TO THE EDITOR Serum-stimulation of quiescent (G0) keratinocytes initiates a temporally regulated program of transcriptional activity required for G0/G1 transit and subsequent entry into the proliferative cycle (Qi and Higgins, 2003). Expression profiling of such “activated” keratinocytes identified physiologically relevant subsets of cell cycle/growth state-regulated genes (Gromov et al., 2002; Gazel et al., 2003). Indeed, non-cycling human (HaCaT) keratinocytes express a differentiated (i.e., super-basal) genetic signature, whereas the serum-stimulated transcriptome approximates that of transient amplifying cells (Pivarcsi et al., 2001; Lemaitre et al., 2004). Clearly, the associated transcriptional responses dictate epidermal cell lineage commitments by impacting the expression of pathway-relevant genes (Banno et al., 2004; Lemaitre et al., 2004). This report provides early evidence regarding the comprehensive inventory of genes expressed by human HaCaT-II4 keratinocytes during the initial stage of cell-cycle re-entry. Re-introduction of serum to quiescent HaCaT-II4 cells stimulates G0 exit and residence in a short-lived “activated G0 substate” (i.e., the kinetically defined G0→G1 transition state) (Qi et al., 2006). Microarray analysis of quiescent and 2 hours fetal bovine serum (FBS)-“ activated” HaCaT-II4 cells defined the transcriptional signature of this early G0→G1 window. A total of 54,675 expressed sequence-tagged genes were analyzed with 41,083 directly compared for groups A (quiescent) and B (2 hours FBS-stimulated) and a total of 35,991 reproducibly assessed for all three experimental conditions (i.e., 66% of the total sequences available; this includes group C [FBS for 2 hours in the presence of puromycin included as a first approximation of the immediate-early response cluster]). Genes exhibiting statistically significant (analysis of variance) changes (two-fold increase or decrease) distributed as follows: 1151 for A versus B, 1241 for A versus C, and 1319 for B versus C. Among the most prominently upregulated mRNA transcripts were those encoding proteins involved in the initial growth response (EGR1-4), extracellular matrix remodeling and tissue invasion (uPA, uPAR, tPA, SERPINE1 (PAI-1), PAI-2, MMP-2, MMP- 12, CYR61), transcription (Myc, Fos, Jun, KLF4, AT3, p300/pTAF), signal transduction (DUSP1, -4, -8, -10, MAPK3, TGF-α), proliferation (GADD45a, GADD45b, CDK7, cyclins), and apoptosis (CASP9, MCL2) (e.g., Figure 1a). Reverse transcription-PCR and northern blotting validated the expression data for selected genes. When more stringent criteria were applied to data filtering (i.e., set threshold of ≥ 10-fold increase), 79 genes were identified (Table 1) of which 75 also partitioned to the puromycin-resistant subset. Rank order analysis indicated that PAI-1 (SERPINE1) and the uPA receptor (PLAUR) were the most significantly elevated transcripts. uPA increased (by 12-fold) as well (by microarray and northern analyses), although maximal uPA expression occurred several hours later than PAI-1. Additional genes that comprise the “tissue repair” subset and that were upregulated > 10-fold within the first 2 hours included DTR, IL8, EREG, HBEGF, IL11, CTGF, LIF, IL6, IL1A, HAS3, SERPINB1, and TGFA. Northern blotting confirmed that PAI-1 transcripts were low to undetectable in quiescent HaCaT-II4 cells, peaked in puromycin-resistant manner 2 hours after serum addition (during residence in the initial activated G0 substate; Qi et al., 2006), and then rapidly decreased (Figure 1b and c). Expression required EGFR/MEK/rho-ROCK signaling during the G0→G1 transition (Figure 1d and e). Actinomycin chase/mRNA decay and temporal assessments of mRNA abundance indicated, moreover, that PAI-1 transcripts were substantially reduced (from a maximum at 2 hours) as early as 4 hours post-stimulation decreasing further by 6–8 hours post-stimulation (i.e., approximately mid-G1) likely due to E2F1-mediated suppression (Koziczak et al., 2001), consistent with a narrow window of serum-initiated transcription and short mRNA half-life (1.5–2 hours) (Mu et al., 1998; White et al., 2000). Figure 1 A significant fraction of FBS-induced transcripts encode proteins involved in cell proliferation, transcriptional reprogramming, and control of pericellular proteolysis Table 1 Genes exhibiting a ≥ 10-fold increase in expression 2 hours after serum stimulation of quiescent HaCaT-II4 cells Activation of a wound repair transcript profile appears to be a general response to serum addition (Iyer et al., 1999; this study). The present findings indicate, furthermore, that PAI-1 is the most prominent member of the keratinocyte “serum response transcriptome”. Several SERPINS (i.e., PAI-1, protease nexin-1), in fact, modulate the complex process of injury resolution through control of focalized plasminmediated matrix remodeling, cell migration, and apoptosis (e.g., Bajou et al., 2001; Li et al., 2000; Deng et al., 2001; Degryse et al., 2004; Rossignol et al., 2004; Wang et al., 2005). Targeted PAI-1 knockdown/overexpression and protein add-back approaches, moreover, support the contention that PAI-1 participates within the global program of injury to coordinate cycles of cell-to-substrate adhesion/detachment and/or maintain a stromal “scaffold” to satisfy the prerequisites for both G1/S transition and effective cellular migration (Planus et al., 1997; Chazaud et al., 2002; Palmeri et al., 2002; Providence et al., 2002; Czekay et al., 2003; Providence and Higgins, 2004). PAI-1 is also expressed at high levels in senescent cells where it likely interferes with uPA-dependent growth factor activation (Mu et al., 1998; Kortlever et al., 2006). Certain “senescence-associated” genes (i.e., p16INK4a, PAI-1) may actually function in the wound repair program by inhibiting proliferation while promoting migration (Chan et al., 2001; Ploplis et al., 2004; Darbro et al., 2005; Kortlever et al., 2006; Natarajan et al., 2006). Indeed, keratinocytes at the leading edge during wound re-epithelialization are less mitotically active than cells more displaced from the motile front and express relatively high levels of PAI-1 (Garlick and Taichman, 1994; Jensen and Lavker, 1996; Providence and Higgins, 2004). Collectively, these data suggest that PAI-1 may regulate the temporal cadence of cell-cycle progression in replicatively competent cells as part of the injury repair program.