Es measured in one system do not directly translate into consistent differences in virus replication capacity in another system, in this case in tissues from various donors [7]. Furthermore, the observed differences in TCID50 of different viruses are much less than the variability that is seen for replication of a given virus stock in tissues from different donors [5,8].determined by staining with a KC57 FITC labeled anti HIV-1 p24 antibody (Beckman Coulter, Miami, FL).Statistical AnalysesAnalyses were conducted using JMP 9.0 (SAS Institute, Cary, NC). Data were analyzed for normality using the Shapiro-Welsh test. When 3 or more groups were compared, we performed an ANOVA with the post-hoc correction of Tukey-kramer Honestly Significant Difference. When data were not normally distributed, we performed a non-parametric multiple comparison with Dunn’s correction for joined ranks. The proportion of successful infection (.100 pg p24) in tissues infected with T/F or C/R viruses were compared using Fishers’ exact test for two group comparisons or the likelihood ratio when successful infection proportions were compared across several groups. In several cases, for the reader’s information, we present both mean 6 SEM and median with IQR. However, in cases of non-normal distribution of the variable, only the medians were used for statistical analysis.ResultsIn an ex vivo cervical tissue system we analyzed biological properties of eight HIV-1 constructs that contained env sequences derived from mucosally transmitted T/F HIV-1 and three constructs that contained envelopes derived from control reference HIV-1 variant (C/R) viruses: NL-SF162.ecto, NL-YU-2.ecto, and NL-BaL.ecto. All env sequences were expressed in otherwise isogenic NL4-3-based backbones [4]. Also, in several experiments we used two full-length T/F viruses, CH077.t and RHPA.c [6]and the laboratory-adapted HIV-1BaL isolate, which we used as the reference. Earlier, we had shown that the HIV-1BaL isolate and the Env-IMC cognate NL-BaL.ecto were similar in cellular tropism and virus replication in various primary target cells ([6] and unpublished]). Cervical tissue blocks were inoculated with virus as described earlier [5] and infection was evaluated by determining the fraction of infected T cells as well as the 
amount of p24 released into the culture medium. Overall we performed experiments with cervical tissues from 37 donors. Each donor tissue was infected with at least one C/R virus and at least one T/F virus. According to our optimized protocol for cervical tissue infection, for any given virus stock, 16 tissue blocks per donor per condition have to be inoculated. The amount of cervical tissue obtained from individual donor did not allow for the infection of tissue from each donor with all the used viruses while keeping the number of replicates dictated 
by the protocol. Therefore, to 15755315 increase the statistical power we pooled data from 58 infections with T/F HIV-1 variants and compared them with pooled data from 39 infections with C/R HIV-1 variants. In some experiments, we also compared the data for one T/F HIV-1 variant, NL-1051.TD12.ecto with the data for the control HIV-1 variant NL-SF162.ecto, but replicating in donor-matched cervical tissues. In order to distinguish de 23115181 novo HIV-1 production from the release of virus or free p24 merely adsorbed at inoculation, we treated infected tissues with the RT inhibitor 3TC. For reliably determining that the infection was productive, based o.Es measured in one system do not directly translate into consistent differences in virus replication capacity in another system, in this case in tissues from various donors [7]. Furthermore, the observed differences in TCID50 of different viruses are much less than the variability that is seen for replication of a given virus stock in tissues from different donors [5,8].determined by staining with a KC57 FITC labeled anti HIV-1 p24 antibody (Beckman Coulter, Miami, FL).Statistical AnalysesAnalyses were conducted using JMP 9.0 (SAS Institute, Cary, NC). Data were analyzed for normality using the Shapiro-Welsh test. When 3 or more groups were compared, we performed an ANOVA with the post-hoc correction of Tukey-kramer Honestly Significant Difference. When data were not normally distributed, we performed a non-parametric multiple comparison with Dunn’s correction for joined ranks. The proportion of successful infection (.100 pg p24) in tissues infected with T/F or C/R viruses were compared using Fishers’ exact test for two group comparisons or the likelihood ratio when successful infection proportions were compared across several groups. In several cases, for the reader’s information, we present both mean 6 SEM and median with IQR. However, in cases of non-normal distribution of the variable, only the medians were used for statistical analysis.ResultsIn an ex vivo cervical tissue system we analyzed biological properties of eight HIV-1 constructs that contained env sequences derived from mucosally transmitted T/F HIV-1 and three constructs that contained envelopes derived from control reference HIV-1 variant (C/R) viruses: NL-SF162.ecto, NL-YU-2.ecto, and NL-BaL.ecto. All env sequences were expressed in otherwise isogenic NL4-3-based backbones [4]. Also, in several experiments we used two full-length T/F viruses, CH077.t and RHPA.c [6]and the laboratory-adapted HIV-1BaL isolate, which we used as the reference. Earlier, we had shown that the HIV-1BaL isolate and the Env-IMC cognate NL-BaL.ecto were similar in cellular tropism and virus replication in various primary target cells ([6] and unpublished]). Cervical tissue blocks were inoculated with virus as described earlier [5] and infection was evaluated by determining the fraction of infected T cells as well as the amount of p24 released into the culture medium. Overall we performed experiments with cervical tissues from 37 donors. Each donor tissue was infected with at least one C/R virus and at least one T/F virus. According to our optimized protocol for cervical tissue infection, for any given virus stock, 16 tissue blocks per donor per condition have to be inoculated. The amount of cervical tissue obtained from individual donor did not allow for the infection of tissue from each donor with all the used viruses while keeping the number of replicates dictated by the protocol. Therefore, to 15755315 increase the statistical power we pooled data from 58 infections with T/F HIV-1 variants and compared them with pooled data from 39 infections with C/R HIV-1 variants. In some experiments, we also compared the data for one T/F HIV-1 variant, NL-1051.TD12.ecto with the data for the control HIV-1 variant NL-SF162.ecto, but replicating in donor-matched cervical tissues. In order to distinguish de 23115181 novo HIV-1 production from the release of virus or free p24 merely adsorbed at inoculation, we treated infected tissues with the RT inhibitor 3TC. For reliably determining that the infection was productive, based o.