Sandage and Tammann (e.g. Sandage & Tammann (1990)) have long argued that for values of H0 < 55, the Virgo Cluster must be at a distance of about 22 Mpc since the cosmic velocity of Virgo is ~ 1170 km/s. Additionally, M87 is usually claimed to be representative of the Virgo distance since its recessional velocity is close to the Virgo average, it is projected near the center of the cluster, and it appears to be at the bottom of the potential well as evidenced by an X-ray cloud typical of X-ray clusters.
A reasonable question to ask is "what must one give up to derive a PNLF distance of 22 Mpc for M87?" The simple answer is that the constancy of the maximum PN luminosity must be abandoned. And, it must be abandoned in a big way - by ~ 0.9 mag.
The astrophysical question becomes "can the PNLF be enhanced by 0.9 mag, and if so, what are the observable manifestations?" The latter turn out to be severe and obvious. Models of the PNLF by Jacoby (1989), Méndez et al. (1993), and Stanghellini (1995) all show that a shift in the PNLF as large as 0.9 mag is almost impossible because there is inadequate UV flux being radiated by any reasonable collection of central stars. A 0.9 mag enhancement demands that the central stars be very massive, originating from 4-5 M progenitors. A population this young (< 100 Myrs) should be evident.
The presence of a very young population in ellipticals is not unreasonable; many ellipticals exhibit evidence for gas and dust (Goudfrooij (1995)). Usually, though, the mass fraction of the elliptical involved in the young population is tiny. To enhance the PNLF by 0.9 mag, though, requires that ~ 40% of the total luminosity of the galaxy comes from the young population. (The 40% value derives from the PN production rate for bright PN in M87. This rate is 40% of the maximum value that is attainable under the assumption that all stars produce PN. Since some stars may not produce PN, the actual fraction of luminosity from young stars could be higher.)
A trivial observational test for a young population is galaxy color. A young population (< 100 Myr) has a color (B - V) < 0.0 while an old population (> 11 Gyr) has a color (B - V) ~ 1.0. Thus a mix where 40% of the luminosity derives from a young population has a color (B - V) ~ 0.6. This color is the direct consequence of insisting that M87 be at 22 Mpc, combined with stellar evolution predictions (e.g. Ciotti et al. (1991)); Vassiliadis & Wood (1994)) for the initial-to-final mass relation that was devised to explain the observed relation (Weidemann & Koester (1983)).
Comparing the predicted color of 0.6 to the observed (B - V) ~ 1.0 (which depends slightly on radial position), it is evident that a sizable young population is untenable. Thus, we are forced to accept that either M87 is much closer than 22 Mpc in order to alleviate the pressure to enhance the PNLF luminosity, or that the initial-to-final mass relation predicted by stellar evolution models and observed in young Galactic clusters is seriously flawed in such a way that old stars can produce massive central stars. Since there is no evidence for the latter, the more likely conclusion is that the existence of PN in M87 at the observed apparent magnitudes demands a distance much smaller than 22 Mpc. (It is possible to push the distance as large as 17 Mpc before the implied population color becomes inconsistently blue.)