NK cells are part of an innate immune system. It is thought to lack antigen-specificity and memory. Generally antigen-specific memory is defined as an augmented and faster response to antigen re-exposure.
The data I have seen so far suggest that NK cells do have much wider antigen-specificity but I am still not entirely convinced that NK cells have a long-term memory as T cells do.
Here is a new article from JEM that points to this conclusion. Now, this study came from Lewis Lanier's lab. Initial studies of NK cells memory was based on experiments with Ly49H+ NK cells and mouse CMV antigen m157. In this new study, however, the authors has examined antigen-specificity and memory formation for Ly49D+ NK cells specific for alloantigen H-2Dd.
In this model system, purified Ly49D+ NK cells were transferred into Ly49D and Ly49H double deficient B6 host mice depleted of CD8 T cells (to prevent donor NK cell rejection). The host mice were then challenged with H-2Dd expressing splenocytes (from balb/c mouse). Donor NK cells showed no response to alloantigen alone. However, when host mice were exposed to m157-deficient mouse CMV infection in conjunction with H-2Dd expressing splenocytes (to mimic conditions for an actual clinical tissue transplantation when human CMV becomes re-activated), the authors observed expansion of Ly49D+ NK cells (but at day 28 post infection, there was minimal difference between responding and non-responding NK cell populations).
This expansion of Ly49D+ NK cells was alloantigen-specific, required simultaneous infection with mouse CMV and was dependent of IL-12.
Interestingly, expansion of Ly49D+ NK cells was abolished when NK cells expressed another inhibitory molecule Ly49A. Both Ly49D and A recognize H-2Dd.
Surprisingly, at day 28 post infection, there was no difference between Ly49D+/A- and Ly49D+/A+ NK cell population and upon secondary challenge with alloantigen, Ly49D+/A- NK cell expanded at the same rate as during primary challenge (compare Fig.1 and Fig. 4).
An ex vivo analyses of effector functions of NK cells showed that presence of Ly49A inhibited NK cell's effector functions in both naive and primed NK cells (and only in presence of H2-Dd). Interestingly, in absence of Ly49A, difference between naive and primed NK cells disappeared.
To show that initial priming leads to antigen-specific enhancement during the secondary response (rather than general increase in NK cell responsiveness), the authors primed the host mice either with (a) m157-deficient mouse CMV infection in conjunction with H-2Dd expressing splenocytes, and then challenged with wild type mouse CMV, or (b) primed the host mice with wild type mouse CMV and then challenged with m157-deficient mouse CMV infection in conjunction with H-2Dd expressing splenocytes. Strangely, In both situations only NK cells primed during initial antigenic stimulation undergo secondary expansion but independent of secondary antigen-specificity. However, for some reason naive NK cells failed to expand in response to specific antigenic challenge in this setting (compare Fig. 1 and Fig. 6).
Finally, the authors showed that primed NK cells showed slightly better protection against wild type mouse CMV in adoptive transfer experiment.
While I agree with the authors that NK cells have antigen-specificity, this study did not convince me that primed NK cells show enhanced secondary recall response (enhanced proliferative response) as typically observed for T cells.