in CA-p24 concentration (Fig. 3). RND-CA-p24 ELISA also detected the isolate SI22 with more than a 25-fold reduction in CA-p24 concentration when compared to ABR-CA-p24 ELISA, suggesting a very low reactivity of this system towards this isolate. SB-CA-p24 ELISA showed better reactivity towards the isolates when compared to ANG- and RND-CA-p24 ELISA by detecting only 4 out of 16 isolates (25%) with more than a fivefold reduction in CA-p24 concentration (Fig. 3). However, SB-CA-p24 ELISA detected the LAI isolate with more than a 25-fold reduction in CA-p24 concentration when compared to ABR-CA-p24 ELISA, implying very low reactivity of this assay for this particular isolate.
Correlations between ELISA systems
To further test the reactivity of the ANG-, SB-, and RND-CA-p24 ELISAs when compared to our standard ABR-CA-p24 ELISA, we determined the Pearson r correlation of the CA-p24 concentrations of all tested isolates measured by each assay (Fig. 4). Here we observed that ANG-CA-p24 ELISA measurements exhibited a significantly high correlation with ABR-CA-p24 ELISA (r = 0.69 [95% CI 0.40–0.85], p = 0.002), possibly due to the use of the same detection antibody. Similarly, SB-CA-p24 ELISA showed a significantly high correlation with ABR-CA-p24 ELISA (r = 0.64 [95% CI 0.31–0.82], p = 0.0007). RND-CA-p24 ELISA also exhibited a significant high correlation with ABR-CA-p24 ELISA (r = 0.61 [95% CI 0.27–0.81], p = 0.0015) (Fig. 4). We found the highest correlation between RND- and ANG-CA-p24 ELISAs (r = 0.83 [95% CI, 0.65–0.92], p < 0.0001), suggesting a similar sensitivity between both assays, possibly due to their similar reactivity for HIV-1 B subtypes. The lowest correlations were observed between RND- and SB-CA-p24 ELISAs (r = 0.34 [95% CI − 0.07 to 0.65], p = 0.1024) and between ANG- and SB-CA-p24 ELISAs (r = 0.21 [95% CI − 0.20 to 0.57], p = 0.3039) (Fig. 4).
Reproducibility of ELISA systems and influence of viral inactivation
We then assessed the reproducibility of the ELISA systems, examining both their consistency within individual assays and their reliability between different assays. In order to assess within-assay reproducibility, we employed the NL4-3 HIV-1 isolate as a test model. We quantified its concentration in each assay by utilizing technical replicates and applying different inactivation treatments, including heat and/or detergent. We observed that the concentration of the NL4-3 isolate remained consistent between the replicates within the tested conditions (Additional file 1: Fig. S3), and across previous ELISA runs (Fig. 2) indicating that all ELISA systems demonstrate within-assay reproducibility. We observed that most ELISA systems are not affected by the inactivation method, whereas SB and RND protocols showed a significant drop in reactivity when samples were heat-inactivated. However, this effect was not observed when heat inactivation was combined with detergent-based inactivation (Additional file 1: Fig. S3). Additionally, we noted that the SB-CA-p24 ELISA exhibited higher reactivity for NL4-3 HIV-1 compared to other systems, resulting in a higher measured CA-p24 concentration, which is consistent with previous findings (Figs. 2, 3). Since we observed that the reactivity of the assays was mostly isolate-dependent (Figs. 2, 3), we evaluated the between-assay reproducibility using the standard CA-p24 protein (Additional file 1: Fig. S4). By analyzing the concentrations of CA-p24 based on the generated luminescence by assay, we noticed an excellent correlation of all concentrations between assays (Pearson r > 0.97, p < 0.002), suggesting a high between-assay reproducibility for the ELISA systems.