While it has been known for more than two decades that there is a global upregulation of the inflammatory and innate immunity pathways in the brains of AD patients, it has not been clear what role, if any, the inflammatory response plays in AD neurodegeneration 
. Over the years a number of findings have raised the possibility that the inflammatory response plays an active role in promoting neurodegeneration. One of the first indications that it could contribute to disease progression came from epidemiological studies indicating that prolonged treatment with anti-inflammatory drugs is correlated with a reduced risk of developing AD 
. Experiments using either mammalian tissue culture cell or whole organism AD models have also pointed to a connection between Aβ induced neurodegeneration and activation of the inflammatory response 
. These studies have shown that expression of the cytokines IL-1, IL-6 and TNF-α are upregulated by the human Aβ peptide and that induction of these cytokines both by Aβ and by other mechanisms can have deleterious effects on neuronal cell viability and cognition. Moreover, downregulation of the innate immune response pathway in a mouse AD model by treatment with anti-inflammatory drugs or disruption of the tumor necrosis factor death receptor gene was found to ameliorate the effects of ectopic Aβ 
. On the other hand, not all findings have been consistent with the notion that the inflammatory response/innate immunity pathways promote neurodegeneration. In fact, a number of reports have suggested instead that these pathways play a beneficial rather than deleterious role 
. Thus, it is not clear at this point whether inflammation is a causative agent in the process of neurodegeneration, is an essentially benign response to AD, or is actually beneficial and acts to retard the progression of the disease 
The findings reported here strongly support the hypothesis that the Tl→NFκB innate immunity pathway plays a critical role in mediating the neuropathological effects of the human Aβ42 polypeptide. We identified this pathway in an unbiased genetic screen for mutations that either promote or suppress the neurpathological effects of the Aβ42 polypeptide on the Drosophila eye. One of the suppressors recovered in our genetic screen was the Tl receptor, which is a key component of the fly innate immunity pathway.
Several lines of evidence demonstrate a direct connection between the neuropathological effects of Aβ42 and the activation of the Tl
→NFκB signaling cascade. First, the Aβ42 polypeptide induces the accumulation of one of the well known downstream transcriptional targets of the Tl
→NFκB pathway, the fly IκB homolog cact
. We also found that the levels of the fly NFκB protein Dl are upregulated in pGMR-Aβ42 transgenic flies as well. Thus, the response of flies to the neurotoxic effects of the Aβ42 polypeptide appears to mimic the upregulation of the inflammatory pathways evident in both AD patients and in mouse AD models 
. Moreover, the extent of cact
induction depends upon the dose of the Aβ42 polypeptide. In flies carrying two copies of the pGMR-Aβ42 transgene cact
accumulation is substantially elevated in virtually every fly. By contrast, in flies carrying only a single pGMR-Aβ42 transgene there is typically a smaller increase in the level of the Cact protein, and in a subset of the transgenic flies little change in Cact accumulation is evident.
Second, as would be predicted if activation of the inflammatory response mediates the neuropathological effects of the Aβ42 polypeptide, we found that loss-of-function mutations in Tl
dominantly suppress neurodegeneration of the eye induced by ectopic Aβ42. Conversely, a Tl
gain-of-function allele that signals constitutively independent of ligand exacerbates the degenerative effects of Aβ42. In this context, it is interesting to note that Tl
was identified in the screen of EP insertions 
. The Tl
insertion was found to be a strong enhancer of Aβ42 induced degeneration of the eye and it substantially upregulates the expression of Tl
mRNA. Thus, it is possible to enhance the sensitivity of the fly to the pathological effects of Aβ42 not only by increasing the activity of the Tl receptor, but also by increasing the amount of the receptor.
Third, Aβ42 induced neurodegeneration is mediated by the downstream target for the Tl→NFκB signaling cascade, the dl transcription factor. This is most clearly demonstrated in pGMR-Aβ42 flies that are homozygous for dl mutations. When there is only a single copy of the transgene, suppression of the Aβ42 rough eye phenotype by the dl mutation is nearly complete and the eyes resemble wild type. Degeneration of the eye is dependent on the dose of the Aβ42 polypeptide and is much more severe when there are two copies of the pGMR-Aβ42 transgene. However, even in this case strong suppression is observed when the transgenic flies are homozygous for a dl loss of function mutation. Moreover, this is not due to genetic background as three independent dl alleles substantially reduce the much more extreme neuropathological effects produced by two copies of the pGMR-Aβ42 transgene. The effects of dl are not limited to homozygous mutant flies; we also found that the pGMR-Aβ42 induced disruptions in eye development are suppressed when the flies are heterozygous for a dl mutation. In this case, suppression is less complete than that observed in flies that are homozygous for the same dl mutation.
Fourth, like Tl
, mutations in three other components of the Tl
→NFκB signaling cascade, tub
and the partner of the dl
transcription factor, dif
also dominantly suppress the degenerative effects of Aβ42 on eye development. In addition, one of the chromatin modifiers recovered in the EP screen, Dsp1, functions as a Dl dependent co-repressor 
. It should be noted that not all genes in the Tl
→NFκB signaling cascade show dominant genetic interactions with pGMR-Aβ42. We tested three different mutations in the Tl
; however, no alterations in the rough eye phenotype were observed. Similarly no effects were observed with a mutation in either cact
or the adaptor protein myd88
. Since both cact
are thought to be canonical cell autonomous components of the Tl
→NFκB signaling pathway a plausible explanation is that these genes are not haploinsufficient in the eye assay. In the case of spz
we cannot exclude the possibility that some other effector molecule mediates the activation of the Tl
While our results clearly demonstrate that the Tl→NFκB pathway plays a key role in facilitating the degenerative effects of the Aβ42 polypeptide on the eye, the evidence that it also mediates the Aβ42-dependent reduction in life span is less clear cut. The Tl allele that is the strongest suppressor in the eye assay, Tlr4, appears to extend the life span of UAS-Aβ42/elav-GAL4 flies by about a quarter. However, much more modest effects, if any, were observed for two other Tl alleles, as well as for mutations in tub and pll. Several factors could explain why these other mutations didn't greatly extend the life span. For one, it is possible that the mutations we tested do not sufficiently reduce the overall activity of the Tl→NFκB signaling pathway as heterozygotes to have much of an impact on the average life span of the UAS-Aβ42/elav-GAL4 flies. Both of the Tl alleles are hypomorphs and neither was a strong suppressor in the eye assay. This is also true for the tub and pll mutations. Given the very strong suppression of the rough eye phenotype observed pGMR- Aβ42 flies that are homozygous for dl mutations, it is possible that we would observe a significant suppression of the life span defects in mutants that lack tub or pll altogether. Another difference is that the life span assay is likely to be a much more indirect measure of the degenerative effects of the Aβ42 polypeptide. While the deleterious effects of the Aβ42 polypeptide on individual cells in each ommatidia can be observed directly, life span depends upon a complex combination of genetic background, environmental conditions and chance circumstances and the neurodegeneration induced by Aβ42 is just one element among many that determine mortality rates. Additionally, there may be specific neuronal circuits whose activity must be maintained above a critical threshold in order to prolong survival. In this case, even if the overall neurodegenerative effects of the Aβ42 polypeptide in the CNS are greatly ameliorated by reducing the activity of the Tl→NFκB pathway, this might not be sufficient to substantially increase life span.
It is also possible that reducing the activity of the Tl→NFκB signaling pathway by two-fold (in animals heterozygous for a null allele) will not be sufficient in itself to strongly suppress the deleterious effects of the Aβ42 polypeptide on life span. One reason why this might be the case is that neurodegenerative effects of activating the Tl→NFκB signaling pathway in the CNS likely depend upon triggering other, cell autonomous pathways or processes after a prolonged or chronic inflammatory response. In this case, it might be necessary to simultaneously alter the activity of one or more of these interacting pathways in addition to Tl→NFκB.
Some potential candidates for the participating pathways were uncovered in our genetic screen for genes that modulate the Aβ42 induced rough eye phenotype. For example, we found that a small deficiency, Df(3)H99,
which removes three pro-apoptotic genes 
, is a strong suppressor of the rough eye phenotype. This would suggest that the cell death pathway promotes the neurodegenerative effects of the human Aβ42 polypeptide in the Drosophila
eye. Consistent with an induction of apoptosis, we found that a mutation in croquemort
, which is required for the phagocytosis of apoptotic cells in Drosophila 
, dominantly enhances the rough eye phenotype. A role for apoptosis in the degenerative process would also be consistent with studies on brains of AD patients which have shown that markers for apoptosis are elevated. In this regard it is interesting to note that c-Jun N-terminal kinase (JNK) signaling cascade, which is thought to promote apoptosis, is upregulated in flies by both the Tl
→NFκB and IMD innate immunity pathways 
. Moreover, a number of observations support a possible role for the JNK pathway in Aβ42 induced neurodegeneration in the fly as well. First, one of the chromatin modifiers identified in the EP screen was the histone deacetylase Rpd3 (Hdac1)
. Rpd3 histone deacetylase activity is required to downregulate the JNK pathway and it functions in complexes with Sin3A and Sap130 
. EP insertions that disrupt the function of Rpd3, Sin3A and Sap130 were found to enhance the rough eye phenotype. Second, we found that two mutations in pucker
), which encodes a tyrosine phosphatase that antagonizes the JNK kinase enhance the rough eye phenotype. It has also recently been suggested that activation of the apoptosis cascade, by JNK, promotes the cleavage and hyperphosphorylation of Tau 
. Tau is a microtubule-associated protein which accumulates in large aggregates in the neurofibrillary tangles of AD patients 
. The Tau protein can also have neurotoxic effects independent of Aβ42 
. The Aβ42 induced rough eye phenotype seems to be connected to the fly tau
gene as well as since we found that it is enhanced by two independent tau
mutations. Further studies will be required to elucidate the role of these and other genes in modulating the neuropathological effects of Aβ42 and their connection to the functioning of the Tl
→NFκB pathway. It will also be important to determine if it is possible to more substantially ameliorate the life span defects of Aβ42 flies by simultaneously manipulating the activity of the Tl
→NFκB pathway and one of these other pathways.