pathways involved in drug responses. In addition, it points to potential target 
genes that might underlie dependence, such as neuropeptide genes (e.g., 
proenkephalin and prodynorphin), neurotransmitter enzymes (e.g., TH), 
neurotrophic factors and their receptors (e.g., NGF, BDNF, and trkB), as well 
as genes encoding proteins involved in modifying the extracellular matrix 
(e.g., collagenase and stromelysin). At the structural level these genes may 
contribute to a general synaptic growth response that modifies the behavior 
of a neural circuit over the long term (i.e., the CIE genes might elicit a chain of 
transcriptional events that culminate in the sprouting of axons with concomitant 
changes in synaptic density, size, and efficiency). 
Several criticisms have been leveled at the significance of CIE genes. First, 
many researchers find it difficult to see any specificity in the system since it is 
apparently induced quite ubiquitously in most cell types by a diverse array of 
agents (reviewed in Morgan and Curran 1991a). Thus, it might be argued that 
this is merely a cellular stress mechanism, analogous to the induction of heat 
shock proteins, designed to protect neurons when they are at risk for neurotoxic 
damage. This argument is becoming less valid. First, these genes are induced 
by normal physiological stimuli, not just the pharmacological and traumatic 
stimuli used in early experiments. Second, many new CIE genes have been 
discovered, some of which show restricted sites of expression and are 
preferentially induced by certain agents in a particular cell type. Third, the 
complex, differential, posttranslational modification of CIE gene products, 
combined with their staggered induction patterns and multiple dimerization 
properties, provides a startling array of diversity. That is, it is likely that diversity 
is generated at one level by a combination of a relatively few restrictively 
induced CIE gene products interacting with many more ubiquitously induced 
members. In addition, specificity is determined by various protein-protein 
interactions between induced and constitutive proteins as well as their 
subsequent posttranslational modification. The presence or absence of 
particular kinases, resident transcription factors, as well as specific signaling 
pathways will dictate this latter aspect of the response. 
Second, although the scenario outlined above nicely accounts for several 
aspects of neurophysiology and neuropathology, it is unproven except for 
correlations. Herein lies a fundamental difficulty in pursuing the function 
of CIE genes in higher vertebrates, namely, the ability to specifically eliminate 
them or their action in vivo. Since many of the CIE genes encode proteins 
that are expressed naturally during development, it is likely that if they 
were to be eliminated by homologous recombination the animals would die 
in utero. Therefore, more subtle ways must be designed to impair their 
function or expression. At one level this could be viewed as a novel branch 
of pharmacology (i.e., designing agents that specifically impair the CIE 
48 
