Lessons Learned in Developing a Commercial FIV Vaccine: The Immunity Required for an Effective HIV-1 Vaccine.

The feline immunodeficiency virus (FIV) vaccine called Fel-O-Vax® FIV is the first commercial FIV vaccine released worldwide for the use in domestic cats against global FIV subtypes (A–E). This vaccine consists of inactivated dual-subtype (A plus D) FIV-infected cells, whereas its prototype vaccine consists of inactivated dual-subtype whole viruses. Both vaccines in experimental trials conferred moderate-to-substantial protection against heterologous strains from homologous and heterologous subtypes. Importantly, a recent case-control field study of Fel-O-Vax-vaccinated cats with outdoor access and ≥3 years of annual vaccine boost, resulted in a vaccine efficacy of 56% in Australia where subtype-A viruses prevail. Remarkably, this protection rate is far better than the protection rate of 31.2% observed in the best HIV-1 vaccine (RV144) trial. Current review describes the findings from the commercial and prototype vaccine trials and compares their immune correlates of protection. The studies described in this review demonstrate the overarching importance of ant-FIV T-cell immunity more than anti-FIV antibody immunity in affording protection. Thus, future efforts in developing the next generation FIV vaccine and the first effective HIV-1 vaccine should consider incorporating highly conserved protective T-cell epitopes together with the conserved protective B-cell epitopes, but without inducing adverse factors that eliminate efficacy.


Introduction
Feline immunodeficiency virus (FIV) was discovered in the fall of 1986 from a stray cat cattery in northern California [1]. Cats residing in one of the five multi-cat pens were succumbing to immunodeficiency syndromes despite being negative for feline leukemia virus (FeLV), the only immunodeficiency causing virus known at the time [2]. The new virus initially referred as feline T-lymphotropic virus; however, later renamed FIV based on its closeness with HIV [3][4][5]. FIV exhibit similar morphology and genomic organization of HIV-1 and serologically related to lentivirus than the oncogenic retrovirus [5][6][7][8][9]. FIV preferentially infect T cells with the hallmark giant cell cytopathology and encode Mg 2+ -dependent reverse transcriptase, unlike a Mn 2+ -dependent RT of FeLV [1,5]. Additionally, the infected CD4 + T cells die of apoptosis causing the CD4/CD8 T-cell inversion, the hallmark of all three AIDS-causing lentiviruses (HIV-1, FIV, and SIV) [4,[10][11][12].
HIV-1 and FIV cause AIDS in their natural hosts, humans and domestic cats, respectively. Conversely, the simian immunodeficiency virus (SIV) live symbiotically without causing simian AIDS (SAIDS) in its natural host, the African macaques, but causes an acute manifestation of SAIDS in FIV (especially subtype A) with certain hotspots of FIV infection as great as 20% infection, Australia may be an excellent testing site for field efficacy trial of the current and the future FIV vaccines for commercial market [37].
The current review will discuss the following areas: (1) the efficacy results that led to the United States Department of Agriculture (USDA) approval of Fel-O-Vax ® FIV vaccine; (2) the efficacy results comparing the Fel-O-Vax ® FIV vaccine to the prototype dual-subtype FIV vaccine; (3) the immune mechanisms of vaccine protection observed with the commercial and prototype FIV vaccines; and (4) the recommendations to the FIV and HIV vaccine researchers that may contribute towards their effort in generating an effective second-generation FIV vaccine and the first effective HIV-1 vaccine for the global use.

Background on the Vaccine Components of the Commercial and Prototype FIV Vaccines
Both Fel-O-Vax ® and its prototype FIV vaccines contain inactivated two subtypes, FIV Pet (subtype A) and FIV Shi (subtype D) [41]. The commercial vaccine is composed of inactivated FIV Pet -infected whole cells (FIV Pet -IWC) and FIV Shi -IWC in FD-1 adjuvant (2 × 10 7 cells total with about 50 µg of total viruses in the fluid). Whereas, the prototype vaccine comprises of paraformaldehyde-inactivated pelleted whole viruses (IWV) (250 µg of each virus) in the same adjuvant supplemented with either human or feline IL-12 to enhance anti-FIV T-cell immunity. Additionally, the commercial vaccine carries a higher concentration of surface envelope glycoproteins (SU Env, gp100) and/or the varying configurations of SU Env expression compared to the prototype vaccine. Consequently, the cats immunized with commercial vaccine develop higher levels of antibodies against SU Env as shown by the FIV Pet immunoblot analysis of the sera of the vaccinated cats [42]. The commercial vaccine uses FIV Pet -infected T cell clone (FL-6) (former Fort Dodge Animal Health, Division of Wyeth) as a source of FIV Pet viral components, whereas the prototype vaccine utilizes a different T cell clone (FL-4) infected with FIV Pet , keeping the common FIV Shi -infected T-cell line (FIV Shi /FeT-J) in both vaccines [42,43]. FIV Pet virus generated from FL-6 has more intact SU Env than the virus generated from the FL-4 cell line causing the difference in antibody response between the vaccines [42]. The IL-2-independent FL-4 and FL-6 cells are clones derived from the IL-2-dependent FIV Pet -FeT1 cell line, whereas the FIV Shi /FeT-J cell line derived by infecting IL-2-independent FeT-J cells with FIV Shi [44]. Since the FeT-J cell line is a derivative of the IL-2-dependent uninfected FeT1 cell line, all the cell lines in vaccine generation have a common lineage of the FeT-1 cell and do not differ in their major histocompatibility complex (MHC) allotypes. Phenotypically, FL-4 and FL-6 cells are CD3 + CD4 −/± CD8 + and lack MHC-II, whereas FeT-J and FIV Shi /FeT-J cells are CD3 + CD4 − CD8 − with the expression of MHC-II molecule [42]. Although the commercial vaccine has more FIV Env glycoproteins than the prototype vaccine, the prototype FIV vaccine had more protective efficacy than the commercial Fel-O-Vax ® FIV vaccine as described in Section 3 below.
Earlier work demonstrated equivalent protection using FIV Pet -infected cell vaccine (Commercial) and inactivated FIV Pet -IWV vaccine (Prototype) against homologous FIV Pet and heterologous FIV Dixon (subtype A; 11% SU and 4% TM aa sequence difference from those of FIV Pet ) [45,46] and induced cellular immunity including T-cell immunity [46]. Furthermore, inactivated subtype-B FIV M2 -IWC vaccine conferred complete protection of 100% (12/12) in the vaccinated cats compared to the non-vaccinated control cats (7/14) when exposed to field cats over 1.5-1.8 years [47]. These promising findings were based on experimental single-strain-IWC and -IWV vaccines. However, the commercial and prototype FIV vaccines are composed of inactivated dual-subtype A & D FIV strains to broaden the protective immunity and efficacy beyond those afforded by single-strain vaccines. Regrettably, not much is known about these vaccines in regards to the immune correlates of protection, especially in the field conditions.

Comparing the Prophylactic Efficacy of the Commercial and Prototype FIV Vaccines
Prototype vaccine containing dual-subtype IWVs demonstrated slightly better protection against heterologous subtype strains even at moderately high challenge doses than the commercial vaccine (Table 1, Studies 1-5) [21,[47][48][49][50][51][52][53][54]. For example, statistically significant protection against New Zealand isolate FIV NZ1 (subtype F /C) was observed with prototype vaccine (p = 0.0015) but not with the commercial vaccine (p = 0.0952) (Study 5a versus 5b). Notably, the prototype vaccine conferred significant protection when none of the vaccinated cats had detectable neutralizing antibodies (Nabs) against FIV NZ1 but had NAbs to FIV Pet before challenge [48]. Furthermore, the NAb titers to the heterologous challenge viruses (Studies 2, 4, 5) did not correlate with the level of protection with the exception of FIV Bang (Study 3). Vaccine-induced NAbs to only FIV Pet and FIV Bang appeared to correlate with protection observed in the vaccinated cats against the FIV Pet and FIV Bang , respectively (Studies 1, 3), suggesting cellular immunity may be the cause of protection against other heterologous viruses, such as FIV NZ1 . However, it is still possible that cellular immunity, such as T-cell and NK-cell activities, in combination with NAbs may work more effectively than cellular immunity alone.
Although FIV does not transmit through the mucosal route as discussed earlier, prototype IWVvaccinated female cats conferred significant protection (83% protection, p = 0.0047) when challenged intravaginally with homologous FIV Pet , suggesting protection against mucosal challenges could be achieved when a similar strategy used in HIV-1 vaccine development where mucosa is the primary site of transmission (Table 1) [21]. The IWV-induced anti-FIV Pet IgG and IgA were detected in the serum and vaginal tract but the titer of anti-FIV IgA was considerably lower than IgG [55]. As a small animal vaccine model for HIV/AIDS, this study was performed to test whether the vaccine can protect against mucosal transmission, for example sexual interaction which is the most common transmission mode for HIV-1 [36].
The commercial vaccine conferred no protection against systemic IM challenge [50], whereas the prototype vaccine showed some protection against systemic IV challenge with the same dose of FIV UK8 [48] (Table 1). Both studies showed minimal (10 NAb titer) anti-FIV UK8 NAb titer in a small percentage of vaccinated cats. Such low NAb titers induced by commercial vaccine did not protect any Fel-O-Vax ® FIV-vaccinated cats, and only one of the two IWV-vaccinated cats with a titer of 10 was among the six IWV-vaccinated protected cats [48]. Thus, no detectable NAb titer was detected in the remaining five prototype vaccinated/protected cats suggesting that vaccine-induced NAbs are not involved in vaccine protection against FIV UK8 . Unfortunately, anti-FIV UK8 NAb post-challenge in Study 2b was not performed to determine any amnestic NAb levels develop shortly after FIV UK8 challenge, much like the observation made in a prime-boost study with canarypox virus-vectored ALVAC-FIV gag/pro/env prime followed by single Pet-ICV boost [56]. The difference in protection rates may be attributed to the different challenge routes used (IM versus IV) or possibly the unique feature of the FIV UK8 . The two USDA trials used the Fel-O-Vax ® FIV vaccine against IM challenge with subtype-A viruses from U.S. (Table 1, Trial 10) [41,49,52] and Netherland (Trial 11) [41,49], and both trials conferred significant protection. These observations suggested that the commercial vaccine has prophylactic efficacy against subtype-A viruses in the U.S. and the Netherlands but perhaps not against subtype-A viruses in the UK. Since no vaccine can confer protection against excessively high challenge dose, the natural transmission dose must be determined to set appropriate challenge dose in experimental studies. Contact exposure with naturally FIV-infected cats is the ideal setting to evaluate the vaccine efficacy close to those performed in human clinical vaccine trials [33,47,51]. Such studies may take over two years for the sufficient number of infection to develop in the non-vaccinated controls, especially when the FIV prevalence is too low in the area used for natural contact exposure (Table 2) [24,33,47,51,57,58]. In one study, contact exposure of uninfected (non-vaccinated) shelter cats (9 females and 5 males) were performed in a less-restrictive free-roaming shelter facility with FIV prevalence rate of 52% (FIV incidence rate of 17%) ( Table 2, Study 1) [47]. In Japan, experimental contact exposure was used in a more-restrictive closed housing system keeping uninfected and infected males together ( Table 2, Study 2) [51]; additionally, in U.S. uninfected females with naturally infected males were kept together to evaluate natural transmission ( Table 2, Study 3) [33]. Based on these studies, the natural transmission doses were on the average 0.48 mean cat infectious dose (CID 50 ) with a range of 0.29-0.75 CID 50 in one year of contact exposure. As expected, more uninfected cats became infected with extended contact exposure time (Table 2). Interestingly, Study 2 included an additional group of Fel-O-Vax-vaccinated laboratory cats that were all protected in the identical contact exposure to subtype-B FIV Ao2 -infected laboratory cats (Table 1, Study 9) [51]. Such contact exposure study may be more close to the natural transmission but require 1.5 years to achieve enough infection in the control cats to determine the protection rate.
As previously discussed, hotspots of FIV prevalence (>20%) in Western Australia such as in Perth [37] may provide an ideal condition for contact exposure studies against subtype-A FIV viruses. Such a study was recently performed in Australia using client-owned cats with outdoor access from five Australian states/territories where Fel-O-Vax ® FIV vaccine has been available since 2004. The case-control study of 139 vaccinated cats receiving ≥3 years of annual boost and 212 non-vaccinated control cats showed protection rate of 56% upon the field exposure [54]. This protection rate was low compared to the laboratory efficacy trial performed for only 1.5 years with one annual boost (Table 1, Study 9). However, this rate observed in vaccinated cats with annual boosts was far better than the 31.2% protection rate observed in the best human HIV-1 vaccine (RV144) trial [59]. Nonetheless, to improve this efficacy of FIV vaccines beyond 56% protection, the protective immunity generated by the commercial and prototype dual-subtype FIV vaccines must be analyzed carefully towards refining and improving the commercial vaccine. To this end, the following sections attempt to determine the protective immunity and the adverse factors that can eliminate vaccine efficacy.

The Role of Vaccine-Induced Antibody (Ab) Immunity in Prophylactic Protection
The prototype IWV vaccine showed significantly more protection than the Fel-O-Vax ® FIV vaccine when the combined results (36/51) of the prototype vaccine was compared to the combined results (11/24) of the Fel-O-Vax ® FIV vaccine (p = 0.0453) ( Table 1). The percentage protection of each vaccine did not correlate with the level of NAbs produced by the vaccine to the challenge virus with exception to homologous FIV Pet and heterologous FIV Bang . The commercial vaccine produced FIV NAbs more consistently in more vaccinated cats and generally at higher titers than the prototype vaccine ( Table 1). The studies comparing the NAb titers between these vaccines demonstrated quite complex findings [48]. The commercial vaccine induced two-fold higher average NAb titers to homologous FIV Pet but 23% lower percent responders (percent of total animals showed titer ≥10 upon vaccination) with NAbs than the prototype vaccine (Tables 1 and 3). (Table 3 includes the summarized results of Table 1 and three additional NAb results to FIV Dix (subtype A), FIV MD subtype B), and vaccine strain FIV Shi (subtype D). Both vaccines induced more NAb titers to FIV Pet than those to FIV Shi (Table 3). Likewise, both vaccines had similar percent NAb responders to FIV Shi but the average NAb titer to FIV Shi of the prototype vaccine was higher than that of the commercial vaccine. This trend differed from those to FIV Pet (Table 3).
In another study, the dual-subtype FIV Pet -ICV (FL-4) plus FIV Shi -ICV (FIV Shi -FeT-1) vaccine conferred 100% protection against FIV Shi when neither single-virus FIV Pet -ICV nor FIV Shi -ICV was able to achieve such protection [60]. Notably, only 2/4 dual-subtype vaccinated cats had NAb titer to FIV Pet before FIV Shi challenge, whereas none (0/4) had NAbs to FIV Shi . These observations suggested that the dual-subtype vaccine, but not single-virus vaccines, had a synergistic effect on efficacy and perhaps on the types of immunity produced and that the vaccine-induced NAb to FIV Shi did not correlate with protection. The latter assumption is made with a cautionary note since the authors of this article did not evaluate the NAb titers to FIV Shi after challenge. Since their assay did not detect NAb levels of below 10, the possibility exists for extremely low titers of anti-FIV Shi NAbs present before challenge may be detected post-challenge as an amnestic response which can potentially clear the FIV Shi infection. Such anamnestic response has been previously observed in ALVAC study [56].
Most importantly, both vaccines produced low average NAb titers (<5-20) to heterologous strains, and the commercial vaccine produced slightly more percent NAb responders to challenge virus vaccine than the prototype vaccine ( Table 1). The NAb titers to subtype-A FIV UK8 were ≤10 in all vaccinated cats before the challenge even though high Nab titers to FIV Pet which also belong to subtype A. Both vaccines developed low NAb titers to FIV UK8 in a low number of Nab responders. Furthermore, these vaccines conferred either no protection (commercial vaccine) or marginal protection (prototype vaccine). In contrast, both vaccines afforded significant protection to heterologous subtype-B FIV FC1 (Table 1, Study 4) even with the similarly low level of NAbs to FIV FC1 in a small percentage of vaccinated cats.
Passive transfer studies using the pooled sera and purified antibodies from Fel-O-Vax-vaccinated cats were performed on naïve specific-pathogen-free (SPF) cats against challenge infection with either FIV Pet or FIV FC1 [48]. The passive transfer of vaccine antibodies with high titers of anti-FIV Pet NAbs conferred significant protection against homologous FIV Pet but conferred no protection against heterologous FIV FC1 (Table 4). These findings suggested that protection against vaccine virus FIV Pet was mainly mediated by antibody immunity including NAbs. In contrast, the protection against distinctly heterologous subtype-B FIV FC1 was not mediated by the vaccine-induced antibody immunity. The latter finding is noteworthy since prototype vaccine afforded 100% protection against FIV FC1 (Table 1, Study 4b). Since prototype vaccine-induced FIV-specific CTL and polyfunctional T-cell activities [61,62], the likelihood of vaccine-induced T-cell immunity playing a major role in protection against heterologous subtype viruses was a strong possibility.   Table 3 is the summary of Table 1 with emphasis on NAb titers generated by Fel-O-Vax ® FIV vaccine and prototype IWV vaccine. b Additional data provided are on the average NAb titers to FIV Dix [55] and to FIV MD [49], and on the dual-subtype FIV Pet -ICV/FIV Shi -ICV vaccine against FIV Shi challenge [60]. c Vaccine strains (bolded). d Substantial NAb titers (bolded) or significant percent preventative fraction (bolded) when p < 0.05; Not available (-).

The Role of Vaccine-Induced T-Cell Immunity in Prophylactic Protection
The first prime-boost FIV vaccine study held in 1996 provided the first indication of the role of T-cell immunity in vaccine protection [56]. The 2X ALVAC (canary poxvirus)-vectored subtype-A FIV Ville Frache gag/pro/env prime followed by a single ICV-FIV Pet boost conferred slightly more protection against FIV Pet (100% vs. 67%) than those boosted with ALVAC-FIV gag/pro/env; however, only three animals per group were tested to make such conclusion (Table 5 adapted from [56]). Before FIV Pet challenge (50 CID 50 ), all cats in ICV-FIV Pet boost group and two cats in ALVAC-FIV boost group developed FIV antigen-specific cytotoxic T lymphocyte (CTL) and/or T-helper activity but did not develop any detectable anti-FIV Pet NAbs. The anti-p24 Ab titers were very high in the ICV boost group and in the control group that received empty ALVAC vector prime with ICV boost, but none detected in the ALVAC-FIV boost group. At 8 months post-first challenge, the protected/vaccinated cats and one non-vaccinated control cat without additional boost were challenged with a high dose (75 CID 50 ) of recombinant subtype-A gag-pol /B env FIV Bang . All cats in the ICV-FIV Pet group was protected, but no protection was observed in the ALVAC-FIV gag/pro/env group and the single control cat. Before the second challenge, only one protected cat had low levels of FIV Pet and FIV Bang titers, whereas the others had no NAb titer to either virus. All protected cats developed no NAbs to FIV Bang post-FIV Bang challenge, and therefore, such protection was not mediated by amnestic anti-FIV Bang NAbs. The authors concluded that vaccine-induced NAbs and T-cell immunity are important in prophylactic protection against FIV Pet which had Env sequence closely related to the vaccine strain (3% and 1% differences in Env and Gag aa sequences, respectively) [56]. The low levels of anti-FIV Pet NAbs which developed or persisted after FIV Bang challenge, suggested a possibility of the synergistic role of non-NAbs and anti-FIV T-cell immunity in protection [55]. This assumption was strengthened with the recent phase-III HIV vaccine trial (RV144 trial) that showed 31.2% protection in the overall risk group [59]. This protection correlated or associated with high vaccine-induced antibody-dependent cellular cytotoxicity (ADCC) Abs and anti-V2(Env) CD4 + T-cell immunity including polyfunctional CD4 + T cells and CD4 + CTLs [59,63,64]. Moreover, in a failed phase-III AIDSVAX B/E (VAX004) vaccine trial, HIV-specific ADCC-Abs, but not Nabs, were associated with lower infection risk [65,66]. These HIV vaccine trials suggested that ADCC Abs can play a role in vaccine protection when only low levels of NAbs are detected [66,67].
The ALVAC-FIV prime-boost study (Table 5) and the commercial and prototype vaccine trials (Table 1) did not test for vaccine-induced ADCC-Abs. Hence, ADCC-Abs together with anti-FIV T-cell immunity and even low titers of NAbs may play a role in vaccine protection against heterologous FIV challenge. The initial T-cell studies consisted of adoptive transfer (AT) of washed blood cells containing 1-2 × 10 7 peripheral blood mononuclear cells (PBMC) from a subtype-A FIV Pet -IWV vaccinated parent (AT donor) to MHC half-matched offspring (AT recipient) to prevent rejection of the immune cells by the AT recipients ( Table 6, Studies 1-3) [68]. Only MHC half-matched or identical matched recipients of blood immune cells showed partial-to-complete AT protection against homologous FIV Pet challenge. Partial protection was defined as no development of FIV-p24 antibodies (Abs) and a major delay in virus isolation from PBMC throughout 24-32 weeks post-challenge (wpc). None of the AT recipients had p24 Abs at 1 wpc, indicating that the p24 Abs were washed off and the transfer of p24-specific B and the plasma cells were minimal during the adoptive transfer. When the results from Studies 1-3 were combined and compared to the protection rate of the combined control group, significant protection rate was achieved in 10/13 partial-to-complete protection (p < 0.0001) and 6/13 complete protection (p = 0.0013). The combined control group consisted of nonvaccinated/half-matched, nonvaccinated/unrelated and vaccinated/unrelated recipients, and nonvaccinated/non-AT cats. These studies indicated that the AT protection was MHC restricted and that the transferred FIV-specific T cells from donor prevented FIV infection in the AT recipient when the MHC was at least partially matched between the donor and recipient. Above Table 5 adapted and summarized from [56].   Above studies, however, did not demonstrate whether the transferred vaccine-induced memory B cells were working in synergy with the transferred anti-FIV T cells. In recent studies, purified T and B cells from vaccinated AT donors were adoptively transferred to partial-to-compete MHC matched recipients (Table 6, Studies 4-7) [62]. In Study 4 that was based on MHC matching by mixed leukocyte reaction (MLR) ( Table 6, Study 4), 3/4 (75%) AT recipients of MLR-matched immune T cells were protected, whereas none (0/4) of the recipients receiving either MLR-matched immune B cells or unrelated non-immune PBMC were protected. Based on MHC-I, MHC-II, and MLR matching (Study 5), the recipients of the completely-matched immune T cells were all (4/4) protected against homologous FIV Pet challenges, whereas partial protection was observed (2/3) in the recipients each of either partial-to-completely-matched immune CD8 + T cells or CD4 + T cells. In this study, one recipient of partially MHC-matched immune CD8 + T cells, and another recipient of completely MHC-matched immune CD4 + T cells were infected. Thus, complete AT protection was consistently observed when both immune CD4 + and CD8 + T cells were transferred together as a whole T-cell population, and when the donor and recipient were completely MHC matched. This study further suggested that the prophylactic protection of prototype vaccine is mediated by vaccine-induced T-cell immunity which appears to require both CD4 + and CD8 + T-cell immunity to exert complete protection against homologous FIV Pet challenge. In Study 6, all recipients (5/5) of completely MHC-matched immune T cells were protected again, resulting in a combined protection rate of 100% (9/9) in the AT recipients of completely MHC-matched T cells from Studies 5 and 6 (p < 0.0001) ( Table 6). Current results together with those from humoral and passive immunity studies support the concept that the vaccine should induce both B-cell and T-cell immunity in synergy to confer complete protection against FIV strains closely related in the Env sequence to subtype-A and -D vaccine strains.

Determining the Importance of T-Cell Immunity Produced by the Prototype Vaccine
In the humoral immunity studies described earlier, vaccine-induced minimal-to-nil titers of NAbs were detected against heterologous subtype (B and F /C; Table 1) strains. These findings suggested that immunity other than B-cell immunity were important in mounting the protective response observed against viruses of heterologous subtypes. Passive transfer (PT) of purified antibodies did not confer protection in PT recipients when challenged with heterologous subtype-B FIV FC1 , but significant protection was observed when challenged with homologous FIV Pet . This observation demonstrated that FIV FC1 is resistant to Fel-O-Vax-induced Abs and any other Ab-mediated antiviral activity (e.g., ADCC-Abs). This finding also suggested that the immunity other than Ab immunity was the real cause for the protection against heterologous challenges (e.g., FIV FC1 ). This assumption was confirmed when purified T cells from prototype IWV-vaccinated cats showed significant AT protection in (6/9) recipients of completely MHC-matched immune T cells but no protection in the control recipients of non-immune/MHC-unmatched T cells, PBS, or without AT (0/8) (p = 0.009). The authors concluded that the AT protection was mediated by vaccine-induced T cells based on the following three observations: (1) the AT preparation of the purified T cells contained minimal-to-nil contaminating B cells; (2) no vaccine-induced antibodies were detected before or at 3 wpc, the initial time when only one control cat developed FIV immunoblot Abs; and (3) AT protection required complete MHC matching, demonstrating MHC-restricted T-cell immunity instead of MHC-unrestricted Ab immunity [62]. Hence, vaccine-induced T-cell immunity is vital for the prophylactic protection afforded by the prototype IWV vaccine as well as perhaps by the Fel-O-Vax ® FIV vaccine. The latter is most likely the case since Fel-O-Vax-induced Abs had no PT efficacy against FIV FC1 , nevertheless, the vaccine attained significant (71.8-75.0%) prophylactic protection against FIV FC1 ( Table 3).
The FIV-specific T-cell immunity generated by the prototype vaccine consisted of FIV-specific polyfunctional CD4 + T-cell activities (perforin, granzyme A, granzyme H, TNFα, IFNγ, IL-2, IL-4, IL-10), polyfunctional CD8 + T-cell activities (perforin, granzyme A, granzyme B, granzyme H, IFNγ, IL-2), CD4 + CTL activities (perforin, granzymes A and H), CD8 + CTL activities (perforin, granzymes A, B, H), and FIV-specific CD4 + and CD8 + T-cell proliferation responses [61,62]. A note of caution must be placed on IFNγ and possibly on TNFα. Overproduction of these two cytokines may be concerning due to their pro-inflammatory nature that could enhance HIV-1 replication in the host (described in review [69]) as well in primary macrophage, CD4 + cells, and PBMC [70]. In the failed phase-IIb HIV vaccine (STEP) trial, the non-specific IFNγ secretion was associated with increased HIV infection risk [71]. The addition of feline IFNγ together with FIV enhanced the FIV replication in primary PBMC cultures [72]. Likewise, treatment with human IFNγ enhanced HIV-1 replication in human PBMC [73]. Furthermore, in an adjuvant study using the same dual-subtype IWV immunogen as the commercial and prototype vaccines, IWV-vaccinated cats with adjuvants other than FD-1 adjuvant were less protective when their PBMC had higher levels of FIV-specific IFNγ production than those from the prototype IWV-vaccinated cats [74]. However, a moderate production of IFNγ and TNFα together with IL-2, perforin, and granzymes may be beneficial as long as inflammatory cytokines are not overproduced [69].

A Novel Approach to Confirm the Value of T-Cell Immunity in Conferring Protection
The FIV peptides that induce potent CD8 + CTL, CD4 + CTL, and polyfunctional CD4 + and CD8 + T-cell activities, along with strong FIV-specific CD4 + and CD8 + T-cell proliferation responses should be included as vaccine immunogens. The FIV epitopes that induce these protective activities should be short linear peptides for T-cell recognition. T cells only recognize the peptides presented on the MHC-I molecules of infected cells or MHC-II molecules of the antigen presenting cells (APC) (e.g., macrophages, dendritic cells, lesser extent by B cells) [75]. More importantly, such vaccine peptide epitopes should be conserved among FIV subtypes and induce protective T-cell immunity without generating escape variants. Therefore, the protective conserved FIV epitopes for an effective vaccine must be essential to the viral survival since any substantial mutation(s) should affect the fitness of the virus [76][77][78].
The conserved epitopic (8-15mer aa) sequences among the FIV viruses within the same subtype and even among the different subtypes provide numerous potential intra-subtype and inter-subtype conserved epitopes but do not identify epitopes which resist mutation or induce protective T-cell responses. One approach in focusing potential T-cell epitopes on FIV proteins is to identify those epitopes that are conserved among two or three AIDS lentiviruses (HIV-1, FIV, SIV) and that possess the ability to induce anti-FIV T-cell activity(s) [79,80]. Furthermore, such conserved epitopes are most likely to have withstood the evolutionary pressure in their respective host but have been retained as part of the virus since any mutation(s) most likely affected the survival of the virus. The first step in pursuing this approach of T-cell epitope selection is to identify the most conserved FIV/HIV-1 proteins.
The most conserved proteins on the FIV was determined by comparing the sequences of HIV-1 and FIV and identified to be Pol-RT (polymerase-RT, 72.0%) followed by Pol-IN (Pol-integrase, 64.7%) and Gag-CA/p24 (Gag-capsid or p24, 63.1%) [79]. This evolutionary approach in the assessment of conserved AIDS virus epitopes was validated by the findings of others who compared genes and amino acid (aa) sequences of domestic-cat FIV to wild-cat FIVs. The genetic divergence of lion FIV PLe subtype E compared to Pallis-cat FIV Oma , puma FIV Pco , and domestic-cat FIV determined that FIV gag/Gag and pol/Pol genes and aa are more conserved than FIV vif /Vif, orfA/OrfA, and env/Env, suggesting that host selection or immune pressure may more readily mutate vif, orfA, and env genes without affecting viral replication [19]. Similar analyses have been performed with different FIV isolates from domestic cats, which also shows that Gag sequences (Pol sequence not tested) are more conserved than Env sequences [58]. Similarly, protective conserved HIV-1 epitopes were identified predominantly on Gag and Pol when compared to those on Env, Nef and accessory proteins [81][82][83][84]. Hence, major changes by mutations in the gag and pol genes may affect the survival of the virus. Since Gag-p24 and Pol-RT antigens are not involved in inducing anti-FIV/HIV neutralizing antibodies (NAbs), immunoprotective epitopes on Gag and Pol proteins may in place induce FIV/HIV-specific T-cell immunity such as anti-FIV/HIV CD8 + cytotoxic T lymphocytes (CTLs), CD4 + CTLs, and polyfunctional T cells [85]. However, simply vaccinating with recombinant Gag and/or Pol proteins or with a vector containing FIV gag/pol/env genes have been unsuccessful in conferring protection against FIV challenge [41,49,86]. Such results suggest that selecting protective T-cell epitopes from the Gag-p24 and Pol-RT may be a better strategy since expression of whole viral proteins do not appear to confer protection. The need for such a selection also infers that viral proteins are composed of epitopes that enhance viral infection as well as epitopes that readily mutate to escape vaccine immunity [79,85,87,88].
The former case has been demonstrated in a preliminary in vitro study which demonstrated the presence of FIV enhancing epitope nearby a protective T-cell epitope on FIV p24 [89].
The T-cell functional parameters derived from those observed in the protected IWV-vaccinated cats have been used to identify the conserved T-cell epitopes on FIV p24 and RT that stimulate anti-FIV/HIV T-cell activities in the PBMC of both HIV + human subjects and IWV-vaccinated cats [87,88]. Two FIV p24 peptides and two FIV RT peptides were selected by this method [89]. These peptides were combined in a multiple antigenic peptide (MAP) formulation with a lysine backbone linked to four identical peptides. The combined MAP vaccine consisted of four MAPs (100-120 µg per MAP; a total of 400-480 combined MAPs) in FD-1 adjuvant [89]. The MAP vaccine conferred protection in 15/19 vaccinated cats against subtype-B FIV FC1 challenge, whereas 10/16 control cats were infected (% Preventable Fraction of 66.3%, p = 0.0181). Since no Env peptide was used in this study, T cells are most likely mediating the vaccine protection against FIV. In support of this assessment, the protected/vaccinated cats developed strong FIV-specific T-cell immunity to the p24 and RT peptide epitopes in the MAP vaccine before challenge [89]. This preliminary study demonstrates that the T-cell immunity induced by the prototype vaccine is essential in the successful prophylaxis and more likely this may be the case for Fel-O-Vax ® FIV vaccine.

Summary of Lessons Learned in Developing a Commercial and Prototype FIV Vaccines
The major lesson learned from understanding the immune mechanisms of protection provided by the prototype vaccine is the need for the vaccine to produce high levels of multiple anti-FIV T-cell responses to confer protection against all subtypes of FIV. Much like HIV-1 vaccine research, current efforts on developing a T cell-based FIV vaccine has been insufficient since the major focus of vaccine research has been to develop a B cell-based or an Ab-based FIV vaccine [31, 41,49,80,86]. Furthermore, the failed vaccine trials to develop a consistent and high levels of anti-FIV T-cell immunity used vector vaccines with FIV gag/pol containing constructs, inactivated whole-virus (FIV UK8 , FIV UT113 ) vaccines, and proviral DNA RT and IN which should have generated FIV Gag-and Pol-specific T-cell responses [41,49,56]. Therefore, one major challenge is to weed out non-protective, replication enhancing epitopes and to select for evolutionary conserved (non-mutable), protective T-cell epitopes on FIV Gag and Pol. Another challenge is to identify an approach which combines T-cell epitopes of FIV Gag and Pol with broad NAb (bNAb)-and ADCC-Ab-inducing B-cell epitopes on FIV Env, without inducing blocking Abs to the NAb and ADCC epitopes and without stimulating excessive CD4 + T-cell activation and cytokines which enhance viral replication [85,89]. Overall, the task of developing FIV and HIV vaccines with selected protective B-and T-cell epitopes will require major innovations at the selection of the epitopes, developing a strategy in combining of the B-cell and T-cell epitopes, and choosing the appropriate vaccine delivery platform targeting the APCs. Lessons learned from the commercial and the prototype FIV vaccines indicate that most of their vaccine protection was mediated by anti-FIV T cell immunity more than anti-FIV Ab immunity (NAbs, ADCC-Abs). However, inducing both T-and B-cell immunity against all subtypes of FIV will be an ideal scenario. Thus, it is anticipated that a highly effective HIV vaccine will need to produce potent anti-HIV T-cell immunity in addition to protective anti-HIV Abs (NAbs, bNAbs, ADCC-Abs), but without concomitant production of blocking antibodies and HIV-enhancing cytokines.
Author Contributions: B.S. and J.K.Y wrote the review.